ISP-CACHING Digest for Wednesday, August 02, 2000.
1. Re: Caching Products and Skycaching
2. Re: isp-caching digest: August 01, 2000
3. Re: Caching Products and Skycaching
4. Re: Caching Products and Skycaching
5. Re: Caching Products and Skycaching
6. Re: Caching Products and Skycaching
7. Re: Caching Products and Skycaching
8. Re: Caching Products and Skycaching
9. RE: Caching Products and Skycaching
10. RE: Callwave problems w/ cache
11. Wireless ISP - complex question
12. Re: Caching Products and Skycaching
13. Re: Caching Products and Skycaching
14. Re: Caching Products and Skycaching
15. Re: Caching Products and Skycaching
16. Re: Caching Products and Skycaching
17. Re: Caching Products and Skycaching
18. Re: Caching Products and Skycaching
19. Re: Caching Products and Skycaching
20. RE: Caching Products and Skycaching
21. Re: Caching Products and Skycaching
22. Re: Caching Products and Skycaching
23. caching seminars?
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: Joe Cooper <[EMAIL PROTECTED]>
Date: Wed, 02 Aug 2000 00:46:12 -0500
X-Message-Number: 1
> Subject: Re: Caching Products and Skycaching
> From: "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>
> Date: Wed, 2 Aug 2000 10:40:42 +0800 (PHT)
> X-Message-Number: 7
>
> Hi,
>
> I think cacheflow will fit your need. Based on my experience, it can make
> a big difference and saving in BW. It can make a 58% localhit compared to
> squid with around 40% localhit. They are almost equal in bytes
> transferred.
Cacheflow's statistics regarding hit rate are known to be quite suspect.
If the relatively new HTML parsing pre-fetch (wherein the box prefetchs
objects based on them probably being needed soon by the browser) is in
use, it seems Cacheflow counts those pre-fetched objects as cache hits,
even though they only came into the cache seconds before. This is not
representative of bandwidth savings. Others on this list have expressed
concerns over there being other obfuscations of fact going on in
Cacheflow generated statistics.
The simple fact is that there are definite hard limits on how much
bandwidth can be saved, given a certain amount of storage in a
cache...and for bandwidth savings some types of pre-fetching can be a
lose--though non-peak pre-fetch for 'freshness' can be a valid way to
improve bandwidth savings. (Other limits are present as well...like how
long objects can be saved, and how good the replacement policy is, and
whether the cache is overloaded because it is too small for the
network.)
But an ISP will not save 58% of their bandwidth no matter how much their
cache costs. Real ISP hit rates from other caching vendors (whose
numbers are a little more straightforward and whose numbers aren't
juggled so much to look good) are usually somewhere in the 35% to 45%
range. Those are good hit ratios and should be possible given enough
cache storage space from any quality caching product.
That's not to say that Cacheflow doesn't make nice boxes. Quite pricey,
but neat to look at, and they do their job pretty well from what I've
heard. Though, I do have to wonder at their refusal to take part in the
cacheoff events hosted by the Measurement Factory. One must assume they
aren't proud of their performance as it compares to other vendors. If I
were shopping for caches, I don't think it's what I'd choose...but then,
I'm biased.
--
Joe Cooper <[EMAIL PROTECTED]>
Affordable Web Caching Proxy Appliances
http://www.swelltech.com
----------------------------------------------------------------------
Subject: Re: isp-caching digest: August 01, 2000
From: Joe Cooper <[EMAIL PROTECTED]>
Date: Wed, 02 Aug 2000 00:55:45 -0500
X-Message-Number: 2
> Subject: Caching Products and Skycaching
> From: "Tim Maveal" <[EMAIL PROTECTED]>
> Date: Tue, 1 Aug 2000 19:08:08 -0700
> X-Message-Number: 5
>
> We are a Regional ISP and thinking of a caching solution. We are looking at
> Squid, CacheFlow and Imimic. We want to achieve a 45% reduction in our BW.
> What are your views? We have about 25,000 subscribers and feel there is
> going to be a dialup markup in the future and want to provide a better
> dialup ISP. We want to to know if Skycaching with any of these appliances
> can help us. Of course, the alternative is redundant Backbones. Any
> ideas?
If you opt for Squid, you may want to check out my tuning instructions
at:
http://www.swelltech.com/pengies/joe
There is a text file linked from there that gives some pretty complete
instructions for making Squid into a real screamer that competes pretty
well with the proprietary vendors. If it presents any problems for you,
feel free to write me with your questions, I'll be happy to help you get
Squid running at your site.
There are also some new Linux 2.2.16 kernel RPMS and SRPMS tuned for
Squid, plus some Squid RPMS (the Squids need an update as the
configuration in them was an experimental config that has to be fiddled
with a lot to work well).
Let me know if you do decide to build your own caches, as there are some
new optimizations that I've implemented in our boxes that I haven't
gotten around to documenting yet that makes Squid more solid and
effective for satellite links. Probably would help in your case.
Hope this helps. BTW- I think Squid is usually the most cost effective
solution. It works well, and with tuning can compete with the
proprietary vendors quite well. And if support is a concern, there are
several Squid OEM's that can provide it (Swell being one of them).
--
Joe Cooper <[EMAIL PROTECTED]>
Affordable Web Caching Proxy Appliances
http://www.swelltech.com
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>
Date: Wed, 2 Aug 2000 14:14:45 +0800 (PHT)
X-Message-Number: 3
Maxi Bernales
Systems Engineer
Pacific Internet=20
637-1700
"A razor may be sharper than an ax, but it cannot cut wood."=20
On Wed, 2 Aug 2000, Joe Cooper wrote:
> > Subject: Re: Caching Products and Skycaching
> > From: "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>
> > Date: Wed, 2 Aug 2000 10:40:42 +0800 (PHT)
> > X-Message-Number: 7
> >=20
> > Hi,
> >=20
> > I think cacheflow will fit your need. Based on my experience, it can ma=
ke
> > a big difference and saving in BW. It can make a 58% localhit compared=
to
> > squid with around 40% localhit. They are almost equal in bytes
> > transferred.
>=20
> Cacheflow's statistics regarding hit rate are known to be quite suspect.
>=20
> If the relatively new HTML parsing pre-fetch (wherein the box prefetchs
> objects based on them probably being needed soon by the browser) is in
> use, it seems Cacheflow counts those pre-fetched objects as cache hits,
> even though they only came into the cache seconds before. This is not
> representative of bandwidth savings. Others on this list have expressed
> concerns over there being other obfuscations of fact going on in
> Cacheflow generated statistics.
>=20
> The simple fact is that there are definite hard limits on how much
> bandwidth can be saved, given a certain amount of storage in a
> cache...and for bandwidth savings some types of pre-fetching can be a
> lose--though non-peak pre-fetch for 'freshness' can be a valid way to
> improve bandwidth savings. (Other limits are present as well...like how
> long objects can be saved, and how good the replacement policy is, and
> whether the cache is overloaded because it is too small for the
> network.) =20
got your point here! =3D) I think I have to verify your statements with
cacheflow.
>=20
> But an ISP will not save 58% of their bandwidth no matter how much their
> cache costs. Real ISP hit rates from other caching vendors (whose
> numbers are a little more straightforward and whose numbers aren't
> juggled so much to look good) are usually somewhere in the 35% to 45%
> range. Those are good hit ratios and should be possible given enough
> cache storage space from any quality caching product.
Yeah I know that. What I am trying to say is not saving 58% of the BW. It
is a fairytale story. I am pertaining to the local hitrate compared to
squid. I based my computations based on my own daily activities.
>=20
> That's not to say that Cacheflow doesn't make nice boxes. Quite pricey,
> but neat to look at, and they do their job pretty well from what I've
> heard. Though, I do have to wonder at their refusal to take part in the
> cacheoff events hosted by the Measurement Factory. One must assume they
> aren't proud of their performance as it compares to other vendors. If I
> were shopping for caches, I don't think it's what I'd choose...but then,
> I'm biased.
> --
> Joe Cooper <[EMAIL PROTECTED]>
> Affordable Web Caching Proxy Appliances
> http://www.swelltech.com
>=20
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> NEWS YOU CAN USE
> Check Out http://www.siliconvalley.internet.com
> For the Latest Internet News From Silicon Valley
>=20
> ____________ =95 The ISP-CACHING Discussion List =95 ___________
> To Join: mailto:[EMAIL PROTECTED]
> To Remove: mailto:[EMAIL PROTECTED]
> Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
>=20
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: "Matthew Watkins" <[EMAIL PROTECTED]>
Date: Wed, 2 Aug 2000 10:36:15 +0100
X-Message-Number: 4
> How do you measure the hit rate? Based on the Java applet GUI?
> Personally I find the object hit rate is high, but byte hit rate is less
so.
> Cacheflow's statistics regarding hit rate are known to be quite suspect.
> I have my own sets of program to have the figures based on the CF logs.
If you don't trust the statistics provided by the Java GUI, you will
probably want to analyse the log file data instead. Enable hit logging in
squid compatible format, then use a publicly available tool such as
Calamaris to generate reports. You can automate the process by setting
Cacheflow devices to regularly post logs off the cache to an FTP services
running on a remote Unix system. A few lines in Cron, set Calamaris to place
the output files into the directory of a webserver, and you've got the
system up and running...
http://cord.de/tools/squid/calamaris/
You should be able to compare the result to those provided by the Cacheflow
interface. It is possible that the data reported by the interface
occasionally changes between software versions, as the dynamic refresh
algorithms, object deletion policies etc. will occasionally get tweaked.
All this is a useful exercise in emancipation, but I imagine the best place
to look for your bandwidth saving would really be the router interfaces,
sometime after deployment when the cache has had the opportunity to fill.
Tools like MRTG make graphing this stuff trivial....
http://ee-staff.ethz.ch/~oetiker/webtools/mrtg/mrtg.html
In every Cacheflow deployment I've seen, we've managed a better hit-rate
than squid caches. How much better varies per customer, and the
manufacturer-generated statistics will always be open to some debate.
However, you should expect a commercial product to produce better
performance.
- Matt
Matthew Watkins
Technical Consultant
Knowware UK Ltd
Mobile: 07968 755807
Fax: 01223 301280
Email: [EMAIL PROTECTED]
http://www.knowware.co.uk
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: [EMAIL PROTECTED]
Date: Wed, 02 Aug 2000 18:12:24 +0800
X-Message-Number: 5
HI Matt,
I agree with you. I based my comments here on my experience using squid and
cacheflow.
Have you read www.swelltech.com? They are using squid and tweaked
something and benchmarked it using polygraph. And the results are
satisfactory! I wonder if it can support streaming video and real player
and still perform well.
At 10:36 AM 8/2/00 +0100, you wrote:
>> How do you measure the hit rate? Based on the Java applet GUI?
>> Personally I find the object hit rate is high, but byte hit rate is less
>so.
>
>> Cacheflow's statistics regarding hit rate are known to be quite suspect.
>
>> I have my own sets of program to have the figures based on the CF logs.
>
>If you don't trust the statistics provided by the Java GUI, you will
>probably want to analyse the log file data instead. Enable hit logging in
>squid compatible format, then use a publicly available tool such as
>Calamaris to generate reports. You can automate the process by setting
>Cacheflow devices to regularly post logs off the cache to an FTP services
>running on a remote Unix system. A few lines in Cron, set Calamaris to=
place
>the output files into the directory of a webserver, and you've got the
>system up and running...
>
>http://cord.de/tools/squid/calamaris/
>
>You should be able to compare the result to those provided by the Cacheflow
>interface. It is possible that the data reported by the interface
>occasionally changes between software versions, as the dynamic refresh
>algorithms, object deletion policies etc. will occasionally get tweaked.
>
>All this is a useful exercise in emancipation, but I imagine the best place
>to look for your bandwidth saving would really be the router interfaces,
>sometime after deployment when the cache has had the opportunity to fill.
>Tools like MRTG make graphing this stuff trivial....
>
>http://ee-staff.ethz.ch/~oetiker/webtools/mrtg/mrtg.html
>
>In every Cacheflow deployment I've seen, we've managed a better hit-rate
>than squid caches. How much better varies per customer, and the
>manufacturer-generated statistics will always be open to some debate.
>However, you should expect a commercial product to produce better
>performance.
>
>- Matt
>
>Matthew Watkins
>Technical Consultant
>Knowware UK Ltd
>
>Mobile: 07968 755807
>Fax: 01223 301280
>Email: [EMAIL PROTECTED]
>http://www.knowware.co.uk
>
>
>
>^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> NEWS YOU CAN USE
> Check Out http://www.siliconvalley.internet.com
> For the Latest Internet News From Silicon Valley
>
>____________ =95 The ISP-CACHING Discussion List =95 ___________
>To Join: mailto:[EMAIL PROTECTED]
>To Remove: mailto:[EMAIL PROTECTED]
>Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
>
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: "Chris Riegel" <[EMAIL PROTECTED]>
Date: Wed, 02 Aug 2000 09:15:46 -0400
X-Message-Number: 6
>In every Cacheflow deployment I've seen, we've managed a better hit-rate
>than squid caches. How much better varies per customer, and the
>manufacturer-generated statistics will always be open to some debate.
>However, you should expect a commercial product to produce better
>performance.
Why should you "expect" a commercial product to produce better performance?=
??
Just because it is a commercial product does not mean that it is naturally =
better...
In fact, I would wager that Squid has a 10 to 1 advantage in the number of =
people developing and enhancing the product worldwide when compared to =
Cacheflow....never underestimate the power of open source.
Again, lets not put Cacheflow marketing skills above Cacheflow technical =
skills....
Also, how come the only manufacturer I see who's stats are frequently =
questioned on this list are those of the Great Oz (Cacheflow)??
I can't remember the last time anybody asked about Netapp or Infolibria's =
statistics.....
Chris Riegel
Stratacache
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: "Matthew Watkins" <[EMAIL PROTECTED]>
Date: Wed, 2 Aug 2000 14:50:03 +0100
X-Message-Number: 7
> Why should you "expect" a commercial product to produce better
performance???
> Just because it is a commercial product does not mean that it is naturally
better...
> In fact, I would wager that Squid has a 10 to 1 advantage in the number of
people developing
> and enhancing the product worldwide when compared to Cacheflow....never >
underestimate the power of open source.
> Again, lets not put Cacheflow marketing skills above Cacheflow technical
skills....
Chris, regarding your comments above...
>From what I've seen from this list, you advocate performance as one of the
most important factors in a purchasing decision. I know I'm not the only
person that has reason to question your motives. Perhaps this is because you
optimise your product to perform exceptionally in Polymix, rather than face
the challenges of caching in the real world. I'd like to one of your boxes
submitted to the bakeoffs performing in the real world. You can happily
tweak TCP/IP parameters and object deletion policies for polymix, but does
it really work in the wild?
As a company that produces a commercial product, you already know that your
product has to offer performance comparable to or exceeding that of squid,
or people won't buy it. In many ways, Squid is the software benchmark that
many commercial cache vendors would accept they have to beat.
Perhaps you should re-read some of the list archives, to remind yourself
what you've said in the past. Your personal crusade against Cacheflow seems
to be obstructing real discussion. The message below was one you posted to
this list back in May this year, if I'm not mistaken.
- Matt
***
>I also rather enjoy the Squid/ICS battle that rages at each bakeoff.
>;-) In case you haven't been following Squid development, the Novell
>guys are really going to have to stay on their toes to keep their
>current speed lead.
Challenge Accepted.
>There are massive improvements going on to Squid's
>FS and replacement policy layers, among other interesting improvements.
I have been watching....and you guys are defiantly headed in the right
direction.
>I suspect that in a bakeoff not too distant we will see a Squid box
>(with the Swell logo of course!) that matches or beats a hardware
>comparable ICS box.
Maybe....keep in mind that we will push ICS to new levels as well...moving
targets are always harder to hit.
>Even now, I think our price performance is right on target, simply
>because we haven't got the expense of an ICS license, so the small
>amount of extra hardware needed to bring the performance up to match ICS
>can be had for about the cost of the license.
Your price performance is decent, but with the new low end ICS boxes we have
and the "super high end" clusters we have in the lab, we see pushing the
envelope on both sides of the performance spectrum.
>> One point to correct however is the misconception that ICS is only a
"high =
>> end" technology and that there is a big expense associated with it. Our
=
>> Stratacache Express E-55 (200TPS+ in Polymix-2 at 45% hit) is $ 3,395 =
>> list, an Express E-50 (150-170TPS in PM2 at 45% hit is $ 2,795).
>
>> At that ratio of performance to cost, we have a pretty strong comparison
=
>> to Squid at the low end.
>Out of curiosity, do you guys produce any IDE based boxes? Is that what
>the E-50 uses? I believe the only IDE based ICS box at the bakeoff was
>the Microbits box. And they were pushing their box so hard to get a
>high request rate, it's difficult to tell exactly what kind of
>performance they would get with reasonable hit rates and response
>times. I'd be curious to see how squid fares when compared to ICS boxes
>running on IDE drives.
The Stratacache E-50 is an IDE based box. The E-55 is SCSI Based.
>I feel like I should interject some angry wrestler style banter here to
>keep this from turning into a total love fest between competitors! (My
>box is gonna eat your box. Nanner nanner nanner.)
<In the voice of WWF Star the Rock>
You've got to smell what the guys at Stratacache have got cookin'
At the next Bakeoff...ICS will be opening a can of 'whoop ass all over you
Squid boys....watch out for folding chairs..
15 Vendors in the Cage Match Bakeoff of The Century.....All caches go in,
but only one comes out with the title...
Hey Alex...maybe you guys could sell the rights to the next Bakeoff on
Pay-per-view....
Cheers
C. Riegel
Stratacache
***
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: Solomon Sokolovsky <[EMAIL PROTECTED]>
Date: Wed, 02 Aug 2000 23:48:10 +1000
X-Message-Number: 8
Hi All,
First of all, Chris and Mathew, as you both know I follow both your comments and
I must say that both of you contradict each other at times, but would have to
say one thing.......... How the hell did we get to the subject of Cacheflow from
a question regarding caching servers and skycaching????
Of the subject, I am keen if somebody could help me out with some benchmark
results of NetApps 760 vs Cacheflow vs Inktomi. Chris, I would ask for ICS but
due to no media streaming capability, sorry!!
Thanks and hope somebody has some info for me and maybe some feedback from
people using Inktomi would be wonderful!
Thanks
Solomon
Matthew Watkins wrote:
> > Why should you "expect" a commercial product to produce better
> performance???
> > Just because it is a commercial product does not mean that it is naturally
> better...
> > In fact, I would wager that Squid has a 10 to 1 advantage in the number of
> people developing
> > and enhancing the product worldwide when compared to Cacheflow....never >
> underestimate the power of open source.
> > Again, lets not put Cacheflow marketing skills above Cacheflow technical
> skills....
>
> Chris, regarding your comments above...
>
> >From what I've seen from this list, you advocate performance as one of the
> most important factors in a purchasing decision. I know I'm not the only
> person that has reason to question your motives. Perhaps this is because you
> optimise your product to perform exceptionally in Polymix, rather than face
> the challenges of caching in the real world. I'd like to one of your boxes
> submitted to the bakeoffs performing in the real world. You can happily
> tweak TCP/IP parameters and object deletion policies for polymix, but does
> it really work in the wild?
>
> As a company that produces a commercial product, you already know that your
> product has to offer performance comparable to or exceeding that of squid,
> or people won't buy it. In many ways, Squid is the software benchmark that
> many commercial cache vendors would accept they have to beat.
>
> Perhaps you should re-read some of the list archives, to remind yourself
> what you've said in the past. Your personal crusade against Cacheflow seems
> to be obstructing real discussion. The message below was one you posted to
> this list back in May this year, if I'm not mistaken.
>
> - Matt
>
> ***
>
> >I also rather enjoy the Squid/ICS battle that rages at each bakeoff.
> >;-) In case you haven't been following Squid development, the Novell
> >guys are really going to have to stay on their toes to keep their
> >current speed lead.
>
> Challenge Accepted.
>
> >There are massive improvements going on to Squid's
> >FS and replacement policy layers, among other interesting improvements.
>
> I have been watching....and you guys are defiantly headed in the right
> direction.
>
> >I suspect that in a bakeoff not too distant we will see a Squid box
> >(with the Swell logo of course!) that matches or beats a hardware
> >comparable ICS box.
>
> Maybe....keep in mind that we will push ICS to new levels as well...moving
> targets are always harder to hit.
>
> >Even now, I think our price performance is right on target, simply
> >because we haven't got the expense of an ICS license, so the small
> >amount of extra hardware needed to bring the performance up to match ICS
> >can be had for about the cost of the license.
>
> Your price performance is decent, but with the new low end ICS boxes we have
> and the "super high end" clusters we have in the lab, we see pushing the
> envelope on both sides of the performance spectrum.
>
> >> One point to correct however is the misconception that ICS is only a
> "high =
> >> end" technology and that there is a big expense associated with it. Our
> =
> >> Stratacache Express E-55 (200TPS+ in Polymix-2 at 45% hit) is $ 3,395 =
> >> list, an Express E-50 (150-170TPS in PM2 at 45% hit is $ 2,795).
> >
> >> At that ratio of performance to cost, we have a pretty strong comparison
> =
> >> to Squid at the low end.
>
> >Out of curiosity, do you guys produce any IDE based boxes? Is that what
> >the E-50 uses? I believe the only IDE based ICS box at the bakeoff was
> >the Microbits box. And they were pushing their box so hard to get a
> >high request rate, it's difficult to tell exactly what kind of
> >performance they would get with reasonable hit rates and response
> >times. I'd be curious to see how squid fares when compared to ICS boxes
> >running on IDE drives.
>
> The Stratacache E-50 is an IDE based box. The E-55 is SCSI Based.
>
> >I feel like I should interject some angry wrestler style banter here to
> >keep this from turning into a total love fest between competitors! (My
> >box is gonna eat your box. Nanner nanner nanner.)
>
> <In the voice of WWF Star the Rock>
> You've got to smell what the guys at Stratacache have got cookin'
> At the next Bakeoff...ICS will be opening a can of 'whoop ass all over you
> Squid boys....watch out for folding chairs..
> 15 Vendors in the Cage Match Bakeoff of The Century.....All caches go in,
> but only one comes out with the title...
>
> Hey Alex...maybe you guys could sell the rights to the next Bakeoff on
> Pay-per-view....
>
> Cheers
>
> C. Riegel
> Stratacache
>
> ***
>
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> NEWS YOU CAN USE
> Check Out http://www.siliconvalley.internet.com
> For the Latest Internet News From Silicon Valley
>
> ____________ • The ISP-CACHING Discussion List • ___________
> To Join: mailto:[EMAIL PROTECTED]
> To Remove: mailto:[EMAIL PROTECTED]
> Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
----------------------------------------------------------------------
Subject: RE: Caching Products and Skycaching
From: [EMAIL PROTECTED]
Date: Wed, 2 Aug 2000 16:44:48 +0200
X-Message-Number: 9
Hello,
First of all, i am new here. So let me say that my name is Henno Kleijwegt
from the Netherlands and working for TELiNDUS. Snd, why isnt InfoLibria part
of your list?
Kind regards
Henno
-----Original Message-----
From: Solomon Sokolovsky [mailto:[EMAIL PROTECTED]]
Sent: woensdag 2 augustus 2000 15:48
To: [EMAIL PROTECTED]
Subject: Re: Caching Products and Skycaching
Hi All,
First of all, Chris and Mathew, as you both know I follow both your comments
and
I must say that both of you contradict each other at times, but would have
to
say one thing.......... How the hell did we get to the subject of Cacheflow
from
a question regarding caching servers and skycaching????
Of the subject, I am keen if somebody could help me out with some benchmark
results of NetApps 760 vs Cacheflow vs Inktomi. Chris, I would ask for ICS
but
due to no media streaming capability, sorry!!
Thanks and hope somebody has some info for me and maybe some feedback from
people using Inktomi would be wonderful!
Thanks
Solomon
----------------------------------------------------------------------
Subject: RE: Callwave problems w/ cache
From: "Rafferty Lengle" <[EMAIL PROTECTED]>
Date: Wed, 2 Aug 2000 08:14:19 -0700
X-Message-Number: 10
Dale,
I do not have any answers for this problem but I wanted to see if your ISP
offers Filtered Internet Access? My name is Rafferty and I work for N2H2,
Inc. and we provide such solutions to ISP,s, Enterprise and Education. I am
on this listserve because I manage our relationship with Compaq and their
caching appliance (TaskSmart C-Series, ICS). Our filtering technology is
fully integrated into the ICS software. Anyhow, ISP's are quickly adopting
filtering services so I just wanted to take this opportunity to introduce
myself and N2H2. If you have any questions please feel free to give me a
call or email.
Regards,
Rafferty Lengle
Strategic Alliance Manager/Compaq
N2H2, Inc.
206.336.2968
877.560.3124
[EMAIL PROTECTED]
-----Original Message-----
From: Dale [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, August 01, 2000 4:44 AM
To: [EMAIL PROTECTED]
Subject: Callwave problems w/ cache
Hi,
We have recently implemented the Cidera Cache Turbocharging system at
our ISP and are quite please with it for the most part. There are some
gaming sites such as zone.com and some call answering services such as
callwave.com and pagoo.com that have caused problems for our
customers. We have found it very difficult to get timely info from the
tech folks at these places so we can exclude the IPs that don't play
well with the cache. We finally had Callwave working for our customers
by excluding a bunch of IPs from caching, only to have it start
causing problems again recently. Apparently they made some changes to
their network, but we can't get anyone from Callwave with clue to
respond. In the meantime their techs are asking our customers to tell
us to remove our firewall. We are waiting on word from pagoo.com as
well.
Any words of encouragement or techniques that can help us quickly get
around issues such as this. Thanks.
Later,
Dale
Megalink Tech Support
____________ • The ISP-CACHING Discussion List • ___________
To Join: mailto:[EMAIL PROTECTED]
To Remove: mailto:[EMAIL PROTECTED]
Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
----------------------------------------------------------------------
Subject: Wireless ISP - complex question
From: "Bogdan Florin" <[EMAIL PROTECTED]>
Date: Wed, 02 Aug 2000 18:12:38 +0300
X-Message-Number: 11
Dear Sirs,
I'm going to start a Wireless ISP and since hardware has been already
chosen I'm a bit concernd about the software which will make this
Network powefull and flexible at the same time.
The users will be connected to linux servers and all linux servers will
be connected together. We expect that there will be around 2000 users
and 50 servers alltogether.
I'm writing to you because I would like some facilities for the
administration of the network and I will also need some advices.
A. About the service
I would like:
1 to be abble to provide for every user different speed inside
the LAN and on the INTERNET. ATTENTION ! some of the users will be
MOBILE so I need the specification for every user to be propagated over
the entire LINUX servers.
2 LINUX servers to be configured in such a way that its
replacement with another one (that I already have standing-by
(prepared)) will take as little time as possible
3 to have a powerfull PROXY CACHE SERVER working on LINUX OS
that is abble to display detail statisctics about the traffic,
bandwith, sites visited by Wireless users, etc.
4 to use a powerfull firewall program abble to propagate its
settings for every user on all LINUX SERVERS
5 to have a report page for every user on "TRAFFIC SERVER"
ATTENTION ! some of the users will be MOBILE so it's possible to have
information about the user's traffic on 2 or more servers within few
minutes only !
6 to be abble to easily create tunelling betwen 2 or more
customers (in case they want to use traffic only between them or in
case some customers want to be seen on the same WORKGROUP in network
neiborhood)
7 to use Windows 95, 98, NT, 2000 and LINUX on users' computers
B. About the management:
1 a user payes 15 USD to acces XX mb internet and YY MB
internal traffic in LAN.
. When he uses 80 % of XX internet traffic mb traffic I
would like him to receive an email warning
. When he arrives at XX mb the user shoud receive
another email warning anouncing he reached maximum payed traffic
. When he arrives at 150 % of payed traffic the user
should be disconected from the INTERNET but he should be abble to work
on Intranet if doesn't exceed the YY payed traffic.
2 when the user pays for more traffic this should be completed
in a WEB PAGE and his acces should automaticaly be active again.
I'm very interested in your opinion on my ideas.
I'm also very interested in what soft should I use in the following
problems:
- bandwith management (different speed depending of destination:
INTERNET or Wireless LAN)
- firewall
- traffic report
- cache / proxy servers (what hardware will be recommanded in case 2000
users are active at same time)
- management of users' acces in case they exceede payed traffic
All advices are wellcome. Please forgive me for any inconvenience I
might have caused you with this long letter.
I'm looking forward to hearing from you.
Yours trully,
Bogdan Florin
admin & manager
BIZAR Club
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: "Chris Riegel" <[EMAIL PROTECTED]>
Date: Wed, 02 Aug 2000 11:16:20 -0400
X-Message-Number: 12
Matt=20
>From what I've seen from this list, you advocate performance as one of =
the
>most important factors in a purchasing decision.
Yes I do, but I never assume that because I have a commercial product that =
I am by default better than Squid....
>Perhaps this is because you optimise your product to perform exceptionally=
in Polymix, rather than face
>the challenges of caching in the real world.=20
No....our configuration for Polymix is the same configuration we use for =
the real world....and by the way, at least we have guts to test under =
Polymix, as do the guys supporting Squid.....unlike your patron Cacheflow..=
>I'd like to one of your boxes submitted to the bakeoffs performing in the =
real world.
No problem....there are lots of them out there....and more every day....
>You can happily tweak TCP/IP parameters and object deletion policies for =
polymix, but does
>it really work in the wild?
Again, we don't change our config at all for Polymix vs real world....if =
you would like to come to our performance labs or to our customer sites, =
you get the same software load.....
>As a company that produces a commercial product, you already know that =
your
>product has to offer performance comparable to or exceeding that of =
squid,
>or people won't buy it.=20
Actually, I believe you were the one who said that performance wasn't =
everything when defending Cacheflow's decision not to publically test =
their products at the Bakeoffs....if performance comparable or better than =
Squid was a pre-requisite for staying in business would Cacheflow or =
Inktomi still be around??
>In many ways, Squid is the software benchmark that many commercial cache =
vendors would accept they have to beat.
No doubt that performance is one of our strong points, but overall we have =
several other strong features that make the Stratacache product line a =
well rounded, feature rich platform, as you say, performance is important, =
but not everything.....
> Your personal crusade against Cacheflow seems to be obstructing real =
discussion.=20
My personal crusade against Cacheflow....I like that, makes me sound like =
a hero or something....Small Caching Vendor takes it to the large =
corporate media machine that is the Great Caching Oz......I hope that Mel =
Gibson or Tom Cruise is available to play me in the made for TV mini =
series....:)
As for discussion, I think that my challenge to the Great Oz has sparked =
good discussion, rather than obstructing anything......
>The message below was one you posted to
>this list back in May this year, if I'm not mistaken.
Sure I like to have fun with the Squid guys about performance, and I think =
that a little competition is a good thing......
At least Squid and it's supporters have the backbone to do performance =
testing and improve the product.....Will Cacheflow be at the next =
Bakeoff???
Chris Riegel
Stratacache
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: "Matthew Watkins" <[EMAIL PROTECTED]>
Date: Wed, 2 Aug 2000 16:41:47 +0100
X-Message-Number: 13
> How the hell did we get to the subject of Cacheflow from
> a question regarding caching servers and skycaching????
The reality Riegel Reality distortion field at work again :-)
> Again, we don't change our config at all for Polymix vs real world....
> if you would like to come to our performance labs or to our customer
> sites, you get the same software load.....
Glad to here it. As you already know, I'm very cynical about such things. I
would hate to think of the torment involved... A small software engineering
team tucked away in Utah, working away in the small hours over the next few
months.... Nothing but ICS, Polymix and caffeine to keep them company.
> My personal crusade against Cacheflow....I like that, makes me sound like
> a hero or something....Small Caching Vendor takes it to the large
corporate
> media machine that is the Great Caching Oz......I hope that Mel Gibson or
Tom
> Cruise is available to play me in the made for TV mini series....:)
This sounds like a great idea. I'll start working on a script. We could call
it "Mission impossible three", or perhaps something more along the lines of
U-571, where we re-write history so that ICS saves customers from from the
evil cacheflow tyrants... ;-)
Religion once taught me that the evil in the world eminated from Satan, but
thanks to Chris I now know that it is actually generated by Cacheflow's
marketing team...
- Matt
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: "Chris Riegel" <[EMAIL PROTECTED]>
Date: Wed, 02 Aug 2000 12:20:51 -0400
X-Message-Number: 14
>> How the hell did we get to the subject of Cacheflow from
>> a question regarding caching servers and skycaching????
>The reality Riegel Reality distortion field at work again :-)
Matt...Matt...Matt...How soon you forget....
Actually it was Maxi Bernales who first brought up Cacheflow and the =
knowledgeable Joe Cooper at Swelltech who accurately questioned the =
validity of Cacheflow statistics....
The Riegel Reality distortion field didn't start this one....In fact...mayb=
e I should give the old reality distortion field a try to see if it =
works....
Reality Distortion field generator warming up.....
Bring it to full Power....
Prepare to fire....
Fire: Cacheflow is the fastest caching product in the industry.......
Does anyone believe me??? Did the reality distortion field work.....Darn..=
..I must be lousing my touch.....
Let's Try it again:
Reality Distortion field generator warming up.....
Bring it to full Power....
Prepare to fire....
Fire: Cacheflow performance isn't the problem, it's Polymix...the entire =
benchmark is structured against Cacheflow
Does anyone believe me??? Did the reality distortion field work..... 0 =
for 2....
Darn, I guess the Riegel Reality distortion field is not all it was =
cracked up to be......
>Glad to here it. As you already know, I'm very cynical about such things. =
I
>would hate to think of the torment involved... A small software engineerin=
g
>team tucked away in Utah, working away in the small hours over the next =
few
>months.... Nothing but ICS, Polymix and caffeine to keep them company.
Utah......We are located in the great State of Ohio....great products like =
ICS spread quickly....
And by the way, it is ICS, Polymix and Diet Pepsi around here....
>> My personal crusade against Cacheflow....I like that, makes me sound =
like
>> a hero or something....Small Caching Vendor takes it to the large
>>corporate media machine that is the Great Caching Oz......I hope that =
Mel Gibson or
>>Tom Cruise is available to play me in the made for TV mini series....:)
>This sounds like a great idea. I'll start working on a script. We could =
call
>it "Mission impossible three", or perhaps something more along the lines =
of
>U-571, where we re-write history so that ICS saves customers from the
>evil cacheflow tyrants... ;-)
First of all to be a tyrant you have to dominate the market...and =
Cacheflow doesn't hold that kind of power....
Secondly, as for movies, we already know that the better analogy is the =
Wizard of Oz....I am just the one to pull back the curtain, but I must =
admit that these ruby slippers are a bit tight....
>Religion once taught me that the evil in the world eminated from Satan, =
but
>thanks to Chris I now know that it is actually generated by Cacheflow's
>marketing team...
I am sure that Cacheflow's marketing team is a perfectly nice group of =
people...in fact I have never questioned the issue of good vs evil, that =
is way above my pay grade....I do question the issue of performance =
however....and others besides me have questioned the accuracy of their =
statistics...
Any members of the Cacheflow marketing team who would like to run Polymix =
on Cacheflow boxes so you show your customers the results, my lab is open =
and available.......just let me know....and if you don't trust me, I am =
sure that Duane and Alex still have room at the Bakeoff (September in =
Houston).
Quietly watching for flying monkeys....
Chris Riegel
Stratacache
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: Larry Armstrong <[EMAIL PROTECTED]>
Date: Wed, 2 Aug 2000 09:36:18 -0700 (PDT)
X-Message-Number: 15
Howdy all!
I guess my biggest problem I have with these kinds of threads is that
they degenerate to CacheFlow vs ICS instead of discussing the original
topic (as Matt so nicely pointed out :^)
Even if these mails are entertaining, they should be reigned in a bit,
no? ;^)
Thanks!
---larry >:)
_
--- Chris Riegel <[EMAIL PROTECTED]> wrote:
> >> How the hell did we get to the subject of Cacheflow from
> >> a question regarding caching servers and skycaching????
>
> >The reality Riegel Reality distortion field at work again :-)
>
> Matt...Matt...Matt...How soon you forget....
>
> Actually it was Maxi Bernales who first brought up Cacheflow and the
> knowledgeable Joe Cooper at Swelltech who accurately questioned the
> validity of Cacheflow statistics....
>
> The Riegel Reality distortion field didn't start this one....In
> fact...maybe I should give the old reality distortion field a try to
> see if it works....
>
> Reality Distortion field generator warming up.....
>
> Bring it to full Power....
>
> Prepare to fire....
>
> Fire: Cacheflow is the fastest caching product in the industry.......
>
> Does anyone believe me??? Did the reality distortion field
> work.....Darn....I must be lousing my touch.....
>
> Let's Try it again:
>
> Reality Distortion field generator warming up.....
>
> Bring it to full Power....
>
> Prepare to fire....
>
> Fire: Cacheflow performance isn't the problem, it's Polymix...the
> entire benchmark is structured against Cacheflow
>
> Does anyone believe me??? Did the reality distortion field work.....
> 0 for 2....
>
> Darn, I guess the Riegel Reality distortion field is not all it was
> cracked up to be......
>
> >Glad to here it. As you already know, I'm very cynical about such
> things. I
> >would hate to think of the torment involved... A small software
> engineering
> >team tucked away in Utah, working away in the small hours over the
> next few
> >months.... Nothing but ICS, Polymix and caffeine to keep them
> company.
>
> Utah......We are located in the great State of Ohio....great products
> like ICS spread quickly....
>
> And by the way, it is ICS, Polymix and Diet Pepsi around here....
>
> >> My personal crusade against Cacheflow....I like that, makes me
> sound like
> >> a hero or something....Small Caching Vendor takes it to the large
> >>corporate media machine that is the Great Caching Oz......I hope
> that Mel Gibson or
> >>Tom Cruise is available to play me in the made for TV mini
> series....:)
>
> >This sounds like a great idea. I'll start working on a script. We
> could call
> >it "Mission impossible three", or perhaps something more along the
> lines of
> >U-571, where we re-write history so that ICS saves customers from
> the
> >evil cacheflow tyrants... ;-)
>
> First of all to be a tyrant you have to dominate the market...and
> Cacheflow doesn't hold that kind of power....
>
> Secondly, as for movies, we already know that the better analogy is
> the Wizard of Oz....I am just the one to pull back the curtain, but I
> must admit that these ruby slippers are a bit tight....
>
> >Religion once taught me that the evil in the world eminated from
> Satan, but
> >thanks to Chris I now know that it is actually generated by
> Cacheflow's
> >marketing team...
>
> I am sure that Cacheflow's marketing team is a perfectly nice group
> of people...in fact I have never questioned the issue of good vs
> evil, that is way above my pay grade....I do question the issue of
> performance however....and others besides me have questioned the
> accuracy of their statistics...
>
> Any members of the Cacheflow marketing team who would like to run
> Polymix on Cacheflow boxes so you show your customers the results, my
> lab is open and available.......just let me know....and if you don't
> trust me, I am sure that Duane and Alex still have room at the
> Bakeoff (September in Houston).
>
> Quietly watching for flying monkeys....
>
>
> Chris Riegel
> Stratacache
>
>
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> NEWS YOU CAN USE
> Check Out http://www.siliconvalley.internet.com
> For the Latest Internet News From Silicon Valley
>
> ____________ • The ISP-CACHING Discussion List • ___________
> To Join: mailto:[EMAIL PROTECTED]
> To Remove: mailto:[EMAIL PROTECTED]
> Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
__________________________________________________
Do You Yahoo!?
Kick off your party with Yahoo! Invites.
http://invites.yahoo.com/
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: "Andy Paluch" <[EMAIL PROTECTED]>
Date: Wed, 2 Aug 2000 12:50:14 -0500
X-Message-Number: 16
----- Original Message -----
From: Matthew Watkins <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Wednesday, August 02, 2000 10:41 AM
Subject: Re: Caching Products and Skycaching
> > How the hell did we get to the subject of Cacheflow from
> > a question regarding caching servers and skycaching????
>
> The reality Riegel Reality distortion field at work again :-)
>
> > Again, we don't change our config at all for Polymix vs real world....
> > if you would like to come to our performance labs or to our customer
> > sites, you get the same software load.....
>
> Glad to here it. As you already know, I'm very cynical about such things.
I
> would hate to think of the torment involved... A small software
engineering
> team tucked away in Utah, working away in the small hours over the next
few
> months.... Nothing but ICS, Polymix and caffeine to keep them company.
>
> > My personal crusade against Cacheflow....I like that, makes me sound
like
> > a hero or something....Small Caching Vendor takes it to the large
> corporate
> > media machine that is the Great Caching Oz......I hope that Mel Gibson
or
> Tom
> > Cruise is available to play me in the made for TV mini series....:)
>
> This sounds like a great idea. I'll start working on a script. We could
call
> it "Mission impossible three", or perhaps something more along the lines
of
> U-571, where we re-write history so that ICS saves customers from from
the
> evil cacheflow tyrants... ;-)
>
> Religion once taught me that the evil in the world eminated from Satan,
but
> thanks to Chris I now know that it is actually generated by Cacheflow's
> marketing team...
>
> - Matt
>
>
>
>
> ____________ . The ISP-CACHING Discussion List . ___________
> To Join: mailto:[EMAIL PROTECTED]
> To Remove: mailto:[EMAIL PROTECTED]
> Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
>
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: [EMAIL PROTECTED] (Shane Bennett)
Date: Tue, 1 Aug 2000 23:22:30 -0700
X-Message-Number: 17
Its very childISHH and turns people of this list like the Wireless
(breese-CON ) list
Shane
----- Original Message -----
From: "Larry Armstrong" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Wednesday, August 02, 2000 9:36 AM
Subject: Re: Caching Products and Skycaching
> Howdy all!
>
> I guess my biggest problem I have with these kinds of threads is that
> they degenerate to CacheFlow vs ICS instead of discussing the original
> topic (as Matt so nicely pointed out :^)
>
> Even if these mails are entertaining, they should be reigned in a bit,
> no? ;^)
>
> Thanks!
> ---larry >:)
> _
> --- Chris Riegel <[EMAIL PROTECTED]> wrote:
> > >> How the hell did we get to the subject of Cacheflow from
> > >> a question regarding caching servers and skycaching????
> >
> > >The reality Riegel Reality distortion field at work again :-)
> >
> > Matt...Matt...Matt...How soon you forget....
> >
> > Actually it was Maxi Bernales who first brought up Cacheflow and the
> > knowledgeable Joe Cooper at Swelltech who accurately questioned the
> > validity of Cacheflow statistics....
> >
> > The Riegel Reality distortion field didn't start this one....In
> > fact...maybe I should give the old reality distortion field a try to
> > see if it works....
> >
> > Reality Distortion field generator warming up.....
> >
> > Bring it to full Power....
> >
> > Prepare to fire....
> >
> > Fire: Cacheflow is the fastest caching product in the industry.......
> >
> > Does anyone believe me??? Did the reality distortion field
> > work.....Darn....I must be lousing my touch.....
> >
> > Let's Try it again:
> >
> > Reality Distortion field generator warming up.....
> >
> > Bring it to full Power....
> >
> > Prepare to fire....
> >
> > Fire: Cacheflow performance isn't the problem, it's Polymix...the
> > entire benchmark is structured against Cacheflow
> >
> > Does anyone believe me??? Did the reality distortion field work.....
> > 0 for 2....
> >
> > Darn, I guess the Riegel Reality distortion field is not all it was
> > cracked up to be......
> >
> > >Glad to here it. As you already know, I'm very cynical about such
> > things. I
> > >would hate to think of the torment involved... A small software
> > engineering
> > >team tucked away in Utah, working away in the small hours over the
> > next few
> > >months.... Nothing but ICS, Polymix and caffeine to keep them
> > company.
> >
> > Utah......We are located in the great State of Ohio....great products
> > like ICS spread quickly....
> >
> > And by the way, it is ICS, Polymix and Diet Pepsi around here....
> >
> > >> My personal crusade against Cacheflow....I like that, makes me
> > sound like
> > >> a hero or something....Small Caching Vendor takes it to the large
> > >>corporate media machine that is the Great Caching Oz......I hope
> > that Mel Gibson or
> > >>Tom Cruise is available to play me in the made for TV mini
> > series....:)
> >
> > >This sounds like a great idea. I'll start working on a script. We
> > could call
> > >it "Mission impossible three", or perhaps something more along the
> > lines of
> > >U-571, where we re-write history so that ICS saves customers from
> > the
> > >evil cacheflow tyrants... ;-)
> >
> > First of all to be a tyrant you have to dominate the market...and
> > Cacheflow doesn't hold that kind of power....
> >
> > Secondly, as for movies, we already know that the better analogy is
> > the Wizard of Oz....I am just the one to pull back the curtain, but I
> > must admit that these ruby slippers are a bit tight....
> >
> > >Religion once taught me that the evil in the world eminated from
> > Satan, but
> > >thanks to Chris I now know that it is actually generated by
> > Cacheflow's
> > >marketing team...
> >
> > I am sure that Cacheflow's marketing team is a perfectly nice group
> > of people...in fact I have never questioned the issue of good vs
> > evil, that is way above my pay grade....I do question the issue of
> > performance however....and others besides me have questioned the
> > accuracy of their statistics...
> >
> > Any members of the Cacheflow marketing team who would like to run
> > Polymix on Cacheflow boxes so you show your customers the results, my
> > lab is open and available.......just let me know....and if you don't
> > trust me, I am sure that Duane and Alex still have room at the
> > Bakeoff (September in Houston).
> >
> > Quietly watching for flying monkeys....
> >
> >
> > Chris Riegel
> > Stratacache
> >
> >
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > NEWS YOU CAN USE
> > Check Out http://www.siliconvalley.internet.com
> > For the Latest Internet News From Silicon Valley
> >
> > ____________ . The ISP-CACHING Discussion List . ___________
> > To Join: mailto:[EMAIL PROTECTED]
> > To Remove: mailto:[EMAIL PROTECTED]
> > Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
>
>
> __________________________________________________
> Do You Yahoo!?
> Kick off your party with Yahoo! Invites.
> http://invites.yahoo.com/
>
>
>
> ____________ . The ISP-CACHING Discussion List . ___________
> To Join: mailto:[EMAIL PROTECTED]
> To Remove: mailto:[EMAIL PROTECTED]
> Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: "Michael Masinelli" <[EMAIL PROTECTED]>
Date: Wed, 02 Aug 2000 14:34:16 -0600
X-Message-Number: 18
Yes, I would ask that the issue of Cacheflow & ICS be dropped.=20
I usually get good information from this list and would like to continue =
to do so. Please, let's move on..
Thank you,
MM
>>> [EMAIL PROTECTED] 08/02/00 01:22AM >>>
Its very childISHH and turns people of this list like the Wireless
(breese-CON ) list
Shane
----- Original Message -----
From: "Larry Armstrong" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Wednesday, August 02, 2000 9:36 AM
Subject: Re: Caching Products and Skycaching
> Howdy all!
>
> I guess my biggest problem I have with these kinds of threads is that
> they degenerate to CacheFlow vs ICS instead of discussing the original
> topic (as Matt so nicely pointed out :^)
>
> Even if these mails are entertaining, they should be reigned in a bit,
> no? ;^)
>
> Thanks!
> ---larry >:)
> _
> --- Chris Riegel <[EMAIL PROTECTED]> wrote:
> > >> How the hell did we get to the subject of Cacheflow from
> > >> a question regarding caching servers and skycaching????
> >
> > >The reality Riegel Reality distortion field at work again :-)
> >
> > Matt...Matt...Matt...How soon you forget....
> >
> > Actually it was Maxi Bernales who first brought up Cacheflow and the
> > knowledgeable Joe Cooper at Swelltech who accurately questioned the
> > validity of Cacheflow statistics....
> >
> > The Riegel Reality distortion field didn't start this one....In
> > fact...maybe I should give the old reality distortion field a try to
> > see if it works....
> >
> > Reality Distortion field generator warming up.....
> >
> > Bring it to full Power....
> >
> > Prepare to fire....
> >
> > Fire: Cacheflow is the fastest caching product in the industry.......
> >
> > Does anyone believe me??? Did the reality distortion field
> > work.....Darn....I must be lousing my touch.....
> >
> > Let's Try it again:
> >
> > Reality Distortion field generator warming up.....
> >
> > Bring it to full Power....
> >
> > Prepare to fire....
> >
> > Fire: Cacheflow performance isn't the problem, it's Polymix...the
> > entire benchmark is structured against Cacheflow
> >
> > Does anyone believe me??? Did the reality distortion field work.....
> > 0 for 2....
> >
> > Darn, I guess the Riegel Reality distortion field is not all it was
> > cracked up to be......
> >
> > >Glad to here it. As you already know, I'm very cynical about such
> > things. I
> > >would hate to think of the torment involved... A small software
> > engineering
> > >team tucked away in Utah, working away in the small hours over the
> > next few
> > >months.... Nothing but ICS, Polymix and caffeine to keep them
> > company.
> >
> > Utah......We are located in the great State of Ohio....great products
> > like ICS spread quickly....
> >
> > And by the way, it is ICS, Polymix and Diet Pepsi around here....
> >
> > >> My personal crusade against Cacheflow....I like that, makes me
> > sound like
> > >> a hero or something....Small Caching Vendor takes it to the large
> > >>corporate media machine that is the Great Caching Oz......I hope
> > that Mel Gibson or
> > >>Tom Cruise is available to play me in the made for TV mini
> > series....:)
> >
> > >This sounds like a great idea. I'll start working on a script. We
> > could call
> > >it "Mission impossible three", or perhaps something more along the
> > lines of
> > >U-571, where we re-write history so that ICS saves customers from
> > the
> > >evil cacheflow tyrants... ;-)
> >
> > First of all to be a tyrant you have to dominate the market...and
> > Cacheflow doesn't hold that kind of power....
> >
> > Secondly, as for movies, we already know that the better analogy is
> > the Wizard of Oz....I am just the one to pull back the curtain, but I
> > must admit that these ruby slippers are a bit tight....
> >
> > >Religion once taught me that the evil in the world eminated from
> > Satan, but
> > >thanks to Chris I now know that it is actually generated by
> > Cacheflow's
> > >marketing team...
> >
> > I am sure that Cacheflow's marketing team is a perfectly nice group
> > of people...in fact I have never questioned the issue of good vs
> > evil, that is way above my pay grade....I do question the issue of
> > performance however....and others besides me have questioned the
> > accuracy of their statistics...
> >
> > Any members of the Cacheflow marketing team who would like to run
> > Polymix on Cacheflow boxes so you show your customers the results, my
> > lab is open and available.......just let me know....and if you don't
> > trust me, I am sure that Duane and Alex still have room at the
> > Bakeoff (September in Houston).
> >
> > Quietly watching for flying monkeys....
> >
> >
> > Chris Riegel
> > Stratacache
> >
> >
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > NEWS YOU CAN USE
> > Check Out http://www.siliconvalley.internet.com=20
> > For the Latest Internet News From Silicon Valley
> >
> > ____________ . The ISP-CACHING Discussion List . ___________
> > To Join: mailto:[EMAIL PROTECTED]=20
> > To Remove: mailto:[EMAIL PROTECTED]=20
> > Archives: http://isp-lists.isp-planet.com/isp-caching/archives/=20
>
>
> __________________________________________________
> Do You Yahoo!?
> Kick off your party with Yahoo! Invites.
> http://invites.yahoo.com/=20
>
>
>
> ____________ . The ISP-CACHING Discussion List . ___________
> To Join: mailto:[EMAIL PROTECTED]=20
> To Remove: mailto:[EMAIL PROTECTED]=20
> Archives: http://isp-lists.isp-planet.com/isp-caching/archives/=20
____________ * The ISP-CACHING Discussion List * ___________
To Join: mailto:[EMAIL PROTECTED]=20
To Remove: mailto:[EMAIL PROTECTED]=20
Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: Wayne Eichenberger <[EMAIL PROTECTED]>
Date: Wed, 02 Aug 2000 16:57:23 -0400
X-Message-Number: 19
Agreed, I'm quite sick of this as well.
Wayne
Andy Paluch wrote:
> ----- Original Message -----
> From: Matthew Watkins <[EMAIL PROTECTED]>
> To: <[EMAIL PROTECTED]>
> Sent: Wednesday, August 02, 2000 10:41 AM
> Subject: Re: Caching Products and Skycaching
>
> > > How the hell did we get to the subject of Cacheflow from
> > > a question regarding caching servers and skycaching????
> >
> > The reality Riegel Reality distortion field at work again :-)
> >
> > > Again, we don't change our config at all for Polymix vs real world....
> > > if you would like to come to our performance labs or to our customer
> > > sites, you get the same software load.....
> >
> > Glad to here it. As you already know, I'm very cynical about such things.
> I
> > would hate to think of the torment involved... A small software
> engineering
> > team tucked away in Utah, working away in the small hours over the next
> few
> > months.... Nothing but ICS, Polymix and caffeine to keep them company.
> >
> > > My personal crusade against Cacheflow....I like that, makes me sound
> like
> > > a hero or something....Small Caching Vendor takes it to the large
> > corporate
> > > media machine that is the Great Caching Oz......I hope that Mel Gibson
> or
> > Tom
> > > Cruise is available to play me in the made for TV mini series....:)
> >
> > This sounds like a great idea. I'll start working on a script. We could
> call
> > it "Mission impossible three", or perhaps something more along the lines
> of
> > U-571, where we re-write history so that ICS saves customers from from
> the
> > evil cacheflow tyrants... ;-)
> >
> > Religion once taught me that the evil in the world eminated from Satan,
> but
> > thanks to Chris I now know that it is actually generated by Cacheflow's
> > marketing team...
> >
> > - Matt
> >
> >
> >
> >
> > ____________ . The ISP-CACHING Discussion List . ___________
> > To Join: mailto:[EMAIL PROTECTED]
> > To Remove: mailto:[EMAIL PROTECTED]
> > Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
> >
>
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> NEWS YOU CAN USE
> Check Out http://www.siliconvalley.internet.com
> For the Latest Internet News From Silicon Valley
>
> ____________ • The ISP-CACHING Discussion List • ___________
> To Join: mailto:[EMAIL PROTECTED]
> To Remove: mailto:[EMAIL PROTECTED]
> Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
----------------------------------------------------------------------
Subject: RE: Caching Products and Skycaching
From: [EMAIL PROTECTED] (Richard DeSoto)
Date: Wed, 2 Aug 2000 14:07:03 -0700
X-Message-Number: 20
Lets use the discussion group for what it was intended for and not for
vendor bashing. Only the bashers really care.
-richard
-----Original Message-----
From: Michael Masinelli [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, August 02, 2000 1:34 PM
To: [EMAIL PROTECTED]
Subject: Re: Caching Products and Skycaching
Yes, I would ask that the issue of Cacheflow & ICS be dropped.
I usually get good information from this list and would like to continue to
do so. Please, let's move on..
Thank you,
MM
>>> [EMAIL PROTECTED] 08/02/00 01:22AM >>>
Its very childISHH and turns people of this list like the Wireless
(breese-CON ) list
Shane
----- Original Message -----
From: "Larry Armstrong" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Wednesday, August 02, 2000 9:36 AM
Subject: Re: Caching Products and Skycaching
> Howdy all!
>
> I guess my biggest problem I have with these kinds of threads is that
> they degenerate to CacheFlow vs ICS instead of discussing the original
> topic (as Matt so nicely pointed out :^)
>
> Even if these mails are entertaining, they should be reigned in a bit,
> no? ;^)
>
> Thanks!
> ---larry >:)
> _
> --- Chris Riegel <[EMAIL PROTECTED]> wrote:
> > >> How the hell did we get to the subject of Cacheflow from
> > >> a question regarding caching servers and skycaching????
> >
> > >The reality Riegel Reality distortion field at work again :-)
> >
> > Matt...Matt...Matt...How soon you forget....
> >
> > Actually it was Maxi Bernales who first brought up Cacheflow and the
> > knowledgeable Joe Cooper at Swelltech who accurately questioned the
> > validity of Cacheflow statistics....
> >
> > The Riegel Reality distortion field didn't start this one....In
> > fact...maybe I should give the old reality distortion field a try to
> > see if it works....
> >
> > Reality Distortion field generator warming up.....
> >
> > Bring it to full Power....
> >
> > Prepare to fire....
> >
> > Fire: Cacheflow is the fastest caching product in the industry.......
> >
> > Does anyone believe me??? Did the reality distortion field
> > work.....Darn....I must be lousing my touch.....
> >
> > Let's Try it again:
> >
> > Reality Distortion field generator warming up.....
> >
> > Bring it to full Power....
> >
> > Prepare to fire....
> >
> > Fire: Cacheflow performance isn't the problem, it's Polymix...the
> > entire benchmark is structured against Cacheflow
> >
> > Does anyone believe me??? Did the reality distortion field work.....
> > 0 for 2....
> >
> > Darn, I guess the Riegel Reality distortion field is not all it was
> > cracked up to be......
> >
> > >Glad to here it. As you already know, I'm very cynical about such
> > things. I
> > >would hate to think of the torment involved... A small software
> > engineering
> > >team tucked away in Utah, working away in the small hours over the
> > next few
> > >months.... Nothing but ICS, Polymix and caffeine to keep them
> > company.
> >
> > Utah......We are located in the great State of Ohio....great products
> > like ICS spread quickly....
> >
> > And by the way, it is ICS, Polymix and Diet Pepsi around here....
> >
> > >> My personal crusade against Cacheflow....I like that, makes me
> > sound like
> > >> a hero or something....Small Caching Vendor takes it to the large
> > >>corporate media machine that is the Great Caching Oz......I hope
> > that Mel Gibson or
> > >>Tom Cruise is available to play me in the made for TV mini
> > series....:)
> >
> > >This sounds like a great idea. I'll start working on a script. We
> > could call
> > >it "Mission impossible three", or perhaps something more along the
> > lines of
> > >U-571, where we re-write history so that ICS saves customers from
> > the
> > >evil cacheflow tyrants... ;-)
> >
> > First of all to be a tyrant you have to dominate the market...and
> > Cacheflow doesn't hold that kind of power....
> >
> > Secondly, as for movies, we already know that the better analogy is
> > the Wizard of Oz....I am just the one to pull back the curtain, but I
> > must admit that these ruby slippers are a bit tight....
> >
> > >Religion once taught me that the evil in the world eminated from
> > Satan, but
> > >thanks to Chris I now know that it is actually generated by
> > Cacheflow's
> > >marketing team...
> >
> > I am sure that Cacheflow's marketing team is a perfectly nice group
> > of people...in fact I have never questioned the issue of good vs
> > evil, that is way above my pay grade....I do question the issue of
> > performance however....and others besides me have questioned the
> > accuracy of their statistics...
> >
> > Any members of the Cacheflow marketing team who would like to run
> > Polymix on Cacheflow boxes so you show your customers the results, my
> > lab is open and available.......just let me know....and if you don't
> > trust me, I am sure that Duane and Alex still have room at the
> > Bakeoff (September in Houston).
> >
> > Quietly watching for flying monkeys....
> >
> >
> > Chris Riegel
> > Stratacache
> >
> >
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > NEWS YOU CAN USE
> > Check Out http://www.siliconvalley.internet.com
> > For the Latest Internet News From Silicon Valley
> >
> > ____________ . The ISP-CACHING Discussion List . ___________
> > To Join: mailto:[EMAIL PROTECTED]
> > To Remove: mailto:[EMAIL PROTECTED]
> > Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
>
>
> __________________________________________________
> Do You Yahoo!?
> Kick off your party with Yahoo! Invites.
> http://invites.yahoo.com/
>
>
>
> ____________ . The ISP-CACHING Discussion List . ___________
> To Join: mailto:[EMAIL PROTECTED]
> To Remove: mailto:[EMAIL PROTECTED]
> Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
____________ * The ISP-CACHING Discussion List * ___________
To Join: mailto:[EMAIL PROTECTED]
To Remove: mailto:[EMAIL PROTECTED]
Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RADWARE, Inc.: The only company offering a complete local/
global IP load balancing solution for all Internet/intranet/
extranet environments. http://www.radware.com
____________ * The ISP-CACHING Discussion List * ___________
To Join: mailto:[EMAIL PROTECTED]
To Remove: mailto:[EMAIL PROTECTED]
Archives: http://isp-lists.isp-planet.com/isp-caching/archives/
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: Paul <[EMAIL PROTECTED]>
Date: Wed, 2 Aug 2000 18:13:44 -0400 (EDT)
X-Message-Number: 21
> dialup ISP. We want to to know if Skycaching with any of these appliances
> can help us. Of course, the alternative is redundant Backbones. Any
> ideas?
>
> =================================
> Tim Maveal
> Lifestyle Internet Services
> (775) 954-1600
> =================================
Tim,
Here are some sample statistics from one of Cidera's
"Skycaching" customers.
The log is for 24hrs. The customer uses the new "combo" box which
runs squid and uses a 36G internal cache to store a week's worth
of the broadcast. Their cache peers with the skycache squid.
Looking at the stats, the first hit data shows skycache was there first
over 70% of the time. Several large .swf (shockwave) files were hit
to help boost the bandwidth savings to ~400M on this day.
Here is the second largest item hit.
http://www.joecartoon.com/buddies/safari/safari2.swf
Output from log analysis..
ComboStats 1.1 (c) 2000 Cidera, Inc.
log(s): iwis4
No base time entered -- using present (959876529)...
Processing log: iwis4... done.
-- RAW HIT/MISS RATE ---
Objects: 745703 hits(48%), 790038 misses(51%) of 1535741 total.
Bytes: 2223846820 hits(26%), 6136403365 misses(73%) of 8360250185
--- SKYCACHE SAVINGS DATA ---
* Skycache injected 6% (400184292) of the bytes, and 9% (83814) of the
objects.
* Skycache saved the customer 381.65M of land line retrieval.
* Skycache increased the byte hitrate 17%, and object hitrate 11%.
--- FIRST HIT DISTRIBUTION INFORMATION ---
skycache: 83814 objects(71%), 400184292 bytes(62%).
local cache: 33357 objects(28%), 245260397 bytes(37%).
total: 117171 objects, 645444689 bytes.
----------------------------------------------------------------------
Subject: Re: Caching Products and Skycaching
From: [EMAIL PROTECTED]
Date: Wed, 02 Aug 2000 20:59:07 EDT
X-Message-Number: 22
please dont send me anymore updates THANK YOU
----------------------------------------------------------------------
Subject: caching seminars?
From: "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>
Date: Thu, 3 Aug 2000 13:05:55 +0800 (PHT)
X-Message-Number: 23
Hi all,
I just want to inquire if there are forth coming events or activities on
caching that you know.
I just want to keep myself noted on those events.
=)
Maxi Bernales
"A razor may be sharper than an ax, but it cannot cut wood."
---
END OF DIGEST
---
You are currently subscribed to isp-caching as: archive@jab.org
isp-caching digest: August 02, 2000
ISP-CACHING Discussion List digest Wed, 02 Aug 2000 22:00:06 -0700
