[4124] in comp.os.os2.announce archive
PR: Project Kickin Butt
daemon@ATHENA.MIT.EDU (Dan Porter)
Sun Jun 14 15:18:00 1998
To: os2ann.DISCUSS@bloom-picayune.mit.edu
Date: 14 Jun 1998 14:02:00 -0400
From: innoval@ibm.net (Dan Porter)
Reply-To: innoval@ibm.net (Dan Porter)
Reply-to: innoval@ibm.net (Dan Porter)
[Followups directed to comp.os.os2.networking.www]
---------------------------------------------------------------------
Project Kickin Butt
InnoVal has been working on a Java-based web accelerator, code named Kickin
Butt. We are seriously considering a native OS/2 version, as well. Web
accelerators, properly designed and coded, can provide significant
performance boosts with most browsers such as Netscape Navigator, Opera,
etc.
There are a few win32 web accelerators on the market. Most offer some
performance improvements. Most of them (the impulse-buy-pretty-package
variety for Windows 95 found at office superstores, computer stores, and car
washes) are okay but claim wildly exaggerated boosts in performance. Windows
95 is in fact is a poor platform for web accelerators and the sort of
results promised are not realistic. A large software producer (a client of
ours) has built two web very advanced accelerators, one for NT and one for
Unix. Because they are properly designed, they offer significant performance
advantages. Our client reports the following sustained performance boosts
with Netscape Navigator 2.0 and 4.0 running on Windows NT.
Dial in 28kbs - 2.6x
Dial in 56kbs - 2.9x
T1 Connection - 1.8x
In my own testing with a web accelerator, I have seen performance boosts
ranging from 2.1x to 5.0x in 30-minute sessions with a 56kbs connection.
Web accelerators are generally implemented as client-side caching proxies.
That is the easy part. The hard part is page and image chasing strategies,
the many anomalies of HTTP, applets, frames, etc. Web accelerators,
admittedly do waste bandwidth but it isn't burdensome if "chasing"
strategies and thread priorities are properly implemented.
We hope to bring our Java version of Kickin Butt to market later this
summer. The question I have for the OS/2 user community is this: Is there
enough interest to build a native OS/2 version? If I can get 1,000
individual and SOHO OS/2 users to say, "I'll buy a native OS/2 version of
Kickin Butt for $25.00," then InnoVal will build it and bring it to market
during August. I'm not asking you for any money at this time.
This is something that I personally want to do. The Java version will work
well for many of our customers, particularly in our corporate accounts that
are already committed to Java and have multiple platforms in-house. I
believe that there may enough interest in the OS/2 individual and SOHO
market for a resource-leaner, native OS/2 version of Kickin Butt. It will
require a significant cash outlay and committed resources on our part. I
would love to stand up at Warpstock 98 and in front of OS/2 users groups and
demonstrate Kickin Butt. I'd love, again, to bring a totally new OS/2
product to market.
Let me know if you would spend $25.00 for Kickin Butt. Send an email to
kickbutt@innoval.com.
BTW: I doubt that Kickin Butt will be the final product name.
Dan Porter
InnoVal Systems Solutions, Inc.
http://www.innoval.com
Posted Using: J Street Mailer (build 98.5.3)
_____________________________________________________________________
| NOTE: Please EMAIL submissions to comp-os-os2-announce@moderators.isc.org
| Correspondence to the COOA Moderator: lfirrantello@bix.com .
| Please see: http://www.bix.com/pub/os2ann/pindex.htm for posting guidelines
