date:20210222

Sorry, last line should have been:
"intended to get an impression of how widespread ***knowledge of*** DPDK's
core operating inefficiency is",
not:
"intended to get an impression of how widespread DPDK's core operating
inefficiency is"

On Tue, Feb 23, 2021 at 8:22 AM Etienne-Victor Depasquale 
wrote:

> Beyond RX/TX CPU affinity, in DANOS you can further tune power consumption
>> by changing the adaptive polling rate.  It doesn’t, per the survey, "keep
>> utilization at 100% regardless of packet activity.”
>>
> Robert, you seem to be conflating DPDK
> with DANOS' power control algorithms that modulate DPDK's default
> behaviour.
>
> Let me know what you think; otherwise, I'm pretty confident that DPDK does:
>
>> "keep utilization at 100% regardless of packet activity.”
>
>
> Keep in mind that this is a bare-bones survey intended for busy,
> knowledgeable people (the ones you'd find on NANOG) -
> not a detailed breakdown of modes of operation of DPDK or DANOS.
> DPDK has been designed for fast I/O that's unencumbered by the trappings
> of general-purpose OSes,
> and that's the impression that needs to be forefront.
> Power control, as well as any other dimensions of modulation,
> are detailed modes of operation that are well beyond the scope of a
> bare-bones 2-question survey
> intended to get an impression of how widespread DPDK's core operating
> inefficiency is.
>
> Cheers,
>
> Etienne
>
> On Mon, Feb 22, 2021 at 10:20 PM Robert Bays  wrote:
>
>> Beyond RX/TX CPU affinity, in DANOS you can further tune power
>> consumption by changing the adaptive polling rate.  It doesn’t, per the
>> survey, "keep utilization at 100% regardless of packet activity.”  Adaptive
>> polling changes in DPDK optimize for tradeoffs between power consumption,
>> latency/jitter and drops during throughput ramp up periods.  Ideally your
>> DPDK implementation has an algorithm that tries to automatically optimize
>> based on current traffic patterns.
>>
>> In DANOS refer to the “system default dataplane power-profile” config
>> command tree for adaptive polling settings.  Interface RX/TX affinity is
>> configured on a per interface basis under the “interfaces dataplane” config
>> command tree.
>>
>> -robert
>>
>>
>> > On Feb 22, 2021, at 11:46 AM, Jared Geiger  wrote:
>> >
>> > DANOS lets you specify how many dataplane cores you use versus control
>> plane cores. So if you put a 16 core host in to handle 2GB of traffic, you
>> can adjust the dataplane worker cores as needed. Control plane cores don't
>> stay at 100% utilization.
>> >
>> > I use that technique plus DANOS runs on VMware (not oversubscribed)
>> which allows me to use the hardware for other VMs. NICS are attached to the
>> VM via PCI Passthrough which helps eliminate the overhead to the VMware
>> hypervisor itself.
>> >
>> > I have an 8 core VM with 4 cores set to dataplane and 4 to control
>> plane. The 4 control plane cores are typically idle only processing BGP
>> route updates, SNMP, logs, etc.
>> >
>> > ~Jared
>> >
>> > On Sun, Feb 21, 2021 at 11:30 PM Etienne-Victor Depasquale <
>> ed...@ieee.org> wrote:
>> > Hello folks,
>> >
>> > I've just followed a thread regarding use of CGNAT and noted a
>> suggestion (regarding DANOS) that includes use of DPDK.
>> >
>> > As I'm interested in the breadth of adoption of DPDK, and as I'm a
>> researcher into energy and power efficiency, I'd love to hear your feedback
>> on your use of power consumption control by DPDK.
>> >
>> > I've drawn up a bare-bones, 2-question survey at this link:
>> >
>> > https://www.surveymonkey.com/r/J886DPY.
>> >
>> > Responses have been set to anonymous.
>> >
>> > Cheers,
>> >
>> > Etienne
>> >
>> > --
>> > Ing. Etienne-Victor Depasquale
>> > Assistant Lecturer
>> > Department of Communications & Computer Engineering
>> > Faculty of Information & Communication Technology
>> > University of Malta
>> > Web. https://www.um.edu.mt/profile/etiennedepasquale
>>
>>
>
> --
> Ing. Etienne-Victor Depasquale
> Assistant Lecturer
> Department of Communications & Computer Engineering
> Faculty of Information & Communication Technology
> University of Malta
> Web. https://www.um.edu.mt/profile/etiennedepasquale
>


-- 
Ing. Etienne-Victor Depasquale
Assistant Lecturer
Department of Communications & Computer Engineering
Faculty of Information & Communication Technology
University of Malta
Web. https://www.um.edu.mt/profile/etiennedepasquale

Re: DPDK and energy efficiency

>
> Beyond RX/TX CPU affinity, in DANOS you can further tune power consumption
> by changing the adaptive polling rate.  It doesn’t, per the survey, "keep
> utilization at 100% regardless of packet activity.”
>
Robert, you seem to be conflating DPDK
with DANOS' power control algorithms that modulate DPDK's default behaviour.

Let me know what you think; otherwise, I'm pretty confident that DPDK does:

> "keep utilization at 100% regardless of packet activity.”


Keep in mind that this is a bare-bones survey intended for busy,
knowledgeable people (the ones you'd find on NANOG) -
not a detailed breakdown of modes of operation of DPDK or DANOS.
DPDK has been designed for fast I/O that's unencumbered by the trappings of
general-purpose OSes,
and that's the impression that needs to be forefront.
Power control, as well as any other dimensions of modulation,
are detailed modes of operation that are well beyond the scope of a
bare-bones 2-question survey
intended to get an impression of how widespread DPDK's core operating
inefficiency is.

Cheers,

Etienne

On Mon, Feb 22, 2021 at 10:20 PM Robert Bays  wrote:

> Beyond RX/TX CPU affinity, in DANOS you can further tune power consumption
> by changing the adaptive polling rate.  It doesn’t, per the survey, "keep
> utilization at 100% regardless of packet activity.”  Adaptive polling
> changes in DPDK optimize for tradeoffs between power consumption,
> latency/jitter and drops during throughput ramp up periods.  Ideally your
> DPDK implementation has an algorithm that tries to automatically optimize
> based on current traffic patterns.
>
> In DANOS refer to the “system default dataplane power-profile” config
> command tree for adaptive polling settings.  Interface RX/TX affinity is
> configured on a per interface basis under the “interfaces dataplane” config
> command tree.
>
> -robert
>
>
> > On Feb 22, 2021, at 11:46 AM, Jared Geiger  wrote:
> >
> > DANOS lets you specify how many dataplane cores you use versus control
> plane cores. So if you put a 16 core host in to handle 2GB of traffic, you
> can adjust the dataplane worker cores as needed. Control plane cores don't
> stay at 100% utilization.
> >
> > I use that technique plus DANOS runs on VMware (not oversubscribed)
> which allows me to use the hardware for other VMs. NICS are attached to the
> VM via PCI Passthrough which helps eliminate the overhead to the VMware
> hypervisor itself.
> >
> > I have an 8 core VM with 4 cores set to dataplane and 4 to control
> plane. The 4 control plane cores are typically idle only processing BGP
> route updates, SNMP, logs, etc.
> >
> > ~Jared
> >
> > On Sun, Feb 21, 2021 at 11:30 PM Etienne-Victor Depasquale <
> ed...@ieee.org> wrote:
> > Hello folks,
> >
> > I've just followed a thread regarding use of CGNAT and noted a
> suggestion (regarding DANOS) that includes use of DPDK.
> >
> > As I'm interested in the breadth of adoption of DPDK, and as I'm a
> researcher into energy and power efficiency, I'd love to hear your feedback
> on your use of power consumption control by DPDK.
> >
> > I've drawn up a bare-bones, 2-question survey at this link:
> >
> > https://www.surveymonkey.com/r/J886DPY.
> >
> > Responses have been set to anonymous.
> >
> > Cheers,
> >
> > Etienne
> >
> > --
> > Ing. Etienne-Victor Depasquale
> > Assistant Lecturer
> > Department of Communications & Computer Engineering
> > Faculty of Information & Communication Technology
> > University of Malta
> > Web. https://www.um.edu.mt/profile/etiennedepasquale
>
>

-- 
Ing. Etienne-Victor Depasquale
Assistant Lecturer
Department of Communications & Computer Engineering
Faculty of Information & Communication Technology
University of Malta
Web. https://www.um.edu.mt/profile/etiennedepasquale

Re: LOAs for Cross Connects - Something like PeeringDB for XC

>>> you can sign over something which ways "the person identified by the
>>> following public key is to be permitted to ..."
>>
>> you mean the fraudlent attacker who owned that INR seems to have signed
>> this request for a €1.000.000,49 wire transfer to their iban.  a person
>> is not identified by that signature.
> 
> If someone has a valid CA cert/key from the RIR, it's very hard to
> argue 'fraudulent'.
> It's, however, "easy" for the RIR to reverse the error, right? :)

sorry.  by 'fraudulent' i meant that they have no authority to request
the funds.  you just know they own some INR.  and if they request it
again, you might be confident it is at least the same attacker :)

now, you and i could agree formally, i.e. provably, out of band say
using pgp or whatever, that ownership of some INR identifies you.

or we could use some arbitrary other PKI entirely, e.g., X.400 was meant
for this.  but, as i said, karen, heather, and lucy know the personal
and organisational identity space far better than i.  i just know enough
about the rpki that it is very intentionally not in that identity space.

but think about the dance that prudent folk do to accept pgp keys, and
pgp has fingerprints to make it a bit easier to do oob verification.
but that verification uses an external identity provider, e.g. passport
or whatever makes you comfortable.  now infer what we would need to do
to accept an rpki INR key as a proof of identity.

randy

---
ra...@psg.com
`gpg --locate-external-keys --auto-key-locate wkd ra...@psg.com`
signatures are back, thanks to dmarc header mangling

Re: LOAs for Cross Connects - Something like PeeringDB for XC

On Mon, Feb 22, 2021 at 8:50 PM Randy Bush  wrote:
>
> > you can sign over something which ways "the person identified by the
> > following public key is to be permitted to ..."
>
> you mean the fraudlent attacker who owned that INR seems to have signed
> this request for a €1.000.000,49 wire transfer to their iban.  a person
> is not identified by that signature.

If someone has a valid CA cert/key from the RIR, it's very hard to
argue 'fraudulent'.
It's, however, "easy" for the RIR to reverse the error, right? :)

Re: LOAs for Cross Connects - Something like PeeringDB for XC

> you can sign over something which ways "the person identified by the
> following public key is to be permitted to ..."

you mean the fraudlent attacker who owned that INR seems to have signed
this request for a €1.000.000,49 wire transfer to their iban.  a person
is not identified by that signature.

randt

Re: STOP USING FONT SIZE SMALL Was: Re: LOAs for Cross Connects - Something like PeeringDB for XC

> Really, does anyone here think that it is good form to send email with
> font size *SMALL*?

rofl!

randy

---
ra...@psg.com
`gpg --locate-external-keys --auto-key-locate wkd ra...@psg.com`
signatures are back, thanks to dmarc header mangling

Re: Famous operational issues

2021-02-22 Thread sronan

Let me tell you about my personal favorite.

It’s 2002 and I am working as an engineer for an electronic stock trading 
platform (ECN), this platform happened to be the biggest platform for trading 
stocks electronically, on some days bigger than NASDAQ itself. This platform 
also happened to be run on DOS, FoxPro and a Novell file share, on a cluster of 
roughly 1,000 computers, two of which were the “engine” that matched all of the 
trades.

Well FoxPro has this “feature” where the ESC key halts the running program. We 
had the ability to remote control these DOS/FoxPro machines via some program we 
had written. Someone asked me to check the status of the process running on the 
primary matching engine, and when I was done, out of habit, I hit ESC. Trade 
processing grinds to a halt (phone calls have to be made to the SEC). I 
immediately called the NOC and told them it was me. Next thing I know, someone 
from the NOC is at my desk with a screwdriver putting the ESC key from my 
keyboard. I remained ESC keyless for the next several years until I left the 
company. I was hazed pretty good over it, but was essentially given a one time 
pass.

> On Feb 22, 2021, at 7:30 PM, b...@theworld.com wrote:
> 
> 
> At Boston Univ we discovered the hard way that a security guard's
> walkie-talkie could cause a $5,000 (or $10K for the big machine room)
> Halon dump.
> 
> Took a couple of times before we figured out the connection tho once
> someone made it to the hold button before it actually dumped.
> 
> Speaking of halon one very hot day I'm goofing off drinking coffee at
> a nearby sub shop when the owner tells me someone from the computing
> center was on the phone, that never happened before.
> 
> Some poor operator was holding the halon shot, it's a deadman's switch
> (well, button) and the building was doing its 110db thing could I come
> help? The building is being evac'd.
> 
> So my boss who wasn't the sharpest knife in the drawer follows me down
> as I enter and I'm sweating like a pig with a floor panel sucker
> trying to figure out which zone tripped.
> 
> And he shouts at me over the alarms: WHY TF DOES IT DO THIS?! Angrily.
> 
> I answered: well, maybe THERE'S A FIRE!!!
> 
> At which point I notice the back of my shoulder is really bothering
> me, which I say to him, and he says hmmm there's a big bee on your
> back maybe he's stinging you?
> 
> Fun day.
> 
> -- 
>-Barry Shein
> 
> Software Tool & Die| b...@theworld.com | 
> http://www.TheWorld.com
> Purveyors to the Trade | Voice: +1 617-STD-WRLD   | 800-THE-WRLD
> The World: Since 1989  | A Public Information Utility | *oo*

STOP USING FONT SIZE SMALL Was: Re: LOAs for Cross Connects - Something like PeeringDB for XC

2021-02-22 Thread Mark Andrews

Really, does anyone here think that it is good form to send email with font 
size *SMALL*?
If your MUA does this by default complain to the developers.  The default 
should be “medium”.
If the font is too big on your screen change the magnification *you* choose to 
display to *yourself*,
don’t change the font size you send to everyone else.

Mark

Well... I must confess that I had some diffi=
culty=C2=A0on the first understanding=C2=A0of what is proposed.But =


> On 23 Feb 2021, at 04:03, Douglas Fischer  wrote:
> 
> Well... I must confess that I had some difficulty on the first understanding 
> of what is proposed.
> 
> But after the 4 reads, I saw that this "spaghetti" thing is more powerful 
> than I could imagine!
> 
> 
> Please correct me if I'm no right:
> But it looks like a "crypto sign and publishes" anything related to an 
> organization.
> 
> Yes, I think that with some effort CrossConnect LOAs can be fitted inside of 
> it...
> I'm not sure if it is the better solution for the scope of LOAs, but 
> certainly is a valid discussion.
> 
> 
> What is bubbling in my mind is the standard data model for each type of 
> different attribute that can exist...
> Who will define that?
> 
> 
> 
> Em seg., 22 de fev. de 2021 às 12:26, Christopher Morrow 
>  escreveu:
> On Mon, Feb 22, 2021 at 9:19 AM Douglas Fischer
>  wrote:
> >
> > I believe that almost everyone in here knows that LOAs for Cross Connects 
> > in Datacenters and Telecom Rooms can be a pain...
> >
> > I don't know if I'm suggesting something that already exists.
> > Or even if I'm suggesting something that could be unpopular for some reason.
> >
> > But every time I need to deal with some Cross-Connect LOA, and mostly when 
> > we face some rework on data mistakes, I dream with a "PeeringDB for Cross 
> > Connects".
> >
> 
> are you asking about something like this:
>   https://datatracker.ietf.org/doc/draft-spaghetti-sidrops-rpki-rsc/
> 
> Which COULD be used to, as an AS holder:
>   "sign something to be sent between you and the colo and your intended peer"
> 
> that you could sign (with your rpki stuffs) and your peer could also
> sign with their 'rpki stuffs', and which the colo provider could
> automatically validate and action upon final signature(s) received.
> 
> > So, this mail is a question and also a suggestion.
> >
> >
> > There is something like an "online notary's office" exclusive for 
> > Cross-Connect LOAs?
> >  - Somewhere Organizations can register information authorizing connections 
> > of Port on their Places (Cages, Racks, etc)...
> >
> 
> The RPKI data today doesn't contain information about
> cages/ports/patch-panels, so possibly the spaghetti draft isn't a
> terrific fit?
> 
> > If it doesn't exist. What would be necessary for that?
> > Mostly considering the PeeringDB work model.
> >  - OpenSource.
> >  - Free access to the tool, and sponsors to keep the project alive.
> >  - API driven, with some Web-gui.
> > And considering some data-modeling.
> >  - Most of the data being Open/Public (Organizations, 
> > Facilities(Datacenters and/or Telecom-Rooms), Presence on Facilities, etc).
> >  - Access control to Information that can not be public (A-side 
> > organization, Z-Side Organization, PathPanel/Port).
> > And some workflow
> >  - Cross Connect Requiremento/Authorization from A-Side
> >  - Acceptance/Authorization from Z-side.
> >  - Acceptance/Authorization from Facilities involved (could be more than 
> > one)
> >  - Execution/Activation notice from Facilities.
> >
> >
> > --
> > Douglas Fernando Fischer
> > Engº de Controle e Automação
> 
> 
> -- 
> Douglas Fernando Fischer
> Engº de Controle e Automação

-- 
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742  INTERNET: ma...@isc.org

Re: Famous operational issues

2021-02-22 Thread bzs



At Boston Univ we discovered the hard way that a security guard's
walkie-talkie could cause a $5,000 (or $10K for the big machine room)
Halon dump.

Took a couple of times before we figured out the connection tho once
someone made it to the hold button before it actually dumped.

Speaking of halon one very hot day I'm goofing off drinking coffee at
a nearby sub shop when the owner tells me someone from the computing
center was on the phone, that never happened before.

Some poor operator was holding the halon shot, it's a deadman's switch
(well, button) and the building was doing its 110db thing could I come
help? The building is being evac'd.

So my boss who wasn't the sharpest knife in the drawer follows me down
as I enter and I'm sweating like a pig with a floor panel sucker
trying to figure out which zone tripped.

And he shouts at me over the alarms: WHY TF DOES IT DO THIS?! Angrily.

I answered: well, maybe THERE'S A FIRE!!!

At which point I notice the back of my shoulder is really bothering
me, which I say to him, and he says hmmm there's a big bee on your
back maybe he's stinging you?

Fun day.

-- 
-Barry Shein

Software Tool & Die| b...@theworld.com | http://www.TheWorld.com
Purveyors to the Trade | Voice: +1 617-STD-WRLD   | 800-THE-WRLD
The World: Since 1989  | A Public Information Utility | *oo*

Re: LOAs for Cross Connects - Something like PeeringDB for XC

2021-02-22 Thread George Michaelson

The LOA type model is one of the ones we showed on slideware when we
presented RTA in IETF, and at the CloudFlare RPKI workshop years ago.

The detached signature model inherent in RTA and RSC goes to "you
define the business logic" It's not proscriptive. I saw nothing
proposed here which I thought wasn't a rational thing to try and
certify in this manner. The key point is, the "action" you want to
approve has to vest in "who controls the stated internet number
resources" -If they have no bearing, then its not rational to propose
using (R)PKI to do it. some other PKI? sure.

Randy is correct that the processes are baroque, not well defined,
come with all kinds of corner cases: what does a more specific command
(regarding some IP address) if its not signed and the parent is? or,
if the more specific is, and the parent isn't?)

Randy is also correct that RPKI certificates by design, do not permit
their use in ways which go directly to things like identity proofs.
detached signatures open the door to doing some things here, because
you can sign over something which ways "the person identified by the
following public key is to be permitted to ..." And in like sense, we
removed the uses which go to message encryption, sender or receiver.
You can't directly use an RPKI certificate to do "for your eyes only"
-it can only say "the person controlling these numbers, says the
following"

Obviously, I think this detached signature model is good :-)

On Tue, Feb 23, 2021 at 6:31 AM Randy Bush  wrote:
>
> >> What if PeeringDB would be the CA for the Facilities?
> >> Supposedly this solves the CA problem of the "Colo Folks".
> >
> > I think pushing your security identification out (as the notional
> > equinix) to a third party where you can't revoke/change/etc is asking
> > for dangerous things to happen.
>
> there are a few examples of industry associations with simple, strong,
> and formal ties sufficient to allow forms of trust automation.  folk
> such as karen o'donoghue, lucy lynch, and heather flanagan would be able
> to speak vastly more knowledgeably in this space than i.
>
> > again, that draft is a... draft still and I"m sure we'll have a bunch
> > of chatter/discussion/changes before done, but it smells like it might
> > help.
>
> you might notice that we use it in draft-ietf-opsawg-finding-geofeeds.
> but that application is specifically to use rpki data to attest to ip
> address ownership.  the problem there is that the draft is a cool proof
> of concept, but is not operationally easy to use.
>
> randy
>
> ---
> ra...@psg.com
> `gpg --locate-external-keys --auto-key-locate wkd ra...@psg.com`
> signatures are back, thanks to dmarc header mangling

Re: DPDK and energy efficiency

2021-02-22 Thread Jared Geiger

"set system default dataplane cpu-affinity 3-7" is what I have set for my
use case. Technically its 5 cores out of 8 total, but 4 are polling cores
and 1 manages those 4. Then the control plane is 3 plus the leftover cycles
of the 1 manager core.

On Mon, Feb 22, 2021 at 2:04 PM Etienne Depasquale  wrote:

> Thanks Jared; that's very interesting.
>
>
>
> Earlier today, I had a private exchange of emails regarding the
> progressive development of architectures specific to the domain of
> high-speed networking functions. Your note reinforces the notion that this
> “hard” partitioning of cores is a key part of the DSA (domain-specific
> architecture) here.
>
>
>
> Sent from my Windows 10 device
>
>
>
> *From: *Jared Geiger 
> *Sent: *Monday, 22 February 2021 20:53
> *To: *NANOG 
> *Subject: *Re: DPDK and energy efficiency
>
>
>
> DANOS lets you specify how many dataplane cores you use versus control
> plane cores. So if you put a 16 core host in to handle 2GB of traffic, you
> can adjust the dataplane worker cores as needed. Control plane cores don't
> stay at 100% utilization.
>
>
>
> I use that technique plus DANOS runs on VMware (not oversubscribed) which
> allows me to use the hardware for other VMs. NICS are attached to the VM
> via PCI Passthrough which helps eliminate the overhead to the VMware
> hypervisor itself.
>
>
>
> I have an 8 core VM with 4 cores set to dataplane and 4 to control plane.
> The 4 control plane cores are typically idle only processing BGP route
> updates, SNMP, logs, etc.
>
>
>
> ~Jared
>
>
>
> On Sun, Feb 21, 2021 at 11:30 PM Etienne-Victor Depasquale 
> wrote:
>
> Hello folks,
>
>
>
> I've just followed a thread regarding use of CGNAT and noted a suggestion
> (regarding DANOS) that includes use of DPDK.
>
>
>
> As I'm interested in the breadth of adoption of DPDK, and as I'm a
> researcher into energy and power efficiency, I'd love to hear your feedback
> on your use of power consumption control by DPDK.
>
>
>
> I've drawn up a bare-bones, 2-question survey at this link:
>
>
>
> https://www.surveymonkey.com/r/J886DPY.
>
>
>
> Responses have been set to anonymous.
>
>
>
> Cheers,
>
>
>
> Etienne
>
>
>
> --
>
> Ing. Etienne-Victor Depasquale
> Assistant Lecturer
> Department of Communications & Computer Engineering
> Faculty of Information & Communication Technology
> University of Malta
>
> Web. https://www.um.edu.mt/profile/etiennedepasquale
>
>
>

Re: Famous operational issues

2021-02-22 Thread Patrick W. Gilmore

On Feb 22, 2021, at 7:02 AM, t...@pelican.org wrote:
> On Thursday, 18 February, 2021 22:37, "Warren Kumari"  
> said:
> 
>> 4: Not too long after I started doing networking (and for the same small
>> ISP in Yonkers), I'm flying off to install a new customer. I (of course)
>> think that I'm hot stuff because I'm going to do the install, configure the
>> router, whee, look at me! Anyway, I don't want to check a bag, and so I
>> stuff the Cisco 2501 in a carryon bag, along with tools, etc (this was all
>> pre-9/11!). I'm going through security and the TSA[0] person opens my bag
>> and pulls the router out. "What's this?!" he asks. I politely tell him that
>> it's a router. He says it's not. I'm still thinking that I'm the new
>> hotness, and so I tell him in a somewhat condescending way that it is, and
>> I know what I'm talking about. He tells me that it's not a router, and is
>> starting to get annoyed. I explain using my "talking to a 5 year old" voice
>> that it most certainly is a router. He tells me that lying to airport
>> security is a federal offense, and starts looming at me. I adjust my
>> attitude and start explaining that it's like a computer and makes the
>> Internet work. He gruffly hands me back the router, I put it in my bag and
>> scurry away. As I do so, I hear him telling his colleague that it wasn't a
>> router, and that he certainly knows what a router is, because he does
>> woodwork...
> 
> Here in the UK we avoid that issue by pronouncing the packet-shifter as 
> "rooter", and only the wood-working tool as "rowter" :)

So wrong.

A “root” server is part of the DNS. A “route” server is part of BGP.


> Of course, it raises a different set of problems when talking to the 
> Australians…

Everything is weird down down. But I still like them. :-)

-- 
TTFN,
patrick

RE: DPDK and energy efficiency

2021-02-22 Thread Etienne Depasquale

Thanks Jared; that's very interesting.

Earlier today, I had a private exchange of emails regarding the progressive 
development of architectures specific to the domain of high-speed networking 
functions. Your note reinforces the notion that this “hard” partitioning of 
cores is a key part of the DSA (domain-specific architecture) here.

Sent from my Windows 10 device

From: Jared Geiger
Sent: Monday, 22 February 2021 20:53
To: NANOG
Subject: Re: DPDK and energy efficiency

DANOS lets you specify how many dataplane cores you use versus control plane 
cores. So if you put a 16 core host in to handle 2GB of traffic, you can adjust 
the dataplane worker cores as needed. Control plane cores don't stay at 100% 
utilization. 

I use that technique plus DANOS runs on VMware (not oversubscribed) which 
allows me to use the hardware for other VMs. NICS are attached to the VM via 
PCI Passthrough which helps eliminate the overhead to the VMware hypervisor 
itself.

I have an 8 core VM with 4 cores set to dataplane and 4 to control plane. The 4 
control plane cores are typically idle only processing BGP route updates, SNMP, 
logs, etc.

~Jared

On Sun, Feb 21, 2021 at 11:30 PM Etienne-Victor Depasquale  
wrote:
Hello folks,

I've just followed a thread regarding use of CGNAT and noted a suggestion 
(regarding DANOS) that includes use of DPDK.

As I'm interested in the breadth of adoption of DPDK, and as I'm a researcher 
into energy and power efficiency, I'd love to hear your feedback on your use of 
power consumption control by DPDK.

I've drawn up a bare-bones, 2-question survey at this link: 

https://www.surveymonkey.com/r/J886DPY. 

Responses have been set to anonymous.

Cheers,

Etienne

-- 
Ing. Etienne-Victor Depasquale
Assistant Lecturer
Department of Communications & Computer Engineering
Faculty of Information & Communication Technology
University of Malta
Web. https://www.um.edu.mt/profile/etiennedepasquale

Re: Famous operational issues

2021-02-22 Thread Owen DeLong



> On Feb 18, 2021, at 9:04 PM, Jen Linkova  wrote:
> 
> On Fri, Feb 19, 2021 at 9:40 AM Warren Kumari  wrote:
>> 4: Not too long after I started doing networking (and for the same small ISP 
>> in Yonkers), I'm flying off to install a new customer. I (of course) think 
>> that I'm hot stuff because I'm going to do the install, configure the 
>> router, whee, look at me! Anyway, I don't want to check a bag, and so I 
>> stuff the Cisco 2501 in a carryon bag, along with tools, etc (this was all 
>> pre-9/11!). I'm going through security and the TSA[0] person opens my bag 
>> and pulls the router out. "What's this?!" he asks. I politely tell him that 
>> it's a router. He says it's not. I'm still thinking that I'm the new 
>> hotness, and so I tell him in a somewhat condescending way that it is, and I 
>> know what I'm talking about. He tells me that it's not a router, and is 
>> starting to get annoyed. I explain using my "talking to a 5 year old" voice 
>> that it most certainly is a router. He tells me that lying to airport 
>> security is a federal offense, and starts looming at me. I adjust my 
>> attitude and start explaining that it's like a computer and makes the 
>> Internet work. He gruffly hands me back the router, I put it in my bag and 
>> scurry away. As I do so, I hear him telling his colleague that it wasn't a 
>> router, and that he certainly knows what a router is, because he does 
>> woodwork...
> 
> OK, Warren, achievement unlocked. You've just made a network engineer
> to google 'router'
> 
> P.S. I guess I'm obliged to tell a story if I respond to this thread...so...
> "Servers and the ice cream factory".
> Late spring/early summer in Moscow. The temperature above 30C (86°F).
> I worked for a local content provided.
> Aircons in our server room died, the technician ETA was 2 days ( I
> guess we were not the only ones with aircon problems).
> So we drove to the nearby ice cream factory  and got *a lot* of  dry
> ice. Then we have a roaster: every few hours one person took a deep
> breath, grabbed a box of dry ice, ran into the server room and emptied
> the box on top of the racks. The backup person was watching through
> the glass door - just in case, you know, ready to start the rescue
> operation.
> We (and the servers) survived till the technician arrived. And we had
> a lot of dry ice to cool the beer..
> 
> -- 
> SY, Jen Linkova aka Furry

During a wood-working project for the Southern California Linux Expo (the tech 
team that
(among other things) runs the network for the show was building new equipment 
carts), I
came up with the following meme:



[I don’t know if NANOG will pass the image despite its small size, so textual 
description:
A bandaged hand with the index finger amputated at the second knuckle with 
overlaid red
text stating “Carless Routing May Lead to Urgent Test of Self Healing Network”]

Fortunately, we didn’t have any such issues with the router, though we did have 
one person
suffer a crushed toe from a cabinet tip-over. Fortunately, the person made a 
full recovery.

Owen

Re: DPDK and energy efficiency

2021-02-22 Thread Robert Bays

Beyond RX/TX CPU affinity, in DANOS you can further tune power consumption by 
changing the adaptive polling rate.  It doesn’t, per the survey, "keep 
utilization at 100% regardless of packet activity.”  Adaptive polling changes 
in DPDK optimize for tradeoffs between power consumption, latency/jitter and 
drops during throughput ramp up periods.  Ideally your DPDK implementation has 
an algorithm that tries to automatically optimize based on current traffic 
patterns.

In DANOS refer to the “system default dataplane power-profile” config command 
tree for adaptive polling settings.  Interface RX/TX affinity is configured on 
a per interface basis under the “interfaces dataplane” config command tree.

-robert

> On Feb 22, 2021, at 11:46 AM, Jared Geiger  wrote:
> 
> DANOS lets you specify how many dataplane cores you use versus control plane 
> cores. So if you put a 16 core host in to handle 2GB of traffic, you can 
> adjust the dataplane worker cores as needed. Control plane cores don't stay 
> at 100% utilization. 
> 
> I use that technique plus DANOS runs on VMware (not oversubscribed) which 
> allows me to use the hardware for other VMs. NICS are attached to the VM via 
> PCI Passthrough which helps eliminate the overhead to the VMware hypervisor 
> itself.
> 
> I have an 8 core VM with 4 cores set to dataplane and 4 to control plane. The 
> 4 control plane cores are typically idle only processing BGP route updates, 
> SNMP, logs, etc.
> 
> ~Jared
> 
> On Sun, Feb 21, 2021 at 11:30 PM Etienne-Victor Depasquale  
> wrote:
> Hello folks,
> 
> I've just followed a thread regarding use of CGNAT and noted a suggestion 
> (regarding DANOS) that includes use of DPDK.
> 
> As I'm interested in the breadth of adoption of DPDK, and as I'm a researcher 
> into energy and power efficiency, I'd love to hear your feedback on your use 
> of power consumption control by DPDK.
> 
> I've drawn up a bare-bones, 2-question survey at this link: 
> 
> https://www.surveymonkey.com/r/J886DPY. 
> 
> Responses have been set to anonymous.
> 
> Cheers,
> 
> Etienne
> 
> -- 
> Ing. Etienne-Victor Depasquale
> Assistant Lecturer
> Department of Communications & Computer Engineering
> Faculty of Information & Communication Technology
> University of Malta
> Web. https://www.um.edu.mt/profile/etiennedepasquale

RE: Famous operational issues

2021-02-22 Thread Tony Wicks

Many years ago I experienced a very similar thing. The DC/Integrator I worked 
for outsourced the co-location and operation of mainframe services for several 
banks and government organisations. One of these banks had a significant 
investment in AS/400's and they decided that it was so much hassle and expense 
using our datacentres that they would start putting those nice small AS/400's 
in computer rooms in their office buildings instead. One particular computer 
room contained large line printers that the developers would use to print out 
whatever it is such people print out. One Saturday morning I received a frantic 
call from the customer to say that all their primary production as/400's had 
gone offline. After a short investigation I realised that all the offline 
devices wire in this particular computer room. It turn's out that one of the 
developers had bought his six year old son to work that Saturday and upon 
retrieval of a printout said son had dutifully followed dad in to the computer 
room and was unable to resist the big red button sitting exposed on the wall by 
the door. Shortly thereafter the embarrassed customer decided that perhaps it 
was worth relocating their as/400's to our expensive datacentres.



> 
>  During my younger days, that button was used a few time by the 
> operator of a VM/370 to regain control from someone with a "curious 
> mind" *cought* *cought*...
> 
Two horror stories I remember from long ago when I was a console jockey for a 
federal space agency that will remain nameless :P

1. A coworker brought her daughter to work with her on a Saturday overtime 
shift because she couldn't get a babysitter. She parked the kid with a coloring 
book and a pile of crayons at the only table in the console room with some 
space, right next to the master console for our 3081. I asked her to make sure 
sh was well away from the console, and as she reached over to scoot the girl 
and her coloring books further away she slipped, and reached out to steady 
herself. Yep, planted her finger right down on the IML button (plexi covers? We 
don' need no STEENKIN' 
plexi covers!). MVS and VM vanished, two dozen tape drives rewound and several 
hours' worth of data merge jobs went blooey.

Re: LOAs for Cross Connects - Something like PeeringDB for XC

>> What if PeeringDB would be the CA for the Facilities?
>> Supposedly this solves the CA problem of the "Colo Folks".
> 
> I think pushing your security identification out (as the notional
> equinix) to a third party where you can't revoke/change/etc is asking
> for dangerous things to happen.

there are a few examples of industry associations with simple, strong,
and formal ties sufficient to allow forms of trust automation.  folk
such as karen o'donoghue, lucy lynch, and heather flanagan would be able
to speak vastly more knowledgeably in this space than i.

> again, that draft is a... draft still and I"m sure we'll have a bunch
> of chatter/discussion/changes before done, but it smells like it might
> help.

you might notice that we use it in draft-ietf-opsawg-finding-geofeeds.
but that application is specifically to use rpki data to attest to ip
address ownership.  the problem there is that the draft is a cool proof
of concept, but is not operationally easy to use.

randy

---
ra...@psg.com
`gpg --locate-external-keys --auto-key-locate wkd ra...@psg.com`
signatures are back, thanks to dmarc header mangling

Re: Texas internet connectivity declining due to blackouts

2021-02-22 Thread Ben Cannon

You guys build how you want. At 6x7 we are building to prepare for possible 
climactic shifts.  The origin need not be anthropogenic, but that doesn’t look 
good.

“Doing nothing” isn’t really an option, and “doing what republicans want 
because they say so and they’re my dad” isn’t a good argument.


Nor is it a personality.

Ms. Lady Benjamin PD Cannon, ASCE
6x7 Networks & 6x7 Telecom, LLC 
CEO 
b...@6by7.net
"The only fully end-to-end encrypted global telecommunications company in the 
world.”

FCC License KJ6FJJ

Sent from my iPhone via RFC1149.

> On Feb 22, 2021, at 11:35 AM, Mel Beckman  wrote:
> 
> Rich,
> 
> Calling my opposing argument “trash”, and then falsely linking it to 
> unrelated theories on vaccines, evolution, moon landings, and dietary 
> supplements, is intellectually dishonest and professionally rude. Why don’t 
> you respond to the facts raised in the article? Does your religion not permit 
> that?
> 
> Either weather events are getting worse, or they aren’t. I provided solid 
> evidence that they are diminishing. The truth of this issue is important to 
> NANOG, because we build the infrastructure that runs the Internet, and we 
> can’t afford to waste finite resources on alarmist claims. 
> 
> -mel
> 
>> On Feb 22, 2021, at 10:23 AM, Rich Kulawiec  wrote:
>> 
>>> On Mon, Feb 22, 2021 at 05:48:06PM +, Mel Beckman wrote:
>>> Sorry Global Warmists,
>> 
>> Right.  Sure.  Also, the earth is 6,000 years old (and flat), the moon
>> landings were faked, creationism is real, dinosaurs and humans co-existed,
>> vaccines cause autism, Elvis is alive, and...how does that line go?  Oh,
>> right: artificial sweeteners are safe, WMDs were in Iraq, and Anna Nicole
>> married for love. [shout-out to Levon Helm]
>> 
>> This trash doesn't deserve rebuttal: it deserves ridicule.
>> 
>> ---rsk
>

Re: LOAs for Cross Connects - Something like PeeringDB for XC

On Mon, Feb 22, 2021 at 2:44 PM Douglas Fischer
 wrote:
>
> What if PeeringDB would be the CA for the Facilities?
> Supposedly this solves the CA problem of the "Colo Folks".
>

I think pushing your security identification out (as the notional
equinix) to a third party where you can't revoke/change/etc
is asking for dangerous things to happen.

The 'strength' of the RPKI (vs the web-pki) is that there are a
defined number of ways into the system.
You have ip space (IP Number Assets)? you get CA-cert and can create ROA.

there are surely a host of corner cases with 'use the rpki to sign not
INR things!!!', but at least:
  "Are you sure that's the right foo.bar? not f00.bar? or fOO.bar?"
  "yes, they have a CA cert signed by the RIR, with INRs they can
toggle ROA for.. if that CA cert signed the checklist then 'ok'"

again, that draft is a... draft still and I"m sure we'll have a bunch
of chatter/discussion/changes before done, but it smells like it might
help.

> Would PeeringDB be interested in that?
>
>
> Em seg., 22 de fev. de 2021 às 16:04, Christopher Morrow 
>  escreveu:
>>
>> On Mon, Feb 22, 2021 at 1:39 PM Randy Bush  wrote:
>> >
>> > > are you asking about something like this:
>> > >   https://datatracker.ietf.org/doc/draft-spaghetti-sidrops-rpki-rsc/
>> > >
>> > > Which COULD be used to, as an AS holder:
>> > >   "sign something to be sent between you and the colo and your intended 
>> > > peer"
>> > >
>> > > that you could sign (with your rpki stuffs) and your peer could also
>> > > sign with their 'rpki stuffs', and which the colo provider could
>> > > automatically validate and action upon final signature(s) received.
>> >
>> > chris,
>> >
>> > way back, the rirs were very insistant that their use of rpki authority
>> > was most emphatically not to be considered an identity service.  this
>> > permeated the design; e.g., organization names were specifically
>> > forbidden in certificate CN, Subject Alternative Name, etc.
>> >
>>
>> yup, I agree... though the b2b stuff George/Geoff have written up LOOKS like
>> it could be useful for this LOA type discussion. The spaghetti draft appears
>> to also fill this niche...
>>
>> Neither are particularly rooted in the RPKI except that the CA certs are 
>> being
>> used as a method to attest that a 'thing' exists, and that something signed
>> that 'thing' as proof of knowledge (I guess, really). Effectively this is:
>>   1) I am 'ca-foo' in a tree that you can trust knows I am 'foo'.
>>   2) I signed this blob (LOA)
>>   3) I asked jane at bar.com to sign as well
>>   4) you can verify me (because rpki tree) and you can verify Jane because 
>> she's
>>   also using her RPKI ca cert.
>>
>> this may be a little cumbersome to sort through, especially if all parties 
>> here
>> aren't party to the RPKI (did equinix plumb the RPKI into their customer 
>> portal
>> and all of the things required to make a x-connect work in this manner?), 
>> but I
>> imagine that if this gets wings it could be automated and it could be 
>> reliable
>> and all parties (except the colo folks perhaps?) may already have incentives
>> in places to use their RPKI goop for this function.
>>
>> -chris
>>
>> > aside: of course a few rirs thought that *their* names should be in
>> > their certs as exeptions.  i remember the laughter.
>> >
>> > randy
>> >
>> > ---
>> > ra...@psg.com
>> > `gpg --locate-external-keys --auto-key-locate wkd ra...@psg.com`
>> > signatures are back, thanks to dmarc header mangling
>
>
>
> --
> Douglas Fernando Fischer
> Engº de Controle e Automação

Re: Texas internet connectivity declining due to blackouts

Brandon,

Actually, no, I don’t have to do science to object to claims made by 
scientists. Even when there is a consensus. I can simply cite data, and it is 
the duty of the person making the claim to defend their theory.

If you’re going to defend it for them, then you need to cite countering data, 
not an “argument from authority”. It should be a simple matter to find data 
supporting the claim that weather is getting more severe, rather than just more 
costly, which is the usual conflation by climate warmists.

 -mel

On Feb 22, 2021, at 11:38 AM, Brandon Svec 
mailto:bs...@teamonesolutions.com>> wrote:

OK, I looked closer.  I see it is a self titled opinion piece so there is that. 
 Next, I see all the links in the article go to questionable sites (not .edu or 
scientific organizations, etc.)  except one cherry picked NOAA stat for a 
single event type for a single year.  Last, the writer is the president of a 
right wing anti science lobbying group called "Spark of Freedom" funded by 
Exxon Mobil.

Look, I and most everyone on this list are not qualified, experienced climate 
scientists.  However, I think when you are not an expert you should respect and 
believe what experts say as a group.  Picking outliers and sharing opinions of 
obviously unqualified and biased people is reprehensible and dishonest as far 
as I am concerned.

If you truly believe the scientific consensus around climate change is wrong 
you are going to have to do a lot more than share links.  You will have to do 
science and prove it.

Best.

On Mon, Feb 22, 2021 at 11:27 AM Mel Beckman 
mailto:m...@beckman.org>> wrote:
What offended you? The term “Global Warmist”? It’s an accurate description of 
people who hold that climate change is causing more frequent and severe 
weather, due to heating of the atmosphere.

And your argument about “Forbes for something related to science” fails on the 
classic logical fallacy “appeal to authority”. Just because Forbes states 
easily verifiable public facts doesn’t make them untrustworthy. Scientific 
knowledge is best established by evidence and experiment rather than argued 
through authority by “consensus”. Science is not a consensus enterprise.

 -mel

> On Feb 22, 2021, at 10:16 AM, Brandon Svec via NANOG 
> mailto:nanog@nanog.org>> wrote:
>
>
>
>> On Feb 22, 2021, at 9:56 AM, Mel Beckman 
>> mailto:m...@beckman.org>> wrote:
>>
>> Sorry Global Warmists,
>
>
> Stopped taking you seriously or reading further right there.  Well, that and 
> linking to Forbes for something related to science.
>
> Best.

Re: Texas internet connectivity declining due to blackouts

2021-02-22 Thread Mike Hammett

I can argue about this all day on Facebook or Twitter (and sometimes do, 
whether trolling or serious depends on the day). Let's reign it back in to 
network operations concerns. 

- 
Mike Hammett 
Intelligent Computing Solutions 
http://www.ics-il.com 

Midwest-IX 
http://www.midwest-ix.com 

- Original Message -

From: "Brandon Svec via NANOG"  
To: "Mel Beckman"  
Cc: "Rich Kulawiec" , nanog@nanog.org 
Sent: Monday, February 22, 2021 1:43:22 PM 
Subject: Re: Texas internet connectivity declining due to blackouts 

On Mon, Feb 22, 2021 at 11:37 AM Mel Beckman < m...@beckman.org > wrote: 

Either weather events are getting worse, or they aren’t. 

No, nothing is so black and white. Certainly not science. 

I provided solid evidence that they are diminishing. 

No, you didn't. You shared an opinion piece written by the president of a 
science denying lobbying group funded by Exxon Mobil 

The truth of this issue is important to NANOG, because we build the 
infrastructure that runs the Internet, and we can’t afford to waste finite 
resources on alarmist claims. 

That I can partially agree with. I would say even if science is 100% wrong 
about climate change and what is causing it, it is still a good investment to 
prepare for the unexpected and unprecedented when it comes to building and 
supporting resilient systems. 

Best 

> On Feb 22, 2021, at 10:23 AM, Rich Kulawiec < r...@gsp.org > wrote: 
> 
> On Mon, Feb 22, 2021 at 05:48:06PM +, Mel Beckman wrote: 
>> Sorry Global Warmists, 
> 
> Right. Sure. Also, the earth is 6,000 years old (and flat), the moon 
> landings were faked, creationism is real, dinosaurs and humans co-existed, 
> vaccines cause autism, Elvis is alive, and...how does that line go? Oh, 
> right: artificial sweeteners are safe, WMDs were in Iraq, and Anna Nicole 
> married for love. [shout-out to Levon Helm] 
> 
> This trash doesn't deserve rebuttal: it deserves ridicule. 
> 
> ---rsk

Re: DPDK and energy efficiency

2021-02-22 Thread Jared Geiger

DANOS lets you specify how many dataplane cores you use versus control
plane cores. So if you put a 16 core host in to handle 2GB of traffic, you
can adjust the dataplane worker cores as needed. Control plane cores don't
stay at 100% utilization.

I use that technique plus DANOS runs on VMware (not oversubscribed) which
allows me to use the hardware for other VMs. NICS are attached to the VM
via PCI Passthrough which helps eliminate the overhead to the VMware
hypervisor itself.

I have an 8 core VM with 4 cores set to dataplane and 4 to control plane.
The 4 control plane cores are typically idle only processing BGP route
updates, SNMP, logs, etc.

~Jared

On Sun, Feb 21, 2021 at 11:30 PM Etienne-Victor Depasquale 
wrote:

> Hello folks,
>
> I've just followed a thread regarding use of CGNAT and noted a suggestion
> (regarding DANOS) that includes use of DPDK.
>
> As I'm interested in the breadth of adoption of DPDK, and as I'm a
> researcher into energy and power efficiency, I'd love to hear your feedback
> on your use of power consumption control by DPDK.
>
> I've drawn up a bare-bones, 2-question survey at this link:
>
> https://www.surveymonkey.com/r/J886DPY.
>
> Responses have been set to anonymous.
>
> Cheers,
>
> Etienne
>
> --
> Ing. Etienne-Victor Depasquale
> Assistant Lecturer
> Department of Communications & Computer Engineering
> Faculty of Information & Communication Technology
> University of Malta
> Web. https://www.um.edu.mt/profile/etiennedepasquale
>

Re: LOAs for Cross Connects - Something like PeeringDB for XC

What if PeeringDB would be the CA for the Facilities?
Supposedly this solves the CA problem of the "Colo Folks".

Would PeeringDB be interested in that?


Em seg., 22 de fev. de 2021 às 16:04, Christopher Morrow <
morrowc.li...@gmail.com> escreveu:

> On Mon, Feb 22, 2021 at 1:39 PM Randy Bush  wrote:
> >
> > > are you asking about something like this:
> > >   https://datatracker.ietf.org/doc/draft-spaghetti-sidrops-rpki-rsc/
> > >
> > > Which COULD be used to, as an AS holder:
> > >   "sign something to be sent between you and the colo and your
> intended peer"
> > >
> > > that you could sign (with your rpki stuffs) and your peer could also
> > > sign with their 'rpki stuffs', and which the colo provider could
> > > automatically validate and action upon final signature(s) received.
> >
> > chris,
> >
> > way back, the rirs were very insistant that their use of rpki authority
> > was most emphatically not to be considered an identity service.  this
> > permeated the design; e.g., organization names were specifically
> > forbidden in certificate CN, Subject Alternative Name, etc.
> >
>
> yup, I agree... though the b2b stuff George/Geoff have written up LOOKS
> like
> it could be useful for this LOA type discussion. The spaghetti draft
> appears
> to also fill this niche...
>
> Neither are particularly rooted in the RPKI except that the CA certs are
> being
> used as a method to attest that a 'thing' exists, and that something signed
> that 'thing' as proof of knowledge (I guess, really). Effectively this is:
>   1) I am 'ca-foo' in a tree that you can trust knows I am 'foo'.
>   2) I signed this blob (LOA)
>   3) I asked jane at bar.com to sign as well
>   4) you can verify me (because rpki tree) and you can verify Jane because
> she's
>   also using her RPKI ca cert.
>
> this may be a little cumbersome to sort through, especially if all parties
> here
> aren't party to the RPKI (did equinix plumb the RPKI into their customer
> portal
> and all of the things required to make a x-connect work in this manner?),
> but I
> imagine that if this gets wings it could be automated and it could be
> reliable
> and all parties (except the colo folks perhaps?) may already have
> incentives
> in places to use their RPKI goop for this function.
>
> -chris
>
> > aside: of course a few rirs thought that *their* names should be in
> > their certs as exeptions.  i remember the laughter.
> >
> > randy
> >
> > ---
> > ra...@psg.com
> > `gpg --locate-external-keys --auto-key-locate wkd ra...@psg.com`
> > signatures are back, thanks to dmarc header mangling
>


-- 
Douglas Fernando Fischer
Engº de Controle e Automação

Re: Texas internet connectivity declining due to blackouts

2021-02-22 Thread Brandon Svec via NANOG

On Mon, Feb 22, 2021 at 11:37 AM Mel Beckman  wrote:

>
>
> Either weather events are getting worse, or they aren’t.

No, nothing is so black and white.  Certainly not science.

> I provided solid evidence that they are diminishing.

No, you didn't.  You shared an opinion piece written by the president of a
science denying lobbying group funded by Exxon Mobil

> The truth of this issue is important to NANOG, because we build the
> infrastructure that runs the Internet, and we can’t afford to waste finite
> resources on alarmist claims.
>

That I can partially agree with.  I would say even if science is 100% wrong
about climate change and what is causing it, it is still a good
investment to prepare for the unexpected and unprecedented when it comes to
building and supporting resilient systems.

Best

>
>
> > On Feb 22, 2021, at 10:23 AM, Rich Kulawiec  wrote:
> >
> > On Mon, Feb 22, 2021 at 05:48:06PM +, Mel Beckman wrote:
> >> Sorry Global Warmists,
> >
> > Right.  Sure.  Also, the earth is 6,000 years old (and flat), the moon
> > landings were faked, creationism is real, dinosaurs and humans
> co-existed,
> > vaccines cause autism, Elvis is alive, and...how does that line go?  Oh,
> > right: artificial sweeteners are safe, WMDs were in Iraq, and Anna Nicole
> > married for love. [shout-out to Levon Helm]
> >
> > This trash doesn't deserve rebuttal: it deserves ridicule.
> >
> > ---rsk
>
>

Re: Texas internet connectivity declining due to blackouts

2021-02-22 Thread Brandon Svec via NANOG

OK, I looked closer.  I see it is a self titled opinion piece so there is
that.  Next, I see all the links in the article go to questionable sites
(not .edu or scientific organizations, etc.)  except one cherry picked NOAA
stat for a single event type for a single year.  Last, the writer is the
president of a right wing anti science lobbying group called "Spark of
Freedom" funded by Exxon Mobil.

Look, I and most everyone on this list are not qualified, experienced
climate scientists.  However, I think when you are not an expert you should
respect and believe what experts say as a group.  Picking outliers and
sharing opinions of obviously unqualified and biased people is
reprehensible and dishonest as far as I am concerned.

If you truly believe the scientific consensus around climate change is
wrong you are going to have to do a lot more than share links.  You will
have to do science and prove it.

Best.

On Mon, Feb 22, 2021 at 11:27 AM Mel Beckman  wrote:

> What offended you? The term “Global Warmist”? It’s an accurate description
> of people who hold that climate change is causing more frequent and severe
> weather, due to heating of the atmosphere.
>
> And your argument about “Forbes for something related to science” fails on
> the classic logical fallacy “appeal to authority”. Just because Forbes
> states easily verifiable public facts doesn’t make them untrustworthy.
> Scientific knowledge is best established by evidence and experiment rather
> than argued through authority by “consensus”. Science is not a consensus
> enterprise.
>
>  -mel
>
> > On Feb 22, 2021, at 10:16 AM, Brandon Svec via NANOG 
> wrote:
> >
> >
> >
> >> On Feb 22, 2021, at 9:56 AM, Mel Beckman  wrote:
> >>
> >> Sorry Global Warmists,
> >
> >
> > Stopped taking you seriously or reading further right there.  Well, that
> and linking to Forbes for something related to science.
> >
> > Best.
>
>

Re: Texas internet connectivity declining due to blackouts

Rich,

Calling my opposing argument “trash”, and then falsely linking it to unrelated 
theories on vaccines, evolution, moon landings, and dietary supplements, is 
intellectually dishonest and professionally rude. Why don’t you respond to the 
facts raised in the article? Does your religion not permit that?

Either weather events are getting worse, or they aren’t. I provided solid 
evidence that they are diminishing. The truth of this issue is important to 
NANOG, because we build the infrastructure that runs the Internet, and we can’t 
afford to waste finite resources on alarmist claims. 

 -mel

> On Feb 22, 2021, at 10:23 AM, Rich Kulawiec  wrote:
> 
> On Mon, Feb 22, 2021 at 05:48:06PM +, Mel Beckman wrote:
>> Sorry Global Warmists,
> 
> Right.  Sure.  Also, the earth is 6,000 years old (and flat), the moon
> landings were faked, creationism is real, dinosaurs and humans co-existed,
> vaccines cause autism, Elvis is alive, and...how does that line go?  Oh,
> right: artificial sweeteners are safe, WMDs were in Iraq, and Anna Nicole
> married for love. [shout-out to Levon Helm]
> 
> This trash doesn't deserve rebuttal: it deserves ridicule.
> 
> ---rsk

Re: Texas internet connectivity declining due to blackouts

Rod,

I brought up a single objection to a single claim. You choose to dismiss me 
because I “don’t accept the basic premise.” Why don’t you respond to the 
specific facts I cited? Why resort to personal attacks?

Name calling is the last resort of the man with no argument.

 -mel

On Feb 22, 2021, at 10:19 AM, Rod Beck 
mailto:rod.b...@unitedcablecompany.com>> wrote:

Mel, just please remove yourself from this conversation if you don't accept the 
basic premise. Go and find those missing votes in Georgia.

Best,

-R.

From: NANOG 
mailto:nanog-bounces+rod.beck=unitedcablecompany@nanog.org>>
 on behalf of Brandon Svec via NANOG mailto:nanog@nanog.org>>
Sent: Monday, February 22, 2021 7:16 PM
To: nanog@nanog.org 
mailto:nanog@nanog.org>>
Subject: Re: Texas internet connectivity declining due to blackouts

> On Feb 22, 2021, at 9:56 AM, Mel Beckman 
> mailto:m...@beckman.org>> wrote:
>
> Sorry Global Warmists,

Stopped taking you seriously or reading further right there.  Well, that and 
linking to Forbes for something related to science.

Best.

Re: Texas internet connectivity declining due to blackouts

What offended you? The term “Global Warmist”? It’s an accurate description of 
people who hold that climate change is causing more frequent and severe 
weather, due to heating of the atmosphere. 

And your argument about “Forbes for something related to science” fails on the 
classic logical fallacy “appeal to authority”. Just because Forbes states 
easily verifiable public facts doesn’t make them untrustworthy. Scientific 
knowledge is best established by evidence and experiment rather than argued 
through authority by “consensus”. Science is not a consensus enterprise.

 -mel

> On Feb 22, 2021, at 10:16 AM, Brandon Svec via NANOG  wrote:
> 
> 
> 
>> On Feb 22, 2021, at 9:56 AM, Mel Beckman  wrote:
>> 
>> Sorry Global Warmists,
> 
> 
> Stopped taking you seriously or reading further right there.  Well, that and 
> linking to Forbes for something related to science. 
> 
> Best.

Re: Famous operational issues

2021-02-22 Thread Dovid Bender

On Mon, Feb 22, 2021 at 2:05 PM Warren Kumari  wrote:

>
>
> On Mon, Feb 22, 2021 at 12:50 PM Regis M. Donovan 
> wrote:
>
>> On Thu, Feb 18, 2021 at 07:34:39PM -0500, Patrick W. Gilmore wrote:
>> > And to put it on topic, cover your EPOs
>>
>> I worked somewhere with an uncovered EPO, which was okay until we had a
>> telco tech in who was used to a different data center where a similar
>> looking button controlled the door access, so he reflexively hit it
>> on his way out to unlock the door.  Oops.
>>
>> Also, consider what's on generator and what's not.  I worked in a
>> corporate
>> data center where we lost power.  The backup system kept all the machines
>> running, but the ventilation system was still down, so it was very warm
>> very
>> fast as everyone went around trying to shut servers down gracefully while
>> other folks propped the doors open to get some cooler air in.
>>
>
> That reminds me of another one...
>
> In parts of NYC, there are noise abatement requirements, and so many
> places have their generators mounted on the roof -- it's cheap real-estate,
> the exhaust is easier, the noise issues are less, etc.
>
> The generators usually have a smallish diesel tank, and then a much larger
> one in the basement (diesel is heavy)...
>
> So, one of the buildings that I was in was really good about testing thier
> gensets - they'd do weekly tests (usually at night), and the generators
> always worked perfectly -- right up until the time that it was actually
> needed.
> The generator fired up, the lights kept blinking, the disks kept spinning
> - but the transfer pump that pumped diesel from the basement to the roof
> was one of the few things that was not on the generator
>
>
When we were looking at one of the big carrier hotels in NYC  they said
that they had the same issue (could be it was the same one). The elevators
were also out as well. They resorted to having techs climb up an down 9
flights of stairs all day long with 5 gallon buckets of diesel and throwing
it into the generator.

>

Re: Texas internet connectivity declining due to blackouts

Saku,

I see that not one of your references addresses the facts pointed out by 
Forbes. Rather than a shotgun response, can you counter the evidence cited that 
disproves the claim that climate events are getting more frequent and severe? 

it’s a fair topic for NANOG. The idea that countervailing evidence cannot be 
tolerated is the domain of religion, not science. Science accepts, and in fact 
encourages, such evidence. 

 -mel

> On Feb 22, 2021, at 10:01 AM, Saku Ytti  wrote:
> 
> On Mon, 22 Feb 2021 at 19:57, Mel Beckman  wrote:
> 
>> Sorry Global Warmists, But Extreme Weather Events Are Becoming Less Extreme
>> Just about every type of extreme weather event is becoming less frequent and 
>> less severe in recent years as our planet continues its modest warming in 
>> the wake of the Little Ice Age. While global warming activists attempt to 
>> spin a narrative of ever-worsening weather, the objective facts tell a 
>> completely different story.
>> 
>> https://www.forbes.com/sites/jamestaylor/2013/05/08/sorry-global-warmists-but-extreme-weather-events-are-becoming-less-extreme/?sh=46c3b30e55a4
> 
> https://grist.org/series/skeptics/
> https://skepticalscience.com/argument.php
> http://www.realclimate.org/index.php/archives/2007/05/start-here/
> 
> --
>  ++ytti, not a climate scientist so chooses to believe them

Re: Famous operational issues

2021-02-22 Thread Jethro R Binks

On Fri, 19 Feb 2021, Andy Ringsmuth wrote:

> > I explain using my "talking to a 5 year old" voice that it 
> > most certainly is a router. He tells me that lying to airport security 
> > is a federal offense, and starts looming at me. I adjust my attitude 
> > and start explaining that it's like a computer and makes the Internet 
> > work. He gruffly hands me back the router, I put it in my bag and 
> > scurry away. As I do so, I hear him telling his colleague that it 
> > wasn't a router, and that he certainly knows what a router is, because 
> > he does woodwork…
> 
> Well, in his defense, he wasn’t wrong…   :-)

This is wjy, in the UK, we tend to pronounce "router" as "router", and 
"router" as "router", so there's no confusion.

You're welcome.

Jethro.

.  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .
Jethro R Binks, Network Manager,
Information Services Directorate, University Of Strathclyde, Glasgow, UK

The University of Strathclyde is a charitable body, registered in
Scotland, number SC015263.

Re: LOAs for Cross Connects - Something like PeeringDB for XC

On Mon, Feb 22, 2021 at 2:06 PM Randy Bush  wrote:
>
> >> way back, the rirs were very insistant that their use of rpki authority
> >> was most emphatically not to be considered an identity service.  this
> >> permeated the design; e.g., organization names were specifically
> >> forbidden in certificate CN, Subject Alternative Name, etc.
> >>
> >
> > yup, I agree... though the b2b stuff George/Geoff have written up LOOKS like
> > it could be useful for this LOA type discussion. The spaghetti draft appears
> > to also fill this niche...
> >
> > Neither are particularly rooted in the RPKI except that the CA certs are 
> > being
> > used as a method to attest that a 'thing' exists, and that something signed
> > that 'thing' as proof of knowledge (I guess, really). Effectively this is:
> >   1) I am 'ca-foo' in a tree that you can trust knows I am 'foo'.
> >   2) I signed this blob (LOA)
> >   3) I asked jane at bar.com to sign as well
> >   4) you can verify me (because rpki tree) and you can verify Jane because 
> > she's
> >   also using her RPKI ca cert.
> >
> > this may be a little cumbersome to sort through, especially if all parties 
> > here
> > aren't party to the RPKI (did equinix plumb the RPKI into their customer 
> > portal
> > and all of the things required to make a x-connect work in this manner?), 
> > but I
> > imagine that if this gets wings it could be automated and it could be 
> > reliable
> > and all parties (except the colo folks perhaps?) may already have incentives
> > in places to use their RPKI goop for this function.
>
> this would work only if the LOA is being sent to someone with whom my
> contract is signed with a key validating through the same hierarchy, or
> the credential was associated contractually.  i do not think equinix
> is up to this yet.

I agree that 'today' equinix (or the notional other Colo/etc folk) are
unlikely to be
able to make this work. In the future though, if the colo's customers are RPKI
users, and the colo has some (probably) relatively simple code in hand
they could
perform verifications of this sort.

I guess I didn't ask: "Do you want this to work 'now'? or is '2 to 5
yrs acceptable'?" :)

Of course any new thing must be better than the world-of-faxes we live in today
for this solution space, and it has to be adotable by the parties involved.

-chris

Re: LOAs for Cross Connects - Something like PeeringDB for XC

>> way back, the rirs were very insistant that their use of rpki authority
>> was most emphatically not to be considered an identity service.  this
>> permeated the design; e.g., organization names were specifically
>> forbidden in certificate CN, Subject Alternative Name, etc.
>>
> 
> yup, I agree... though the b2b stuff George/Geoff have written up LOOKS like
> it could be useful for this LOA type discussion. The spaghetti draft appears
> to also fill this niche...
> 
> Neither are particularly rooted in the RPKI except that the CA certs are being
> used as a method to attest that a 'thing' exists, and that something signed
> that 'thing' as proof of knowledge (I guess, really). Effectively this is:
>   1) I am 'ca-foo' in a tree that you can trust knows I am 'foo'.
>   2) I signed this blob (LOA)
>   3) I asked jane at bar.com to sign as well
>   4) you can verify me (because rpki tree) and you can verify Jane because 
> she's
>   also using her RPKI ca cert.
> 
> this may be a little cumbersome to sort through, especially if all parties 
> here
> aren't party to the RPKI (did equinix plumb the RPKI into their customer 
> portal
> and all of the things required to make a x-connect work in this manner?), but 
> I
> imagine that if this gets wings it could be automated and it could be reliable
> and all parties (except the colo folks perhaps?) may already have incentives
> in places to use their RPKI goop for this function.

this would work only if the LOA is being sent to someone with whom my
contract is signed with a key validating through the same hierarchy, or
the credential was associated contractually.  i do not think equinix
is up to this yet.

randy

Re: Famous operational issues

2021-02-22 Thread Warren Kumari

On Mon, Feb 22, 2021 at 12:50 PM Regis M. Donovan 
wrote:

> On Thu, Feb 18, 2021 at 07:34:39PM -0500, Patrick W. Gilmore wrote:
> > And to put it on topic, cover your EPOs
>
> I worked somewhere with an uncovered EPO, which was okay until we had a
> telco tech in who was used to a different data center where a similar
> looking button controlled the door access, so he reflexively hit it
> on his way out to unlock the door.  Oops.
>
> Also, consider what's on generator and what's not.  I worked in a corporate
> data center where we lost power.  The backup system kept all the machines
> running, but the ventilation system was still down, so it was very warm
> very
> fast as everyone went around trying to shut servers down gracefully while
> other folks propped the doors open to get some cooler air in.
>

That reminds me of another one...

In parts of NYC, there are noise abatement requirements, and so many places
have their generators mounted on the roof -- it's cheap real-estate, the
exhaust is easier, the noise issues are less, etc.

The generators usually have a smallish diesel tank, and then a much larger
one in the basement (diesel is heavy)...

So, one of the buildings that I was in was really good about testing thier
gensets - they'd do weekly tests (usually at night), and the generators
always worked perfectly -- right up until the time that it was actually
needed.
The generator fired up, the lights kept blinking, the disks kept spinning -
but the transfer pump that pumped diesel from the basement to the roof was
one of the few things that was not on the generator

W

>
> --r
>
>

-- 
The computing scientist’s main challenge is not to get confused by the
complexities of his own making.
  -- E. W. Dijkstra

Re: LOAs for Cross Connects - Something like PeeringDB for XC

On Mon, Feb 22, 2021 at 1:39 PM Randy Bush  wrote:
>
> > are you asking about something like this:
> >   https://datatracker.ietf.org/doc/draft-spaghetti-sidrops-rpki-rsc/
> >
> > Which COULD be used to, as an AS holder:
> >   "sign something to be sent between you and the colo and your intended 
> > peer"
> >
> > that you could sign (with your rpki stuffs) and your peer could also
> > sign with their 'rpki stuffs', and which the colo provider could
> > automatically validate and action upon final signature(s) received.
>
> chris,
>
> way back, the rirs were very insistant that their use of rpki authority
> was most emphatically not to be considered an identity service.  this
> permeated the design; e.g., organization names were specifically
> forbidden in certificate CN, Subject Alternative Name, etc.
>

yup, I agree... though the b2b stuff George/Geoff have written up LOOKS like
it could be useful for this LOA type discussion. The spaghetti draft appears
to also fill this niche...

Neither are particularly rooted in the RPKI except that the CA certs are being
used as a method to attest that a 'thing' exists, and that something signed
that 'thing' as proof of knowledge (I guess, really). Effectively this is:
  1) I am 'ca-foo' in a tree that you can trust knows I am 'foo'.
  2) I signed this blob (LOA)
  3) I asked jane at bar.com to sign as well
  4) you can verify me (because rpki tree) and you can verify Jane because she's
  also using her RPKI ca cert.

this may be a little cumbersome to sort through, especially if all parties here
aren't party to the RPKI (did equinix plumb the RPKI into their customer portal
and all of the things required to make a x-connect work in this manner?), but I
imagine that if this gets wings it could be automated and it could be reliable
and all parties (except the colo folks perhaps?) may already have incentives
in places to use their RPKI goop for this function.

-chris

> aside: of course a few rirs thought that *their* names should be in
> their certs as exeptions.  i remember the laughter.
>
> randy
>
> ---
> ra...@psg.com
> `gpg --locate-external-keys --auto-key-locate wkd ra...@psg.com`
> signatures are back, thanks to dmarc header mangling

Re: Texas internet connectivity declining due to blackouts

2021-02-22 Thread Stephen Satchell

When I lived in Oklahoma, the mantra of the locals was "if you don't 
like the weather, wait five minutes."  As a member of a Boy Scout troop 
in the northern part of the Sooner State, we were told, repeatedly, to 
expect anything from broiling to deep freeze on our campouts.


One such outing was on fallow farmland.  Because the campsite was in the 
middle of nowhere, we were a small group, and came in three cars.  We 
pitched four tents.  During the night, a gullywasher came through and 
dropped several inches of water in one hour.  Three of the tents were 
inundated with water, and the campers ended up sleeping in the cars.  My 
tent was dry inside, because my tent-mate and I had seen the storm 
clouds, dug a trench around the tent, and loosened the ropes.  It helped 
that we had pitched the tent on a slight mound.


Some disasters are unavoidable, like tornados.  Others allow for 
mitigation by the thoughtful.


On 2/22/21 9:18 AM, Rich Kulawiec wrote:

On Tue, Feb 16, 2021 at 12:23:22PM +, Bret Clark wrote:

Texas doesn't generally experience this type of extreme cold.


That was then; this is now.

As scientist Jeff Masters put it most of a decade ago:

The atmosphere I grew up with no longer exists.  My new motto
with regards to the weather is, "expect the unprecedented."

In the years since he's said that we've seen a number of unprecedented
events: Sandy, Harvey, California wildfires, last year's midwest derecho,
and so on.  This event in Texas is just another one; there will be more;
they'll get worse.

We should probably get ready for that.

---rsk

Re: Texas internet connectivity declining due to blackouts

2021-02-22 Thread Saku Ytti

On Mon, 22 Feb 2021 at 20:28, Rich Kulawiec  wrote:

> right: artificial sweeteners are safe, WMDs were in Iraq, and Anna Nicole

Hope you meant to write 'unsafe', as the conspiracy theory is that
aspartame is unsafe, the science says it is safe.

https://en.wikipedia.org/wiki/Aspartame_controversy


-- 
  ++ytti

Re: LOAs for Cross Connects - Something like PeeringDB for XC

> are you asking about something like this:
>   https://datatracker.ietf.org/doc/draft-spaghetti-sidrops-rpki-rsc/
> 
> Which COULD be used to, as an AS holder:
>   "sign something to be sent between you and the colo and your intended peer"
> 
> that you could sign (with your rpki stuffs) and your peer could also
> sign with their 'rpki stuffs', and which the colo provider could
> automatically validate and action upon final signature(s) received.

chris,

way back, the rirs were very insistant that their use of rpki authority
was most emphatically not to be considered an identity service.  this
permeated the design; e.g., organization names were specifically
forbidden in certificate CN, Subject Alternative Name, etc.

aside: of course a few rirs thought that *their* names should be in
their certs as exeptions.  i remember the laughter.

randy

---
ra...@psg.com
`gpg --locate-external-keys --auto-key-locate wkd ra...@psg.com`
signatures are back, thanks to dmarc header mangling

Re: Texas internet connectivity declining due to blackouts

2021-02-22 Thread Rich Kulawiec

On Mon, Feb 22, 2021 at 05:48:06PM +, Mel Beckman wrote:
> Sorry Global Warmists,
 
Right.  Sure.  Also, the earth is 6,000 years old (and flat), the moon
landings were faked, creationism is real, dinosaurs and humans co-existed,
vaccines cause autism, Elvis is alive, and...how does that line go?  Oh,
right: artificial sweeteners are safe, WMDs were in Iraq, and Anna Nicole
married for love. [shout-out to Levon Helm]

This trash doesn't deserve rebuttal: it deserves ridicule.

---rsk

Re: Texas internet connectivity declining due to blackouts

2021-02-22 Thread Rod Beck

Mel, just please remove yourself from this conversation if you don't accept the 
basic premise. Go and find those missing votes in Georgia.

Best,

-R.

From: NANOG  on behalf 
of Brandon Svec via NANOG 
Sent: Monday, February 22, 2021 7:16 PM
To: nanog@nanog.org 
Subject: Re: Texas internet connectivity declining due to blackouts

> On Feb 22, 2021, at 9:56 AM, Mel Beckman  wrote:
>
> Sorry Global Warmists,

Stopped taking you seriously or reading further right there.  Well, that and 
linking to Forbes for something related to science.

Best.

Re: LOAs for Cross Connects - Something like PeeringDB for XC

> But it looks like a "crypto sign and publishes" anything related to an
> organization.

that is the problem with this discussion.  it does not.  it allows one
to show ownership of an AS or prefix.  it does not show ownership or
authority over an organization.  keep your trust model straight.

randy

Re: Texas internet connectivity declining due to blackouts

2021-02-22 Thread Brandon Svec via NANOG




> On Feb 22, 2021, at 9:56 AM, Mel Beckman  wrote:
> 
> Sorry Global Warmists,


Stopped taking you seriously or reading further right there.  Well, that and 
linking to Forbes for something related to science. 

Best.

Re: Texas internet connectivity declining due to blackouts

2021-02-22 Thread Saku Ytti

On Mon, 22 Feb 2021 at 19:57, Mel Beckman  wrote:

> Sorry Global Warmists, But Extreme Weather Events Are Becoming Less Extreme
> Just about every type of extreme weather event is becoming less frequent and 
> less severe in recent years as our planet continues its modest warming in the 
> wake of the Little Ice Age. While global warming activists attempt to spin a 
> narrative of ever-worsening weather, the objective facts tell a completely 
> different story.
>
> https://www.forbes.com/sites/jamestaylor/2013/05/08/sorry-global-warmists-but-extreme-weather-events-are-becoming-less-extreme/?sh=46c3b30e55a4

https://grist.org/series/skeptics/
https://skepticalscience.com/argument.php
http://www.realclimate.org/index.php/archives/2007/05/start-here/

--
  ++ytti, not a climate scientist so chooses to believe them

Re: Texas internet connectivity declining due to blackouts

Sorry Global Warmists, But Extreme Weather Events Are Becoming Less Extreme
Just about every type of extreme weather event is becoming less frequent and
less severe in recent years as our planet continues its modest warming in the
wake of the Little Ice Age. While global warming activists attempt to spin a
narrative of ever-worsening weather, the objective facts tell a completely
different story.

https://www.forbes.com/sites/jamestaylor/2013/05/08/sorry-global-warmists-but-extreme-weather-events-are-becoming-less-extreme/?sh=46c3b30e55a4

The above article is from 2013 article provides hard data that extreme weather
events are becoming less, not more, frequent and severe. That trend continues
today. So the premise that weather events are getting “increasingly
unprecedented" is demonstrably false. What is true is that cost of extreme
weather events has gone up, but that’s a function of population density, not of
weather itself, nor of climate change.

The data that climate alarmists most often cite to bolster the false narrative
of increasing severe weather is the “Billion-Dollar Weather and Climate
Disaster Events” chart:

https://www.ncdc.noaa.gov/billions/events

The worst event ever, at $170B, was the merely Cat3 hurricane Katrina. It
caused $170B in damage because it hit the highly dense New Orleans urban
center. Yet the massively more severe Cat5 hurricane Michael, which hit a much
less population-dense Mexico Beach, FL, caused only $25B in damage. Weather
intensity doesn’t correlate with damage costs. Population does.

Knowing accurately the true trend of severe weather is important to NANOG
operational planning, because it bears directly on the cost/benefit analysis of
infrastructure hardening. Physical integrity is always a function of cost vs
benefit, and we as network operators can’t afford to spend “unprecedented" sums
for diminishing benefits. We must use facts, not alarmism, to decide on what we
should "probably get ready for”.

-mel

On Feb 22, 2021, at 9:18 AM, Rich Kulawiec wrote:

On Tue, Feb 16, 2021 at 12:23:22PM +, Bret Clark wrote:
Texas doesn't generally experience this type of extreme cold.

That was then; this is now.

As scientist Jeff Masters put it most of a decade ago:

The atmosphere I grew up with no longer exists. My new motto
with regards to the weather is, "expect the unprecedented."

In the years since he's said that we've seen a number of unprecedented
events: Sandy, Harvey, California wildfires, last year's midwest derecho,
and so on. This event in Texas is just another one; there will be more;
they'll get worse.

We should probably get ready for that.

---rsk

Re: Famous operational issues

2021-02-22 Thread Regis M. Donovan

On Thu, Feb 18, 2021 at 07:34:39PM -0500, Patrick W. Gilmore wrote:
> And to put it on topic, cover your EPOs

I worked somewhere with an uncovered EPO, which was okay until we had a
telco tech in who was used to a different data center where a similar
looking button controlled the door access, so he reflexively hit it
on his way out to unlock the door.  Oops.

Also, consider what's on generator and what's not.  I worked in a corporate
data center where we lost power.  The backup system kept all the machines
running, but the ventilation system was still down, so it was very warm very
fast as everyone went around trying to shut servers down gracefully while
other folks propped the doors open to get some cooler air in.

--r

Re: Support for End User Services

2021-02-22 Thread Keith Stokes

I’ve always used wording such as “I’m contacting you on behalf of so-and-so.” 
If they ask further I usually tell them I’m a consultant.



On Feb 20, 2021, at 11:29 AM, Mike Hammett 
mailto:na...@ics-il.net>> wrote:

Leave aside any conversation about whether the business has the ability (or 
approval) to pay for it or not.


Is it appropriate for organizations that provide services to end-users to 
require that you are a paying customer to contact their support?

Is it appropriate to pretend to be your complaining customer to get support on 
network-level issues (IP Geolocation, false VPN notices, buffering, despite a 
clean path to their CDN, etc.)?



-
Mike Hammett
Intelligent Computing Solutions
[http://www.ics-il.com/images/fbicon.png][http://www.ics-il.com/images/googleicon.png][http://www.ics-il.com/images/linkedinicon.png][http://www.ics-il.com/images/twittericon.png]
Midwest Internet Exchange
[http://www.ics-il.com/images/fbicon.png][http://www.ics-il.com/images/linkedinicon.png][http://www.ics-il.com/images/twittericon.png]
The Brothers WISP
[http://www.ics-il.com/images/fbicon.png][http://www.ics-il.com/images/youtubeicon.png]

Re: Famous operational issues

Long ago, in a galaxy far away I worked for a gov't contractor on site
at a gov't site...

We had our own cute little datacenter, and our 4 building complex had
a central power distribution setup from utility -> buildings.
It was really quite nice :) (the job, the buildings, the power and
cute little datacenter)

One fine Tues afternoon ~2pm local time, the building engineers
decided they would make a copy of the key used to turn the main /
utility power off...
Of course they also needed to make sure their copy worked, so... they
put the key in and turned it.

Shockingly, the key worked! and no power was provided to the buildings :(
It was very suddenly very dark and very quiet... (then the yelling started)

Ok, fast forward 7 days... rerun the movie... Yes, the same building
engineers made a new copy, and .. tested that new copy in the same
manner.

For neither of these events did someone tell the rest of us (and our
customers): "Hey, we MAY interrupt power to the buildings... FYI, BTW,
make sure your backups are current..." I recall we got the name of the
engineer the 1st time around, but not the second.

On Mon, Feb 22, 2021 at 12:26 PM Tony Finch  wrote:
>
> Patrick W. Gilmore  wrote:
> >
> >   Me: Did you order that EPO cover?
> >   Her: Nope.
>
> There are apparently two kinds of EPO cover:
>
> - the kind that stops you from pressing the button by mistake;
>
> - and the kind that doesn't, and instead locks the button down to make
> sure it isn't un-pressed until everything is safe.
>
> We had a series of incidents similar to yours, so an EPO cover was
> belatedly installed. We learned about the second kind of EPO cover when a
> colleague proudly demonstrated that the EPO button should no longer be
> pressed by accident, or so he thought.
>
> Tony.
> --
> f.anthony.n.finchhttp://dotat.at/
> the quest for freedom and justice can never end

Re: Famous operational issues

2021-02-22 Thread Warren Kumari

On Mon, Feb 22, 2021 at 7:09 AM t...@pelican.org  wrote:

> On Thursday, 18 February, 2021 22:37, "Warren Kumari" 
> said:
>
> > 4: Not too long after I started doing networking (and for the same small
> > ISP in Yonkers), I'm flying off to install a new customer. I (of course)
> > think that I'm hot stuff because I'm going to do the install, configure
> the
> > router, whee, look at me! Anyway, I don't want to check a bag, and so I
> > stuff the Cisco 2501 in a carryon bag, along with tools, etc (this was
> all
> > pre-9/11!). I'm going through security and the TSA[0] person opens my bag
> > and pulls the router out. "What's this?!" he asks. I politely tell him
> that
> > it's a router. He says it's not. I'm still thinking that I'm the new
> > hotness, and so I tell him in a somewhat condescending way that it is,
> and
> > I know what I'm talking about. He tells me that it's not a router, and is
> > starting to get annoyed. I explain using my "talking to a 5 year old"
> voice
> > that it most certainly is a router. He tells me that lying to airport
> > security is a federal offense, and starts looming at me. I adjust my
> > attitude and start explaining that it's like a computer and makes the
> > Internet work. He gruffly hands me back the router, I put it in my bag
> and
> > scurry away. As I do so, I hear him telling his colleague that it wasn't
> a
> > router, and that he certainly knows what a router is, because he does
> > woodwork...
>
> Here in the UK we avoid that issue by pronouncing the packet-shifter as
> "rooter", and only the wood-working tool as "rowter" :)
>
> Of course, it raises a different set of problems when talking to the
> Australians...
>

Yes. I discovered this while walking around Sydney wearing my "I have root
@ Google" t-shirt got some odd looks/snickers...

W




>
> Cheers,
> Tim.
>
>
>

-- 
The computing scientist’s main challenge is not to get confused by the
complexities of his own making.
  -- E. W. Dijkstra

Re: Texas internet connectivity declining due to blackouts

2021-02-22 Thread Rod Beck

Exactly. The weather is not a stationary time series. The moments of the 
probability distribution are not time invariant.

From: NANOG  on behalf 
of Rich Kulawiec 
Sent: Monday, February 22, 2021 6:18 PM
To: nanog@nanog.org 
Subject: Re: Texas internet connectivity declining due to blackouts

On Tue, Feb 16, 2021 at 12:23:22PM +, Bret Clark wrote:
> Texas doesn't generally experience this type of extreme cold.

That was then; this is now.

As scientist Jeff Masters put it most of a decade ago:

The atmosphere I grew up with no longer exists.  My new motto
with regards to the weather is, "expect the unprecedented."

In the years since he's said that we've seen a number of unprecedented
events: Sandy, Harvey, California wildfires, last year's midwest derecho,
and so on.  This event in Texas is just another one; there will be more;
they'll get worse.

We should probably get ready for that.

---rsk

Re: Famous operational issues

2021-02-22 Thread Tony Finch

Patrick W. Gilmore  wrote:
>
>   Me: Did you order that EPO cover?
>   Her: Nope.

There are apparently two kinds of EPO cover:

- the kind that stops you from pressing the button by mistake;

- and the kind that doesn't, and instead locks the button down to make
sure it isn't un-pressed until everything is safe.

We had a series of incidents similar to yours, so an EPO cover was
belatedly installed. We learned about the second kind of EPO cover when a
colleague proudly demonstrated that the EPO button should no longer be
pressed by accident, or so he thought.

Tony.
-- 
f.anthony.n.finchhttp://dotat.at/
the quest for freedom and justice can never end

Re: Texas internet connectivity declining due to blackouts

2021-02-22 Thread Rich Kulawiec

On Wed, Feb 17, 2021 at 10:34:35AM -0800, Sabri Berisha wrote:
> With apologies to those on the list who still use mutt/pine etc. 

1. "still"?  Competent professionals with security awareness use text-only
email clients as a matter of basic self-defense.  I trust it's obvious
why those of us who are responsible for systems/networks/data need to
protect ourselves in order to protect the resources we run.

2. It's pretty easy to handle attachments gracefully while using mutt.
By default it checks ~/.mailcap, and in that file one can specify external
programs to handle various attachment types, e.g.:

text/html; w3m -dump -T text/html %s | less
application/pdf; /usr/bin/evince %s

Of course this needs to be done carefully, since it subjects the user
to attacks against those applications carried via attachments, but
that's where judicious choices on the part of the user come in --
choices as in "which attachment types, which applications, and
whether or not to exercise this capability on a per-message basis".

---rsk

Re: Texas internet connectivity declining due to blackouts

2021-02-22 Thread Rich Kulawiec

On Tue, Feb 16, 2021 at 12:23:22PM +, Bret Clark wrote:
> Texas doesn't generally experience this type of extreme cold.

That was then; this is now.

As scientist Jeff Masters put it most of a decade ago:

The atmosphere I grew up with no longer exists.  My new motto
with regards to the weather is, "expect the unprecedented."

In the years since he's said that we've seen a number of unprecedented
events: Sandy, Harvey, California wildfires, last year's midwest derecho,
and so on.  This event in Texas is just another one; there will be more;
they'll get worse.

We should probably get ready for that.

---rsk

Re: LOAs for Cross Connects - Something like PeeringDB for XC

Well... I must confess that I had some difficulty on the first
understanding of what is proposed.

But after the 4 reads, I saw that this "spaghetti" thing is more
powerful than I could imagine!


Please correct me if I'm no right:
But it looks like a "crypto sign and publishes" anything related to an
organization.

Yes, I think that with some effort CrossConnect LOAs can be fitted inside
of it...
I'm not sure if it is the better solution for the scope of LOAs, but
certainly is a valid discussion.


What is bubbling in my mind is the standard data model for each type of
different attribute that can exist...
Who will define that?



Em seg., 22 de fev. de 2021 às 12:26, Christopher Morrow <
morrowc.li...@gmail.com> escreveu:

> On Mon, Feb 22, 2021 at 9:19 AM Douglas Fischer
>  wrote:
> >
> > I believe that almost everyone in here knows that LOAs for Cross
> Connects in Datacenters and Telecom Rooms can be a pain...
> >
> > I don't know if I'm suggesting something that already exists.
> > Or even if I'm suggesting something that could be unpopular for some
> reason.
> >
> > But every time I need to deal with some Cross-Connect LOA, and mostly
> when we face some rework on data mistakes, I dream with a "PeeringDB for
> Cross Connects".
> >
>
> are you asking about something like this:
>   https://datatracker.ietf.org/doc/draft-spaghetti-sidrops-rpki-rsc/
>
> Which COULD be used to, as an AS holder:
>   "sign something to be sent between you and the colo and your intended
> peer"
>
> that you could sign (with your rpki stuffs) and your peer could also
> sign with their 'rpki stuffs', and which the colo provider could
> automatically validate and action upon final signature(s) received.
>
> > So, this mail is a question and also a suggestion.
> >
> >
> > There is something like an "online notary's office" exclusive for
> Cross-Connect LOAs?
> >  - Somewhere Organizations can register information authorizing
> connections of Port on their Places (Cages, Racks, etc)...
> >
>
> The RPKI data today doesn't contain information about
> cages/ports/patch-panels, so possibly the spaghetti draft isn't a
> terrific fit?
>
> > If it doesn't exist. What would be necessary for that?
> > Mostly considering the PeeringDB work model.
> >  - OpenSource.
> >  - Free access to the tool, and sponsors to keep the project alive.
> >  - API driven, with some Web-gui.
> > And considering some data-modeling.
> >  - Most of the data being Open/Public (Organizations,
> Facilities(Datacenters and/or Telecom-Rooms), Presence on Facilities, etc).
> >  - Access control to Information that can not be public (A-side
> organization, Z-Side Organization, PathPanel/Port).
> > And some workflow
> >  - Cross Connect Requiremento/Authorization from A-Side
> >  - Acceptance/Authorization from Z-side.
> >  - Acceptance/Authorization from Facilities involved (could be more than
> one)
> >  - Execution/Activation notice from Facilities.
> >
> >
> > --
> > Douglas Fernando Fischer
> > Engº de Controle e Automação
>


-- 
Douglas Fernando Fischer
Engº de Controle e Automação

Re: DPDK and energy efficiency

I forgot to point out that on Friday 26th, I'll share the results collected
through a link or a series of screenshots.

Cheers,

Etienne

On Mon, Feb 22, 2021 at 2:15 PM Pawel Malachowski <
pawmal-na...@freebsd.lublin.pl> wrote:

> Dnia Mon, Feb 22, 2021 at 01:01:45PM +0100, Etienne-Victor Depasquale
> napisał(a):
>
> > It is, after all, Intel's response to the problem of general-purpose
> > scheduling of its processors - which prevents the processor from being
> > viable under high networking loads.
>
> It totally makes sense to busy poll under high networking load.
> By high networking load I mean roughly > 7 Mpps RX+TX per one x86 CPU core.
>
> I partially agree it may be hard to mix DPDK and non-DPDK workload
> on a single CPU, not only because of advanced power management logic
> requirement for the dataplane application, but also due to LLC trashing.
> It heavily depends on usecase and dataset sizes, for example
> optimised FIB may fit nicely into cache and use only tiny, hot part
> of the dataset, but CGNAT Mflow mapping likely won't fit. For such
> a usecase I would recommand dedicated CPU or cache partitioning (CAT),
> if available.
>
> In case of low volume traffic like 20-40G of IMIX one can dedicate
> e.g. 2 cores and interleave busy polling with halt instructions to
> lower the usage significantly (~60-80% core underutilisation).
>
>
>
> --
> Pawel Malachowski
> @pawmal80
>


-- 
Ing. Etienne-Victor Depasquale
Assistant Lecturer
Department of Communications & Computer Engineering
Faculty of Information & Communication Technology
University of Malta
Web. https://www.um.edu.mt/profile/etiennedepasquale

Re: Famous operational issues

2021-02-22 Thread Bruce H McIntosh





On 2/22/21 9:14 AM, Alain Hebert wrote:

*[External Email]*

     Well...

     During my younger days, that button was used a few time by the 
operator of a VM/370 to regain control from someone with a "curious 
mind" *cought* *cought*...


Two horror stories I remember from long ago when I was a console jockey 
for a federal space agency that will remain nameless :P


1. A coworker brought her daughter to work with her on a Saturday 
overtime shift because she couldn't get a babysitter. She parked the kid 
with a coloring book and a pile of crayons at the only table in the 
console room with some space, right next to the master console for our 
3081. I asked her to make sure sh was well away from the console, and as 
she reached over to scoot the girl and her coloring books further away 
she slipped, and reached out to steady herself. Yep, planted her finger 
right down on the IML button (plexi covers? We don' need no STEENKIN' 
plexi covers!). MVS and VM vanished, two dozen tape drives rewound and 
several hours' worth of data merge jobs went blooey.


2. The 3081 was water cooled via a heat exchanger. The building chilled 
water feed had a very old, very clogged filter that was bypassed until 
it could be replaced. One day a new maintenance foreman came through the 
building doing his "clipboard and harried expression" thing, and spotted 
the filter in bypass (NO, I don't know WHY it hadn't been red-tagged. 
Someone clearly dropped that ball.) He thought, "Well that's not right" 
and reset all the valves to put it back inline, which of course, pretty 
much killed the chilled water flow through the heat exchanger. First 
thing we knew about it in Operations was when the 3081 started throwing 
thermal alarms and MVS crashed hard. IBM had to replace several modules 
in the CPUs.


--

Bruce H. McIntosh
Network Engineer II
University of Florida Information Technology
b...@ufl.edu
352-273-1066

Re: LOAs for Cross Connects - Something like PeeringDB for XC

On Mon, Feb 22, 2021 at 9:19 AM Douglas Fischer
 wrote:
>
> I believe that almost everyone in here knows that LOAs for Cross Connects in 
> Datacenters and Telecom Rooms can be a pain...
>
> I don't know if I'm suggesting something that already exists.
> Or even if I'm suggesting something that could be unpopular for some reason.
>
> But every time I need to deal with some Cross-Connect LOA, and mostly when we 
> face some rework on data mistakes, I dream with a "PeeringDB for Cross 
> Connects".
>

are you asking about something like this:
  https://datatracker.ietf.org/doc/draft-spaghetti-sidrops-rpki-rsc/

Which COULD be used to, as an AS holder:
  "sign something to be sent between you and the colo and your intended peer"

that you could sign (with your rpki stuffs) and your peer could also
sign with their 'rpki stuffs', and which the colo provider could
automatically validate and action upon final signature(s) received.

> So, this mail is a question and also a suggestion.
>
>
> There is something like an "online notary's office" exclusive for 
> Cross-Connect LOAs?
>  - Somewhere Organizations can register information authorizing connections 
> of Port on their Places (Cages, Racks, etc)...
>

The RPKI data today doesn't contain information about
cages/ports/patch-panels, so possibly the spaghetti draft isn't a
terrific fit?

> If it doesn't exist. What would be necessary for that?
> Mostly considering the PeeringDB work model.
>  - OpenSource.
>  - Free access to the tool, and sponsors to keep the project alive.
>  - API driven, with some Web-gui.
> And considering some data-modeling.
>  - Most of the data being Open/Public (Organizations, Facilities(Datacenters 
> and/or Telecom-Rooms), Presence on Facilities, etc).
>  - Access control to Information that can not be public (A-side organization, 
> Z-Side Organization, PathPanel/Port).
> And some workflow
>  - Cross Connect Requiremento/Authorization from A-Side
>  - Acceptance/Authorization from Z-side.
>  - Acceptance/Authorization from Facilities involved (could be more than one)
>  - Execution/Activation notice from Facilities.
>
>
> --
> Douglas Fernando Fischer
> Engº de Controle e Automação

RE: CGNAT

2021-02-22 Thread na...@jima.us

While I don't doubt the accuracy of Lee's presentation at the time, at least 
two base factors have changed since then:

- Greater deployment of IPv6 content (necessitating less CGN capacity per user)
- Increased price of Legacy IP space on the secondary market (changing the 
formula) -- strictly speaking, this presentation was still in "primary market" 
era for LACNIC/ARIN/AFRINIC

IPv6 migration is not generally aided by CGNAT, but CGNAT deployment is 
generally aided by IPv6 deployment; to reiterate the earlier point, any ISPs 
deploying CGNAT without first deploying IPv6 are burning cash.

- Jima

From: NANOG On Behalf Of Owen DeLong
Sent: Sunday, February 21, 2021 16:59
To: Steve Saner
Cc: nanog@nanog.org
Subject: Re: CGNAT

On Feb 18, 2021, at 8:38 AM, Steve Saner wrote:

> We are starting to look at CGNAT solutions. The primary motivation at the 
> moment is to extend current IPv4 resources, but IPv6 migration is also a 
> factor.

IPv6 Migration is generally not aided by CGNAT.

In general, the economics today still work out to make purchasing or leasing 
addresses more favorable than CGNAT.

It’s a bit dated by now, but still very relevant, see Lee Howard’s excellent 
research presented at the 2012 Rocky
mountain v6 task force meeting:

https://www.rmv6tf.org/wp-content/uploads/2012/11/TCO-of-CGN1.pdf

Owen

We've been in touch with A10. Just wondering if there are some alternative 
vendors that anyone would recommend. We'd probably be looking at a solution to 
support 5k to 15k customers and bandwidth up to around 30-40 gig as a starting 
point. A solution that is as transparent to user experience as possible is a 
priority.

Thanks

-- 
Steve Saner
ideatek HUMAN AT OUR VERY FIBER
This email transmission, and any documents, files or previous email messages 
attached to it may contain confidential information. If the reader of this 
message is not the intended recipient or the employee or agent responsible for 
delivering the message to the intended recipient, you are hereby notified that 
any dissemination, distribution or copying of this communication is strictly 
prohibited. If you are not, or believe you may not be, the intended recipient, 
please advise the sender immediately by return email or by calling 
tel:620.543.5026. Then take all steps necessary to permanently delete the email 
and all attachments from your computer system.

LOAs for Cross Connects - Something like PeeringDB for XC

I believe that almost everyone in here knows that LOAs for Cross Connects
in Datacenters and Telecom Rooms can be a pain...

I don't know if I'm suggesting something that already exists.
Or even if I'm suggesting something that could be unpopular for some reason.

But every time I need to deal with some Cross-Connect LOA, and mostly
when we face some rework on data mistakes, I dream with a "PeeringDB for
Cross Connects".

So, this mail is a question and also a suggestion.


There is something like an "online notary's office" exclusive for
Cross-Connect LOAs?
 - Somewhere Organizations can register information authorizing connections
of Port on their Places (Cages, Racks, etc)...

If it doesn't exist. What would be necessary for that?
Mostly considering the PeeringDB work model.
 - OpenSource.
 - Free access to the tool, and sponsors to keep the project alive.
 - API driven, with some Web-gui.
And considering some data-modeling.
 - Most of the data being Open/Public (Organizations,
Facilities(Datacenters and/or Telecom-Rooms), Presence on Facilities, etc).
 - Access control to Information that can not be public (A-side
organization, Z-Side Organization, PathPanel/Port).
And some workflow
 - Cross Connect Requiremento/Authorization from A-Side
 - Acceptance/Authorization from Z-side.
 - Acceptance/Authorization from Facilities involved (could be more than
one)
 - Execution/Activation notice from Facilities.


-- 
Douglas Fernando Fischer
Engº de Controle e Automação

Re: Famous operational issues

2021-02-22 Thread Alain Hebert


    Well...

    During my younger days, that button was used a few time by the 
operator of a VM/370 to regain control from someone with a "curious 
mind" *cought* *cought*...


-
Alain Hebertaheb...@pubnix.net
PubNIX Inc.
50 boul. St-Charles
P.O. Box 26770 Beaconsfield, Quebec H9W 6G7
Tel: 514-990-5911  http://www.pubnix.netFax: 514-990-9443

On 2/20/21 4:07 AM, Henry Yen wrote:

On Thu, Feb 18, 2021 at 07:34:39AM -0500, Patrick W. Gilmore wrote:

In 1994, there was a major earthquake near the city of Los Angeles. City hall 
had to be evacuated and it would take over a year to reinforce the building to 
make it habitable again. My company moved all the systems in the basement of 
city hall to a new datacenter a mile or so away. After the install, we spent 
more than a week coaxing their ancient (even for 1994) machines back online, 
such as a Prime Computer and an AS400 with tons of DASD. Well, tons of 
cabinets, certainly less storage than my watch has now.

I was in the DC going over something with the lady in charge when someone 
walked in to ask her something. She said â€œjust a secondâ€. That person took 
one step to the side of the door and leaned against the wall - right on an EPO 
which had no cover.

Have you ever heard an entire row of DASD spin down instantly? Or taken 40 
minutes to IPL an AS400? In the middle of the business day? For the second most 
populous city in the country?

Me: Maybe you should get a cover for that?
Her: Good idea.

Couple weeks later, in the same DC, going over final checklist. A fedex guy 
walks in. (To this day, no idea how he got in a supposedly locked DC.) She says 
â€œjust a secondâ€, and I get a very strong deja vu feeling. He takes one step 
to the side and leans against the wall.

Me: Did you order that EPO cover?
Her: Nope.

some of the ibm 4300 series mini-mainframes came with a console terminal
that had a very large, raised (completely not flush), alternate power
button on the upper panel of the keyboard, facing the operator. in later
versions, the button was inset in a little open box with high sides. in
earlier versions, there was just a pair of raised ribs on either side of the
button. in the earliest version, if that panel needed to be replaced, the
replacement part didn't even have those protective ribs, this huge button
was just sitting there. on our 4341, someone had dropped the keyboard during
installation and the damaged panel was replaced with the
no-protection-whatsoever part.

i had an operator who, working a double shift into the overnight run,
fell asleep and managed to bang his head square on the button.
the overnight jobs running were left in various states of ruin.

third party manufacturers had an easy sell for lucite power/EPO button covers.

--
Henry Yen   Aegis Information Systems, Inc.
Senior Systems Programmer   Hicksville, New York

Re: DPDK and energy efficiency

2021-02-22 Thread Pawel Malachowski

Dnia Mon, Feb 22, 2021 at 01:01:45PM +0100, Etienne-Victor Depasquale 
napisał(a):

> It is, after all, Intel's response to the problem of general-purpose
> scheduling of its processors - which prevents the processor from being
> viable under high networking loads.

It totally makes sense to busy poll under high networking load.
By high networking load I mean roughly > 7 Mpps RX+TX per one x86 CPU core.

I partially agree it may be hard to mix DPDK and non-DPDK workload
on a single CPU, not only because of advanced power management logic
requirement for the dataplane application, but also due to LLC trashing.
It heavily depends on usecase and dataset sizes, for example
optimised FIB may fit nicely into cache and use only tiny, hot part
of the dataset, but CGNAT Mflow mapping likely won't fit. For such
a usecase I would recommand dedicated CPU or cache partitioning (CAT),
if available.

In case of low volume traffic like 20-40G of IMIX one can dedicate
e.g. 2 cores and interleave busy polling with halt instructions to
lower the usage significantly (~60-80% core underutilisation).



-- 
Pawel Malachowski
@pawmal80

Re: Famous operational issues

2021-02-22 Thread t...@pelican.org

On Thursday, 18 February, 2021 22:37, "Warren Kumari"  said:

> 4: Not too long after I started doing networking (and for the same small
> ISP in Yonkers), I'm flying off to install a new customer. I (of course)
> think that I'm hot stuff because I'm going to do the install, configure the
> router, whee, look at me! Anyway, I don't want to check a bag, and so I
> stuff the Cisco 2501 in a carryon bag, along with tools, etc (this was all
> pre-9/11!). I'm going through security and the TSA[0] person opens my bag
> and pulls the router out. "What's this?!" he asks. I politely tell him that
> it's a router. He says it's not. I'm still thinking that I'm the new
> hotness, and so I tell him in a somewhat condescending way that it is, and
> I know what I'm talking about. He tells me that it's not a router, and is
> starting to get annoyed. I explain using my "talking to a 5 year old" voice
> that it most certainly is a router. He tells me that lying to airport
> security is a federal offense, and starts looming at me. I adjust my
> attitude and start explaining that it's like a computer and makes the
> Internet work. He gruffly hands me back the router, I put it in my bag and
> scurry away. As I do so, I hear him telling his colleague that it wasn't a
> router, and that he certainly knows what a router is, because he does
> woodwork...

Here in the UK we avoid that issue by pronouncing the packet-shifter as 
"rooter", and only the wood-working tool as "rowter" :)

Of course, it raises a different set of problems when talking to the 
Australians...

Cheers,
Tim.

Re: DPDK and energy efficiency

>
> It consumes 100% only if you busy poll (which is the default approach).
>
Precisely.

It is, after all, Intel's response to the problem of general-purpose
scheduling of its processors - which prevents the processor from being
viable under high networking loads.

Cheers,

Etienne

On Mon, Feb 22, 2021 at 12:58 PM Pawel Malachowski <
pawmal-na...@freebsd.lublin.pl> wrote:

> Dnia Mon, Feb 22, 2021 at 08:33:35AM -0300, Douglas Fischer napisał(a):
>
> > But IMHO, the questions do not cover the actual reality of DPDK.
> > That característic of "100% CPU" depends on several aspects, like:
> >  - How old are the hardware on DPDK.
> >  - What type of DPDK Instructions are made(Very Dynamic as Statefull
> CGNAT,
> > ou Static ACLs?)
> >  - Using or not the measurements of DPDK Input/Drop/Fowarding.
> >  - CPU Affinity done according to the demand of traffic
> >  - SR-IOV (sharing resources) on DPDK.
>
> It consumes 100% only if you busy poll (which is the default approach).
> One can switch between polling and interrupts (or monitor, if supported),
> or introduce halt instructions, in case of low/medium traffic volume.
>
>
> --
> Pawel Malachowski
> @pawmal80
>


-- 
Ing. Etienne-Victor Depasquale
Assistant Lecturer
Department of Communications & Computer Engineering
Faculty of Information & Communication Technology
University of Malta
Web. https://www.um.edu.mt/profile/etiennedepasquale

Re: DPDK and energy efficiency

Here are a few references.
Strictly speaking, DPDK and SR-IOV are orthogonal. DPDK is intended to
facilitate cloud-native operation through hardware independence. SR-IOV
presumes SR-IOV-compliant hardware.

[1] Z. Xu, F. Liu, T. Wang, and H. Xu, “Demystifying the energy efficiency
of Network Function Virtualization,”
in 2016 IEEE/ACM 24th International Symposium on Quality of Service
(IWQoS), Jun. 2016, pp. 1–10.
DOI: 10.1109/IWQoS.2016.7590429.

[2] S. Fu, J. Liu, and W. Zhu, “Multimedia Content Delivery with Network
Function Virtualization: The Energy Perspective,”
 IEEE MultiMedia, vol. 24, no. 3, pp. 38–47, 2017, ISSN: 1941-0166.
DOI: 10.1109/MMUL.2017.3051514.

[3] X. Li, W. Cheng, T. Zhang, F. Ren, and B. Yang, “Towards Power
Efficient High Performance Packet I/O,”
IEEE Transactions on Parallel and Distributed Systems, vol. 31, no. 4, pp.
981–996, April 2020,
ISSN:1558-2183. DOI: 10.1109/TPDS.2019.2957746.

[4] G. Li, D. Zhang, Y. Li, and K. Li, “Toward energy
efficiency optimization of pktgen-DPDK for green network testbeds,”
China Communications, vol. 15, no. 11, pp. 199–207, November 2018,
ISSN: 1673-5447. DOI: 10.1109/CC.2018.8543100.


On Mon, Feb 22, 2021 at 12:45 PM Etienne-Victor Depasquale 
wrote:

> The way I saw, the questions induce the public to conclude that DPDK
>> ALWAYS has 100% CPU usage, which is not true.
>
>
> I don't concur.
>
> Every research paper I've read indicates that, regardless of whether it
> has packets to process or not, DPDK PMDs (poll-mode drivers) prevent the
> CPU from falling into an LPI (low-power idle).
>
> When it has no packets to process, the PMD runs the processor in a polling
> loop that keeps utilization of the running core at 100%.
>
> Cheers,
>
> Etienne
>
> On Mon, Feb 22, 2021 at 12:33 PM Douglas Fischer 
> wrote:
>
>> I'm very happy to see interest in DPDK and power consumption.
>>
>> But IMHO, the questions do not cover the actual reality of DPDK.
>> That característic of "100% CPU" depends on several aspects, like:
>>  - How old are the hardware on DPDK.
>>  - What type of DPDK Instructions are made(Very Dynamic as
>> Statefull CGNAT, ou Static ACLs?)
>>  - Using or not the measurements of DPDK Input/Drop/Fowarding.
>>  - CPU Affinity done according to the demand of traffic
>>  - SR-IOV (sharing resources) on DPDK.
>>
>> The way I saw, the questions induce the public to conclude that DPDK
>> ALWAYS has 100% CPU usage, which is not true.
>>
>>
>> Em seg., 22 de fev. de 2021 às 04:30, Etienne-Victor Depasquale <
>> ed...@ieee.org> escreveu:
>>
>>> Hello folks,
>>>
>>> I've just followed a thread regarding use of CGNAT and noted a
>>> suggestion (regarding DANOS) that includes use of DPDK.
>>>
>>> As I'm interested in the breadth of adoption of DPDK, and as I'm a
>>> researcher into energy and power efficiency, I'd love to hear your feedback
>>> on your use of power consumption control by DPDK.
>>>
>>> I've drawn up a bare-bones, 2-question survey at this link:
>>>
>>> https://www.surveymonkey.com/r/J886DPY.
>>>
>>> Responses have been set to anonymous.
>>>
>>> Cheers,
>>>
>>> Etienne
>>>
>>> --
>>> Ing. Etienne-Victor Depasquale
>>> Assistant Lecturer
>>> Department of Communications & Computer Engineering
>>> Faculty of Information & Communication Technology
>>> University of Malta
>>> Web. https://www.um.edu.mt/profile/etiennedepasquale
>>>
>>
>>
>> --
>> Douglas Fernando Fischer
>> Engº de Controle e Automação
>>
>
>
> --
> Ing. Etienne-Victor Depasquale
> Assistant Lecturer
> Department of Communications & Computer Engineering
> Faculty of Information & Communication Technology
> University of Malta
> Web. https://www.um.edu.mt/profile/etiennedepasquale
>


-- 
Ing. Etienne-Victor Depasquale
Assistant Lecturer
Department of Communications & Computer Engineering
Faculty of Information & Communication Technology
University of Malta
Web. https://www.um.edu.mt/profile/etiennedepasquale

Re: DPDK and energy efficiency

2021-02-22 Thread Pawel Malachowski

Dnia Mon, Feb 22, 2021 at 08:33:35AM -0300, Douglas Fischer napisał(a):

> But IMHO, the questions do not cover the actual reality of DPDK.
> That característic of "100% CPU" depends on several aspects, like:
>  - How old are the hardware on DPDK.
>  - What type of DPDK Instructions are made(Very Dynamic as Statefull CGNAT,
> ou Static ACLs?)
>  - Using or not the measurements of DPDK Input/Drop/Fowarding.
>  - CPU Affinity done according to the demand of traffic
>  - SR-IOV (sharing resources) on DPDK.

It consumes 100% only if you busy poll (which is the default approach).
One can switch between polling and interrupts (or monitor, if supported),
or introduce halt instructions, in case of low/medium traffic volume.


-- 
Pawel Malachowski
@pawmal80

Re: DPDK and energy efficiency

>
> The way I saw, the questions induce the public to conclude that DPDK
> ALWAYS has 100% CPU usage, which is not true.


I don't concur.

Every research paper I've read indicates that, regardless of whether it has
packets to process or not, DPDK PMDs (poll-mode drivers) prevent the CPU
from falling into an LPI (low-power idle).

When it has no packets to process, the PMD runs the processor in a polling
loop that keeps utilization of the running core at 100%.

Cheers,

Etienne

On Mon, Feb 22, 2021 at 12:33 PM Douglas Fischer 
wrote:

> I'm very happy to see interest in DPDK and power consumption.
>
> But IMHO, the questions do not cover the actual reality of DPDK.
> That característic of "100% CPU" depends on several aspects, like:
>  - How old are the hardware on DPDK.
>  - What type of DPDK Instructions are made(Very Dynamic as
> Statefull CGNAT, ou Static ACLs?)
>  - Using or not the measurements of DPDK Input/Drop/Fowarding.
>  - CPU Affinity done according to the demand of traffic
>  - SR-IOV (sharing resources) on DPDK.
>
> The way I saw, the questions induce the public to conclude that DPDK
> ALWAYS has 100% CPU usage, which is not true.
>
>
> Em seg., 22 de fev. de 2021 às 04:30, Etienne-Victor Depasquale <
> ed...@ieee.org> escreveu:
>
>> Hello folks,
>>
>> I've just followed a thread regarding use of CGNAT and noted a suggestion
>> (regarding DANOS) that includes use of DPDK.
>>
>> As I'm interested in the breadth of adoption of DPDK, and as I'm a
>> researcher into energy and power efficiency, I'd love to hear your feedback
>> on your use of power consumption control by DPDK.
>>
>> I've drawn up a bare-bones, 2-question survey at this link:
>>
>> https://www.surveymonkey.com/r/J886DPY.
>>
>> Responses have been set to anonymous.
>>
>> Cheers,
>>
>> Etienne
>>
>> --
>> Ing. Etienne-Victor Depasquale
>> Assistant Lecturer
>> Department of Communications & Computer Engineering
>> Faculty of Information & Communication Technology
>> University of Malta
>> Web. https://www.um.edu.mt/profile/etiennedepasquale
>>
>
>
> --
> Douglas Fernando Fischer
> Engº de Controle e Automação
>


-- 
Ing. Etienne-Victor Depasquale
Assistant Lecturer
Department of Communications & Computer Engineering
Faculty of Information & Communication Technology
University of Malta
Web. https://www.um.edu.mt/profile/etiennedepasquale

Re: DPDK and energy efficiency