RE: Arista Routing Solutions

[Timothy Creswick]

> We're going to be getting some Arista gear soon and this issue came
> up. They made the same noises and vague overtures of "well, you
> *might* have problems with TAC if you go with 3rd party optics"...
> until I said "Oh really- well, that's a deal breaker, we can't really
> even consider that". And then they backpedaled at light speed and
> reassured me that 3rd party optics would be fine, they just "had to
> have the conversation".

Similar experience here, but the conversation went on far too long and 
ultimately lost Arista the deal. There was a ridiculous amount of insistence 
that we would have to carry a "stock of Arista optics", but every attempt to 
clarify exactly what that meant (how many, what they would cost etc) failed to 
get a straight answer.

It's 2016 and stupid conversations about vendor optics waste time and destroy 
deals.

The slight difference here is that pretty much the first thing we said to 
Arista was that transceivers were out of the question unless they could price 
them reasonably** (they chose not to). On this particular deal we were probably 
only talking about 500 SR 10G transceivers.

We've had similar conversations with Extreme, Brocade, Solarflare and Juniper, 
all of whom are quite happy with us running our own parts. Solarflare even 
certified our parts and put them on their website 
(http://solarflare.com/transceivers-and-cables).

> Also talked to a local Arista customer, much bigger than us and using
> a lot more of their gear. They have 0 Arista optics and 0 problems
> with 3rd party for a few years now. IMHO the whole thing was just
> sales guy FUD to try to squeeze a few extra bucks out.

Doesn't surprise me, and I'm sure if we'd pushed for another week we could have 
got to this position. Unfortunately for Arista, there was another vendor quite 
happy to get the deal done faster and without all the BS so we voted with our 
feet.

Clearly, mileage will vary on this one.

T



** in this context, "reasonably" means no more than _double_ what I currently 
buy at.

Re: Arista Routing Solutions

[Rob van de Logt]

On Thu, Apr 28, 2016 at 2:22 PM, Timothy Creswick
 wrote:
> Not in response to any point specifically, but the major issue which stopped 
> us buying Arista a few months ago was the rather out-dated attitude to 3rd 
> party transceiver support.
>
> I'm sure there are plenty of people running Arista on 3rd party optics, but 
> all the noises that were being made by the sales and technical guys suggested 
> that we could find ourselves abandoned by their support or a policy change in 
> the future.

We're going to be getting some Arista gear soon and this issue came
up. They made the same noises and vague overtures of "well, you
*might* have problems with TAC if you go with 3rd party optics"...
until I said "Oh really- well, that's a deal breaker, we can't really
even consider that". And then they backpedaled at light speed and
reassured me that 3rd party optics would be fine, they just "had to
have the conversation".

Also talked to a local Arista customer, much bigger than us and using
a lot more of their gear. They have 0 Arista optics and 0 problems
with 3rd party for a few years now. IMHO the whole thing was just
sales guy FUD to try to squeeze a few extra bucks out.

Re: Arista Routing Solutions

[Ryan Woolley]

On Thu, Apr 28, 2016 at 1:33 AM, lincoln dale  wrote:

> On Wed, Apr 27, 2016 at 4:41 PM, Peter Kranz 
> wrote:
>
>> Curious if you have any thoughts on the longevity of the 7500R
>> and 7280R survival's with IPv4 full tables? How full are you seeing the
>> TCAM getting today (I'm assuming they are doing some form of selective
>> download)? And if we are currently adding 100k/routes a year, how much
>> longer will it last?
>
> [...]
>
> One could ask Geoff Huston where he thinks combined IPv4+v6 will exceed 1M
> entries but I would expect it to be many years away based on
> http://bgp.potaroo.net/ and we'd welcome discussions about if it you want
> to know our opinion [*] on how we're doing it will scale.  What we're doing
> doesn't explode at 1M, there's headroom in it hence why we say "1M+". Again
> we're happy to talk about it, just ask your friendly arista person and if
> you don't know who to ask, ask me and i'll put you in touch with the right
> folks.
>

Peter, I'd point you to https://labs.apnic.net/?p=767 for more historical
detail and a table with some (recent) predictions.  The summary is that the
rate is mostly linear at around 10% per year and even 1MM routes lasts
quite comfortably beyond 5 years at the current growth rate.  I am not
particularly worried about the table growth rate (or Moore's law) changing
dramatically.

With respect to the utilization of the hardware, our setup is basically the
same as Lincoln's scenario #1 and so utilization looks about the same, on
both platforms.

RE: Arista Routing Solutions

[Timothy Creswick]

> Just wanted to interject, the port density of the Arista switches is quite
> impressive, especially considering the price point they're at.

Not in response to any point specifically, but the major issue which stopped us 
buying Arista a few months ago was the rather out-dated attitude to 3rd party 
transceiver support.

I'm sure there are plenty of people running Arista on 3rd party optics, but all 
the noises that were being made by the sales and technical guys suggested that 
we could find ourselves abandoned by their support or a policy change in the 
future. 

I don't fundamentally have an issue with vendor optics, except when they are 
excessively priced. One or two vendors will actually sell their 10Gbps optics 
at a price that's pretty hard to refuse, given that it's all supported. The 
same couldn't be said in this case.

Additionally, the insistence that we would have to buy a "small number" of 
Arista optics with each device for testing purposes gets old very quickly. 
Again, I could get on-board with this if it's just for troubleshooting, but not 
when these additional optics suddenly add 15% to the overall buy price of each 
switch. At that point, other vendors are firmly back on the table.

YMMV of course, I suspect especially if you're buying 100s of boxes.

T

Re: Arista Routing Solutions

[Patrick Cole]

Laszln,

Thu, Apr 28, 2016 at 12:47:45PM +, Laszlo Hanyecz wrote:

> On 2016-04-28 11:06, Alain Hebert wrote:
> >
> >  Well,
> >
> >  Once you eliminate the ~160k superfluous prefixes (last time I
> > checked)...  This is a none issue.
> >
> >  Some work on some sort summary function would keep those devices
> > alive...  but we all know there is more money to be made the faster the
> > device become obsolete :(
> >
> >
> 
> Can you explain how this works?  How can a router determine which prefix 
> is superfluous?  How does it cope when a suppressed prefix is withdrawn 
> or a more specific prefix is added? Is this just one of those 'it works 
> some of the time' solutions or is this something that can be done safely 
> with an appropriate algorithm?
 
A fair chunk of the routing table is aggregable.  If multiple
aggregable prefixes share the same nexthop, the HW entries can be 
summarised accordingly, reducing the HW resource footprint.

Should one of the smaller prefixes be withdrawn or best path
change to another nexthop, the control plane needs to be smart 
enough to adapt and reprogram the HW accordingly.   It is a fairly
logical and reasonable algorithm to construct

-- 
Patrick Cole 
Senior Network Specialist
World Without Wires
PO Box 869. Palm Beach, QLD, 4221
Ph:  0410 626 630

Re: Arista Routing Solutions

[Laszlo Hanyecz]

On 2016-04-28 11:06, Alain Hebert wrote:

 Well,

 Once you eliminate the ~160k superfluous prefixes (last time I
checked)...  This is a none issue.

 Some work on some sort summary function would keep those devices
alive...  but we all know there is more money to be made the faster the
device become obsolete :(




Can you explain how this works?  How can a router determine which prefix 
is superfluous?  How does it cope when a suppressed prefix is withdrawn 
or a more specific prefix is added? Is this just one of those 'it works 
some of the time' solutions or is this something that can be done safely 
with an appropriate algorithm?


Thanks,
Laszlo

Re: Arista Routing Solutions

[Alain Hebert]

Well,

Once you eliminate the ~160k superfluous prefixes (last time I
checked)...  This is a none issue.

Some work on some sort summary function would keep those devices
alive...  but we all know there is more money to be made the faster the
device become obsolete :(

-
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 04/28/16 01:33, lincoln dale wrote:
> On Wed, Apr 27, 2016 at 4:41 PM, Peter Kranz  wrote:
>
>> Curious if you have any thoughts on the longevity of the 7500R and
>> 7280R survival's with IPv4 full tables? How full are you seeing the TCAM
>> getting today (I'm assuming they are doing some form of selective
>> download)? And if we are currently adding 100k/routes a year, how much
>> longer will it last?
>>
> I can't speak for Ryan or Netflix, but we (Arista) are stating our
> technique is good for 1M+ prefixes of IPv4+v6 combined.  Internet right now
> is at between 575K and 635K IPv4 and between 28K and 35K IPv6 right now and
> its taken many many many years to get there, its foreseeable there's many
> years of growth there.
> Note that we don't do static partitioning between IPv4 and IPv6 and our how
> we do it has more headroom in it than we state, so we're confident.  We're
> also not doing "selective download", this is every prefix in current table.
>
> What I can share is two different scenarios today:
>
> 1. a traditional internet edge router with multiple transit/peer providers,
> Internet as of right now, and a cloud customer  that also has hundreds of
> thousands of prefixes internally
> Ryan's case might be different to others, but here are three scenarios
> deployed today: 1. a large hosting provider with full tables and many
> internal prefixes, 2. a cloud deployment.
>
> The former is at 854K IPv4 and 35K IPv6 of 'internet' as of a few weeks ago:
>
> 7500R# show ip route summary | grep Total
> Total Routes  575127
> 7500R# show ipv6 route summary | grep Total
>  Total Routes  35511
> 7500R# show hardware capacity | grep Routing
> Forwarding Resources Usage
>
> TableFeatureChip Used   Used  Free   Committed   Best
> Case   High
>   Entries(%)   Entries Entries
> Max  Watermark
>
>  Entries
>  -- -  -- - ---
> --- -
> Routing  Resource1  815   39% 1233   0
>  2048817
> Routing  Resource2  469   45%  555   0
>  1024471
> Routing  Resource314074   42%18694   0
> 32768  14098
> Routing  V4Routes696364   88%89753   0
>  786432 697110
> Routing  V6Routes 00%89753   0
>  786432  0
>
>
> The latter is at 854K IPv4 + 45K IPv6:
>
> 7500R# show ip route summary | grep Total
> Total Routes  854393
> 7500R# show ipv6 route summary | grep Total
>  Total Routes  45678
> 7500R# show hardware capacity | grep Routing
> Forwarding Resources Usage
>
> TableFeatureChip Used   Used  Free   Committed   Best
> Case   High
>   Entries(%)   Entries Entries
> Max  Watermark
>
>  Entries
>  -- -  -- - ---
> --- -
> Routing  Resource1   131964%   729   0
>  2048   1320
> Routing  Resource280979%   215   0
>  1024814
> Routing  Resource3  2410273%  8666   0
> 32768  24104
> Routing  V4Routes  64433683%124302   0
>  786432 644364
> Routing  V6Routes   1779212%124302   0
>  786432  17795
>
>
> One could ask Geoff Huston where he thinks combined IPv4+v6 will exceed 1M
> entries but I would expect it to be many years away based on
> http://bgp.potaroo.net/ and we'd welcome discussions about if it you want
> to know our opinion [*] on how we're doing it will scale.  What we're doing
> doesn't explode at 1M, there's headroom in it hence why we say "1M+". Again
> we're happy to talk about it, just ask your friendly arista person and if
> you don't know who to ask, ask me and i'll put you in touch with the right
> folks.
>
>
> cheers,
>
> lincoln.  [*] l...@arista.com
>

Re: Arista Routing Solutions

[lincoln dale]

On Wed, Apr 27, 2016 at 4:41 PM, Peter Kranz  wrote:

> Curious if you have any thoughts on the longevity of the 7500R and
> 7280R survival's with IPv4 full tables? How full are you seeing the TCAM
> getting today (I'm assuming they are doing some form of selective
> download)? And if we are currently adding 100k/routes a year, how much
> longer will it last?
>

I can't speak for Ryan or Netflix, but we (Arista) are stating our
technique is good for 1M+ prefixes of IPv4+v6 combined.  Internet right now
is at between 575K and 635K IPv4 and between 28K and 35K IPv6 right now and
its taken many many many years to get there, its foreseeable there's many
years of growth there.
Note that we don't do static partitioning between IPv4 and IPv6 and our how
we do it has more headroom in it than we state, so we're confident.  We're
also not doing "selective download", this is every prefix in current table.

What I can share is two different scenarios today:

1. a traditional internet edge router with multiple transit/peer providers,
Internet as of right now, and a cloud customer  that also has hundreds of
thousands of prefixes internally
Ryan's case might be different to others, but here are three scenarios
deployed today: 1. a large hosting provider with full tables and many
internal prefixes, 2. a cloud deployment.

The former is at 854K IPv4 and 35K IPv6 of 'internet' as of a few weeks ago:

7500R# show ip route summary | grep Total
Total Routes  575127
7500R# show ipv6 route summary | grep Total
 Total Routes  35511
7500R# show hardware capacity | grep Routing
Forwarding Resources Usage

TableFeatureChip Used   Used  Free   Committed   Best
Case   High
  Entries(%)   Entries Entries
Max  Watermark

 Entries
 -- -  -- - ---
--- -
Routing  Resource1  815   39% 1233   0
 2048817
Routing  Resource2  469   45%  555   0
 1024471
Routing  Resource314074   42%18694   0
32768  14098
Routing  V4Routes696364   88%89753   0
 786432 697110
Routing  V6Routes 00%89753   0
 786432  0


The latter is at 854K IPv4 + 45K IPv6:

7500R# show ip route summary | grep Total
Total Routes  854393
7500R# show ipv6 route summary | grep Total
 Total Routes  45678
7500R# show hardware capacity | grep Routing
Forwarding Resources Usage

TableFeatureChip Used   Used  Free   Committed   Best
Case   High
  Entries(%)   Entries Entries
Max  Watermark

 Entries
 -- -  -- - ---
--- -
Routing  Resource1   131964%   729   0
 2048   1320
Routing  Resource280979%   215   0
 1024814
Routing  Resource3  2410273%  8666   0
32768  24104
Routing  V4Routes  64433683%124302   0
 786432 644364
Routing  V6Routes   1779212%124302   0
 786432  17795


One could ask Geoff Huston where he thinks combined IPv4+v6 will exceed 1M
entries but I would expect it to be many years away based on
http://bgp.potaroo.net/ and we'd welcome discussions about if it you want
to know our opinion [*] on how we're doing it will scale.  What we're doing
doesn't explode at 1M, there's headroom in it hence why we say "1M+". Again
we're happy to talk about it, just ask your friendly arista person and if
you don't know who to ask, ask me and i'll put you in touch with the right
folks.


cheers,

lincoln.  [*] l...@arista.com

RE: Arista Routing Solutions

[Peter Kranz]

Ryan,
Curious if you have any thoughts on the longevity of the 7500R and 
7280R survival's with IPv4 full tables? How full are you seeing the TCAM 
getting today (I'm assuming they are doing some form of selective download)? 
And if we are currently adding 100k/routes a year, how much longer will it last?

-Peter

Re: Arista Routing Solutions

[Paras Jha]

Just wanted to interject, the port density of the Arista switches is quite
impressive, especially considering the price point they're at.

On Tue, Apr 26, 2016 at 12:46 PM, Ryan Woolley 
wrote:

> While the QFX in general is similar to Jericho-based platforms, I think the
> QFX10002 is perhaps not an ideal comparison.  At 100G, there is a
> significant density penalty on that platform, as you can use all 36 ports
> at 40G, but only 12 ports at 100G.
>
> BGP convergence in the newer EOS releases is indeed very, very fast.
>
> On Sun, Apr 24, 2016 at 12:08 PM, Colton Conor 
> wrote:
>
> > Saku,
> >
> > I guess you are right the QFX10002-36Q is probably a better comparison.
> But
> > let's be honest, Juniper is not going to sell a QFX10002-36Q for less
> than
> > $20k like Arista will do for a semi- similar box. Even with a high
> discount
> > (like 90 percent off list), the Juniper QFX10002-36Q at $360k list price
> > comes nowhere close on the price point. Cisco, Juniper, ALU, etc are all
> > not going to see a low cost high density fixed switch because that would
> > cannibalize on their sales on the larger platforms. I really think Arista
> > is kind of unique here as they don't have another routing platform to
> > cannibalize, so they are competitively pricing their platform.
> >
> > So I guess the question becomes, what features are missing that Arista
> does
> > not currently have? They seems to be adding more and more features, and
> > taking more market share. Here is a list of features supported:
> >
> https://www.arista.com/en/support/product-documentation/supported-features
> > I have not personally used Arista myself, but I like what I am seeing as
> > far as price point, company culture, and repruatation in the market
> place.
> > I know their switching is solid, but I am not sure about their routing.
> >
> > Arista claims to have much, much faster BGP convergence time than all the
> > other vendors.
> >
> >
> >
> >
> >
> > On Sat, Apr 23, 2016 at 1:20 PM, Saku Ytti  wrote:
> >
> > > On 23 April 2016 at 10:52, Tom Hill  wrote:
> > > > In broad strokes: for your money you're either getting port density,
> or
> > > > more features per port. The only difference here is that there's
> > > > suddenly more TCAM on the device, and I still don't see the above
> > > > changing too drastically.
> > >
> > > Yeah OP is comparing high touch chip (MX104) to low touch chip
> > > (Jericho) that is not fair comparison. And cost is what customer is
> > > willing to pay, regardless of sticker on the box. No one will pay
> > > significant mark-up for another sticker, I've never seen in RFP
> > > significant differences in comparable products.
> > >
> > > Fairer comparison would be QFX10k, instead of MX104. QFX10k is AFAIK
> > > only product in this segment which is not using Jericho. If this is
> > > competitive advantage or risk, jury is still out, I lean towards
> > > competitive advantage, mainly due to its memory design.
> > >
> > > --
> > >   ++ytti
> > >
> >
>



-- 
Regards,
Paras

President
ProTraf Solutions, LLC
Enterprise DDoS Mitigation

Re: Arista Routing Solutions

[Ryan Woolley]

While the QFX in general is similar to Jericho-based platforms, I think the
QFX10002 is perhaps not an ideal comparison.  At 100G, there is a
significant density penalty on that platform, as you can use all 36 ports
at 40G, but only 12 ports at 100G.

BGP convergence in the newer EOS releases is indeed very, very fast.

On Sun, Apr 24, 2016 at 12:08 PM, Colton Conor 
wrote:

> Saku,
>
> I guess you are right the QFX10002-36Q is probably a better comparison. But
> let's be honest, Juniper is not going to sell a QFX10002-36Q for less than
> $20k like Arista will do for a semi- similar box. Even with a high discount
> (like 90 percent off list), the Juniper QFX10002-36Q at $360k list price
> comes nowhere close on the price point. Cisco, Juniper, ALU, etc are all
> not going to see a low cost high density fixed switch because that would
> cannibalize on their sales on the larger platforms. I really think Arista
> is kind of unique here as they don't have another routing platform to
> cannibalize, so they are competitively pricing their platform.
>
> So I guess the question becomes, what features are missing that Arista does
> not currently have? They seems to be adding more and more features, and
> taking more market share. Here is a list of features supported:
> https://www.arista.com/en/support/product-documentation/supported-features
> I have not personally used Arista myself, but I like what I am seeing as
> far as price point, company culture, and repruatation in the market place.
> I know their switching is solid, but I am not sure about their routing.
>
> Arista claims to have much, much faster BGP convergence time than all the
> other vendors.
>
>
>
>
>
> On Sat, Apr 23, 2016 at 1:20 PM, Saku Ytti  wrote:
>
> > On 23 April 2016 at 10:52, Tom Hill  wrote:
> > > In broad strokes: for your money you're either getting port density, or
> > > more features per port. The only difference here is that there's
> > > suddenly more TCAM on the device, and I still don't see the above
> > > changing too drastically.
> >
> > Yeah OP is comparing high touch chip (MX104) to low touch chip
> > (Jericho) that is not fair comparison. And cost is what customer is
> > willing to pay, regardless of sticker on the box. No one will pay
> > significant mark-up for another sticker, I've never seen in RFP
> > significant differences in comparable products.
> >
> > Fairer comparison would be QFX10k, instead of MX104. QFX10k is AFAIK
> > only product in this segment which is not using Jericho. If this is
> > competitive advantage or risk, jury is still out, I lean towards
> > competitive advantage, mainly due to its memory design.
> >
> > --
> >   ++ytti
> >
>

Re: Arista Routing Solutions

[Ryan Woolley]

IOS-XR on ASR 9k and Junos on MX.

For our use case, there's no longer anything limiting as compared to those
platforms.  BGP policy is perhaps not as rich as you might be used to if
your experience is with the sort of routers traditionally marketed to
service providers, but I'm sure that will get better, and it's probably
irrelevant if your policy is fairly static.

You are correct that we do collect a lot of flow data, both via sflow and
Netflow.  We've been able to do everything we need with Arista's sflow
implementation.

On Mon, Apr 25, 2016 at 6:41 PM, Colton Conor 
wrote:

> Ryan,
>
> What routing platform were you coming from before? What features does
> Arista not have that you find limiting that the old platform did have?
>
> How does Astira's Sflow only compare to having Cisco Netflow or Juniper
> JFlow for traffic monitoring which I assume Netflix does alot of?
>
> On Wed, Apr 20, 2016 at 3:48 PM, Ryan Woolley 
> wrote:
>
>> Colton Conor wrote:
>> > I know Arista is typically a switch manufacturer, but with their
>> recently
>> > announced Arista 7500R Series and soon to be announced but already
>> shipping
>> > 7280R Series Arista is officially getting into the routing game. The
>> fixed
>> > 1U 7280R Series looks quite impressive. The 7500R series is your
>> > traditional chassis and line card based solution.
>> >
>> > Both of these products have the ability to hold the full internet
>> routing
>> > table, and Arista is working on MPLS features. Both of these new
>> products
>> > use the latest Broadcom Jerico chipsets.
>>
>> We (Netflix) have been deploying the previous gen (7500E) as edge routers
>> for about two years in high traffic, low route count applications in our
>> CDN, and have been working with Arista for almost as long to improve route
>> scale so that we could turn off all our traditional routers.
>>
>> The features that enable full routes on Jericho are running in our
>> production network today and we also have the 7500R and 7280R working with
>> full tables.
>>
>> I can't speak to MPLS, but for our use case (all L3, very high-density
>> 10/40/100G, BGP, IS-IS and light QoS), it's working well.
>>
>> So, yes, I'd say those two products are quite viable and competitive
>> options in the edge router space.
>>
>
>

Re: Arista Routing Solutions

[Colton Conor]

Ryan,

What routing platform were you coming from before? What features does
Arista not have that you find limiting that the old platform did have?

How does Astira's Sflow only compare to having Cisco Netflow or Juniper
JFlow for traffic monitoring which I assume Netflix does alot of?

On Wed, Apr 20, 2016 at 3:48 PM, Ryan Woolley 
wrote:

> Colton Conor wrote:
> > I know Arista is typically a switch manufacturer, but with their recently
> > announced Arista 7500R Series and soon to be announced but already
> shipping
> > 7280R Series Arista is officially getting into the routing game. The
> fixed
> > 1U 7280R Series looks quite impressive. The 7500R series is your
> > traditional chassis and line card based solution.
> >
> > Both of these products have the ability to hold the full internet routing
> > table, and Arista is working on MPLS features. Both of these new products
> > use the latest Broadcom Jerico chipsets.
>
> We (Netflix) have been deploying the previous gen (7500E) as edge routers
> for about two years in high traffic, low route count applications in our
> CDN, and have been working with Arista for almost as long to improve route
> scale so that we could turn off all our traditional routers.
>
> The features that enable full routes on Jericho are running in our
> production network today and we also have the 7500R and 7280R working with
> full tables.
>
> I can't speak to MPLS, but for our use case (all L3, very high-density
> 10/40/100G, BGP, IS-IS and light QoS), it's working well.
>
> So, yes, I'd say those two products are quite viable and competitive
> options in the edge router space.
>

Re: Arista Routing Solutions

[Thomas Penrose]

Hey Colton,

Comments inline:

On Wed, Apr 20, 2016 at 3:37 PM, Colton Conor 
wrote:

> NANOG,
>
> I know Arista is typically a switch manufacturer, but with their recently
> announced Arista 7500R Series and soon to be announced but already shipping
> 7280R Series Arista is officially getting into the routing game. The fixed
> 1U 7280R Series looks quite impressive. The 7500R series is your
> traditional chassis and line card based solution.
>

I must admit, i'm not usually excited by new hardware, but this
announcement did catch my eye!

Both of these products have the ability to hold the full internet routing
> table, and Arista is working on MPLS features. Both of these new products
> use the latest Broadcom Jerico chipsets.
>
> I would like to know how viable of a product NANOG thinks these Arista
> routers are compared to service provider grade routers from Cisco, Juniper,
> ALU, and Brocade?
>

Honestly? I think you need to look at what you actually need out of a box.
At the end of the day, its a 1U switch. If you want to terminate a GRT a
the edge of your network and do some basic path selection then it sounds
like it would be an amazing and cheap fit. On the other hand, I don't think
we can start throwing away core routers yet ;)

Cost wise, Arista seems to be much, much less per port. For example, the
> 1U Arista 7280R with 48x10GbE (SFP+) & 6x100GbE QSFP cost about the same as
> what Juniper sells a MX104 with only four 10G ports for (Under 20K).
>

I'm consistently amazed at the density they are achieving for the $$ and I
think it all comes down to what the actual application is here. Most basic
BGP networks do not need all the bells and whistles of the MX104 and will
really benefit from the extra port density. That being said, I wouldn't be
replacing core PoPs in large ISPs with 1U switches!


> Can the Arista EOS software combine with their hardware based on the
> Broadcom Jericho chipset truly compete  with the custom chipsets and
> accompanying software from the big guys?
>

I've used Arista for a while now (Moving from Cisco / Extreme) and I truly
believe that their software is excellent. They just seem to be doing it
'the right way', If you've not watched it, this video is worth a bit of
your time! https://www.youtube.com/watch?v=VdJZq4dRjf4

Thats my $0.02 anyway

Tom

Re: Arista Routing Solutions

[Ryan Woolley]

Colton Conor wrote:
> I know Arista is typically a switch manufacturer, but with their recently
> announced Arista 7500R Series and soon to be announced but already
shipping
> 7280R Series Arista is officially getting into the routing game. The fixed
> 1U 7280R Series looks quite impressive. The 7500R series is your
> traditional chassis and line card based solution.
>
> Both of these products have the ability to hold the full internet routing
> table, and Arista is working on MPLS features. Both of these new products
> use the latest Broadcom Jerico chipsets.

We (Netflix) have been deploying the previous gen (7500E) as edge routers
for about two years in high traffic, low route count applications in our
CDN, and have been working with Arista for almost as long to improve route
scale so that we could turn off all our traditional routers.

The features that enable full routes on Jericho are running in our
production network today and we also have the 7500R and 7280R working with
full tables.

I can't speak to MPLS, but for our use case (all L3, very high-density
10/40/100G, BGP, IS-IS and light QoS), it's working well.

So, yes, I'd say those two products are quite viable and competitive
options in the edge router space.

Re: Arista Routing Solutions

[lincoln dale]

>
> > High Touch / Low Touch
>
> High touch means very general purpose NPU, with off-chip memory. Low
> touch means usually ASIC or otherwise simplified pipeline and on-chip
> memory. Granted Jericho can support off-chip memory too.
>
> L3 switches are canonical example of low touch. EZchip, Trio, Solar,
> FP3 etc are examples of canonical high touch NPUs. What low touch can
> do, it can do fast and economically.
>

Your analogy makes some sense, but what you classify as high-touch /
low-touch is just one dimension and could do with a more modern update.

I'd suggest a more modern analogy would be that historically the difference
between a L3 switch and a router is the former has a fixed processing
pipeline, limited buffering (most are just on-chip buffer) and limited
table sizes.
But more modern packet processors with fixed pipelines often have blocks or
sections that are programmable or flexible. e.g. with a flexible packet
parser, its possible to support new overlay or tunnel mechanisms, flexible
key generation makes it possible to reuse different table resources in
different ways, flexible rewrite engine means egress encap or tunnels or
logic can be done.
There's also often more capacity for recirc or additional stages as
required.

Specific to Jericho, the underlying silicon has all these characteristics.
We [*] used the flexibility in all of the stages both now and in previous
iterations (Arad) to add new features/functionality that wasn't natively
there to start with. And it uses a combination of on-chip & off-chip
buffering with VoQ

Its also not only Arista that call it a router cisco do too (NCS5K5).

Sure, using a NPU for packet processing essentially provided a 100%
programmable packet forwarding pipeline, and maybe even a "run to
completion" kind of packet pipeline where the pipeline could have a long
tail of processing. However, engineering is a zero sum game, and to do that
means you sacrifice power or density, or most often, both.

I agree the lines have been blurred as to the characteristics, and we'd
openly state that its not going to be useful in every use case of where a
router is deployed, but for specific use cases, it fits the bill and has
compelling density, performance and cost dynamics.


To the OPs question, there are people running with this in EFT and others
in production.
My suggestion would be that if you think its of interest, reach out to your
friendly Arista person [*] and try it out or talk through what it is you're
after. We are generally a friendly bunch and often we can be quite creative
in enabling things in different ways to old.



> Yeah they are certainly much behind in features, but if you don't
> need those features, it's probably actually an advantage. For my
> use-cases Arista's MPLS stack is not there.


We've historically had the data-plane but not the control-plane. Thats a
work in progress.
Again, often there are creative solutions to ways of doing things that
aren't necessarily the same as old ways but achieve the same end result.


cheers,

lincoln.
[*] disclosure: i work on said products described l...@arista.com.

Re: Arista Routing Solutions

[Saku Ytti]

On 24 April 2016 at 09:08, Colton Conor  wrote:

Hey,

> I guess you are right the QFX10002-36Q is probably a better comparison. But
> let's be honest, Juniper is not going to sell a QFX10002-36Q for less than
> $20k like Arista will do for a semi- similar box. Even with a high discount
> (like 90 percent off list), the Juniper QFX10002-36Q at $360k list price
> comes nowhere close on the price point. Cisco, Juniper, ALU, etc are all not
> going to see a low cost high density fixed switch because that would
> cannibalize on their sales on the larger platforms. I really think Arista is
> kind of unique here as they don't have another routing platform to
> cannibalize, so they are competitively pricing their platform.

20k seems a stretch, that's like 94.5% discount, it's not unheard off.
If you have volume, I would imagine it being doable.

 > So I guess the question becomes, what features are missing that Arista does
> not currently have? They seems to be adding more and more features, and
> taking more market share. Here is a list of features supported:
> https://www.arista.com/en/support/product-documentation/supported-features I
> have not personally used Arista myself, but I like what I am seeing as far
> as price point, company culture, and repruatation in the market place. I
> know their switching is solid, but I am not sure about their routing.

Yeah they are ccertainly much behind in features, but if you don't
need those features, it's probably actually an advantage. For my
use-cases Arista's MPLS stack is not there.

> Arista claims to have much, much faster BGP convergence time than all the
> other vendors.

I wouldn't be surprised, but honestly the competition does not set the
bar high there.

-- 
  ++ytti

Re: Arista Routing Solutions

[Colton Conor]

Saku,

I guess you are right the QFX10002-36Q is probably a better comparison. But
let's be honest, Juniper is not going to sell a QFX10002-36Q for less than
$20k like Arista will do for a semi- similar box. Even with a high discount
(like 90 percent off list), the Juniper QFX10002-36Q at $360k list price
comes nowhere close on the price point. Cisco, Juniper, ALU, etc are all
not going to see a low cost high density fixed switch because that would
cannibalize on their sales on the larger platforms. I really think Arista
is kind of unique here as they don't have another routing platform to
cannibalize, so they are competitively pricing their platform.

So I guess the question becomes, what features are missing that Arista does
not currently have? They seems to be adding more and more features, and
taking more market share. Here is a list of features supported:
https://www.arista.com/en/support/product-documentation/supported-features
I have not personally used Arista myself, but I like what I am seeing as
far as price point, company culture, and repruatation in the market place.
I know their switching is solid, but I am not sure about their routing.

Arista claims to have much, much faster BGP convergence time than all the
other vendors.





On Sat, Apr 23, 2016 at 1:20 PM, Saku Ytti  wrote:

> On 23 April 2016 at 10:52, Tom Hill  wrote:
> > In broad strokes: for your money you're either getting port density, or
> > more features per port. The only difference here is that there's
> > suddenly more TCAM on the device, and I still don't see the above
> > changing too drastically.
>
> Yeah OP is comparing high touch chip (MX104) to low touch chip
> (Jericho) that is not fair comparison. And cost is what customer is
> willing to pay, regardless of sticker on the box. No one will pay
> significant mark-up for another sticker, I've never seen in RFP
> significant differences in comparable products.
>
> Fairer comparison would be QFX10k, instead of MX104. QFX10k is AFAIK
> only product in this segment which is not using Jericho. If this is
> competitive advantage or risk, jury is still out, I lean towards
> competitive advantage, mainly due to its memory design.
>
> --
>   ++ytti
>

RE: Arista Routing Solutions

[Keith Medcalf]

Got it, thanks for the explanation!

> -Original Message-
> From: Saku Ytti [mailto:s...@ytti.fi]
> Sent: Sunday, 24 April, 2016 11:03
> To: Keith Medcalf
> Cc: nanog list
> Subject: Re: Arista Routing Solutions
>
> On 24 April 2016 at 05:14, Keith Medcalf <kmedc...@dessus.com> wrote:
>
> > High Touch / Low Touch
>
> High touch means very general purpose NPU, with off-chip memory. Low
> touch means usually ASIC or otherwise simplified pipeline and on-chip
> memory. Granted Jericho can support off-chip memory too.
>
> L3 switches are canonical example of low touch. EZchip, Trio, Solar,
> FP3 etc are examples of canonical high touch NPUs. What low touch can
> do, it can do fast and economically.
>
> But like few terms, it's not exact, and borders are hazy and even
> subjective.
>
>
>
> --
>   ++ytti

Re: Arista Routing Solutions

[Saku Ytti]

On 24 April 2016 at 05:14, Keith Medcalf  wrote:

> High Touch / Low Touch

High touch means very general purpose NPU, with off-chip memory. Low
touch means usually ASIC or otherwise simplified pipeline and on-chip
memory. Granted Jericho can support off-chip memory too.

L3 switches are canonical example of low touch. EZchip, Trio, Solar,
FP3 etc are examples of canonical high touch NPUs. What low touch can
do, it can do fast and economically.

But like few terms, it's not exact, and borders are hazy and even subjective.



-- 
  ++ytti

RE: Arista Routing Solutions

[Keith Medcalf]

High Touch / Low Touch

Is this a measure of the amount of fiddle diddling required to get the chip to 
work as documented, or is it some other kind of code?

For example a "High Touch" chip needs lots of fiddle farting because it was 
designed by a moron and every possible thing that can be programmed incorrectly 
is programmed incorrectly, whereas in a "Low Touch" chip all the defaults are 
already set to the most useful and rational setting so that it can be used 
without touching it to fix all the defects?

Perhaps it is a measure of the babysitting required while the chip is running.  
"High Touch" chips require constant attention, nappy changes, positive 
re-inforcement of the settings, etc., while operating because they are 
inherently unreliable and badly designed whereas "Low Touch" chips once set up 
just work and require little ongoing supervision unless you want to change 
something?

Or is it just a strange translation for functionality (as in High End / Low 
End)?


> -Original Message-
> From: NANOG [mailto:nanog-boun...@nanog.org] On Behalf Of Saku Ytti
> Sent: Saturday, 23 April, 2016 14:21
> To: Tom Hill
> Cc: nanog list
> Subject: Re: Arista Routing Solutions
>
> On 23 April 2016 at 10:52, Tom Hill <t...@ninjabadger.net> wrote:
> > In broad strokes: for your money you're either getting port density, or
> > more features per port. The only difference here is that there's
> > suddenly more TCAM on the device, and I still don't see the above
> > changing too drastically.
>
> Yeah OP is comparing high touch chip (MX104) to low touch chip
> (Jericho) that is not fair comparison. And cost is what customer is
> willing to pay, regardless of sticker on the box. No one will pay
> significant mark-up for another sticker, I've never seen in RFP
> significant differences in comparable products.
>
> Fairer comparison would be QFX10k, instead of MX104. QFX10k is AFAIK
> only product in this segment which is not using Jericho. If this is
> competitive advantage or risk, jury is still out, I lean towards
> competitive advantage, mainly due to its memory design.
>
> --
>   ++ytti

Re: Arista Routing Solutions

[Jeff Tantsura]

Saku,

Jericho is in no sense a low end chip, while there are some scale limitations 
(what can be done with SuperFEC, some bridging related stuff), from 
functionality prospective it is a very capable silicon.

One has to:
Understand how to program it properly (recursiveness, ECMP’s, etc) 
Know how to enhance SDK
Have a rather rich control plane, which can be translated into rich forwarding 
functionality :-)

I’m not familiar with Arista’s feature set
NCS with XR would be a good proof 

Watch for Jericho updates from DNX

Cheers,
Jeff 



On 4/23/16, 11:20 AM, "NANOG on behalf of Saku Ytti"  wrote:

>On 23 April 2016 at 10:52, Tom Hill  wrote:
>> In broad strokes: for your money you're either getting port density, or
>> more features per port. The only difference here is that there's
>> suddenly more TCAM on the device, and I still don't see the above
>> changing too drastically.
>
>Yeah OP is comparing high touch chip (MX104) to low touch chip
>(Jericho) that is not fair comparison. And cost is what customer is
>willing to pay, regardless of sticker on the box. No one will pay
>significant mark-up for another sticker, I've never seen in RFP
>significant differences in comparable products.
>
>Fairer comparison would be QFX10k, instead of MX104. QFX10k is AFAIK
>only product in this segment which is not using Jericho. If this is
>competitive advantage or risk, jury is still out, I lean towards
>competitive advantage, mainly due to its memory design.
>
>-- 
>  ++ytti

Re: Arista Routing Solutions

[Saku Ytti]

On 23 April 2016 at 10:52, Tom Hill  wrote:
> In broad strokes: for your money you're either getting port density, or
> more features per port. The only difference here is that there's
> suddenly more TCAM on the device, and I still don't see the above
> changing too drastically.

Yeah OP is comparing high touch chip (MX104) to low touch chip
(Jericho) that is not fair comparison. And cost is what customer is
willing to pay, regardless of sticker on the box. No one will pay
significant mark-up for another sticker, I've never seen in RFP
significant differences in comparable products.

Fairer comparison would be QFX10k, instead of MX104. QFX10k is AFAIK
only product in this segment which is not using Jericho. If this is
competitive advantage or risk, jury is still out, I lean towards
competitive advantage, mainly due to its memory design.

-- 
  ++ytti

Re: Arista Routing Solutions

[Tom Hill]

On 20/04/16 15:37, Colton Conor wrote:
> Can the Arista EOS software combine with their hardware based on the
> Broadcom Jericho chipset truly compete  with the custom chipsets and
> accompanying software from the big guys?

In broad strokes: for your money you're either getting port density, or
more features per port. The only difference here is that there's
suddenly more TCAM on the device, and I still don't see the above
changing too drastically.

If it works* for you, use it. :)


* Assuming that you've done your due diligence before purchasing, and
not just skim-read the vendor PDF.

-- 
Tom