🚀 go-pugleaf

On Thu, 15 Feb 2001 16:00:36 GMT, John Navas
<spamfilter@navasgrp.dublin.ca.us> wrote:

>[POSTED TO comp.dcom.modems.cable; PLEASE REPLY THERE]
>
>In <3a8be731.161275899@news1.news.adelphia.net>, bubba@bubbanet.com
>wrote:
>
>>snip
>>>
>>>1.  While it's possible to screw up anything, I think it rare in
>>>practice that proxy cache would be worse than cache bypass -- transit
>>>capacity is a lot more expensive than cache capacity.  And I'm pretty
>>>sure that ISP's do measure and monitor cache performance.
>>>
>>>2.  I've seen little or no evidence of cases where proxy cache was
>>>actually worse than cache bypass -- have you?  
>>
>>I have experienced it first hand.  It may be that they don't know how
>>to configure it, but when my ISP put in a forced proxy  (Cacheflow) my
>>browsing speeds slowed to a crawl (3-5K/sec) while I could download a
>>file at over T1 speed.
>
>Do all subscribers see such low speeds on cache, or only you?

Everyone did...but my ISP had close to 2000 people on 3 T1's for 6
months so they expect slow speeds.  Also many probably didn't know
that the pages they got were old.
>
>>It also routinely sent old pages.  I sent them
>>enough proof of this and they took it off.  Speeds went back up.  They
>>put it back and now I'm getting old pages all the time again.
>
>Although it's certainly possible that the Cacheflow is misconfigured, if
>reload works to update the page (as you indicated in another article),
>then the problem is probably in your browser (cache) settings, not the
>Cacheflow.

Nope, I turned the cache off to test for that...it didn't change
anything.
>
>FWIW, my own ISP is testing Cacheflow, and I haven't seen any problems
>with stale pages (with IE 5.5 SP1 set to Check for newer versions of
>stored pages Automatically).
>
>>I can't
>>tell about the speed issue, because they also have a router problem
>>that has increased pings by about six times and I generally get 15%
>>packet loss by the time I leave their ATM cloud.  Speeds are slow, but
>>it may not be the proxy this time.   
>
>It sounds like there are problems other than the Cacheflow.  If so, then
>it would be hard to make a fair assessment of the Cacheflow.

When they had the initial problem with speed, there were no router
problems.  I doubt that most of the problems are caused by the
Cacheflow, suspect that it's the lack of expertise on the part of the
people configuring it.  If they can make it seemless for me and not
degrade my service with slow speeds and/or old pages I'm all for it.
If they can't, they should make it optional for customers IMHO. 
>
>>>3.  As I noted in my prior reply, cache bypass is selfish and
>>>counterproductive, since it reduces efficiency to the detriment of other
>>>subscribers.
>>
>>Forced cache inclusion under these circumstances is selfish and works
>>to the detriment of all subscribers.  Also, their cache is attached to
>>the same pipe I am, so there's no real chance for performance
>>improvement, unless I get a thrill from seeing expired/old pages
>>really fast.  This still holds.  Fix it or make optional I say!
>
>The Cacheflow is almost certainly on a much faster pipe than you are,
>and is able to pull full pages faster from remote servers (due to
>efficiency as well as pipe size) when those pages aren't in cache.

If it did that it would be ducky, but it sends me old pages instead
and it's not my browser cache.

I wish it was...but it's not.  My MSO has one DS3, everyone uses it.
There's no second pipe.
>
>-- 
>Best regards,
>John Navas     <http://navasgrp.home.att.net/>
>     CABLE MODEM/DSL GUIDE:  <http://Cable-DSL.home.att.net/>

Bubba