Hi. (Pls excuse, new to this list.) On Thursday, 23 September 1999 16:18, Josh Bailey [SMTP:joshbailey(a)lucent.com] wrote:
Unfortunately the only real way you will be able to tell, for any given environment, is try it and see in a small scale, controlled trail. No vendor can provide you with the proverbial rock-solid gaurantee that a wildly hetreogenous populution of clients from many different manufacturers, firmware versions, etc will respond well to a given implementation.
At my previous employer, we did that, with 5200 and Microcom v.34 (K56Flex not released by then.) All went well with 1 PRI from Clear dms100 and 1 fully equipped 5200. Ran perfectly. Put 3 x 5200's maxed out with Microcom 56K modems (but limited to v.34) and we went through 14 months of bug fixes and reloads and memory leaks and all sorts of shit. At one stage, the Telco switch was signalling the end of a call, and the NAS was ignoring it as it hadn't received a PPP or LAPM disconnect notice. This was of course fixed, memory leaks never were. Needless to say, the 5300's were actually much better when we trialled them, even with MICA. The point? I no no longer trust the accuracy of Pilot's. The pilot of 5200 went perfectly, the implementation was hell. Cheers - --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
On Thu, 23 Sep 1999, Steve Lang wrote:
The point? I no no longer trust the accuracy of Pilot's. The pilot of 5200 went perfectly, the implementation was hell.
Hi Steve; To be fair to Cisco, if I understand your message correctly, your trial did not actually involve the hardware you went live with. You had V.34-only modems, then more-than-V.34 modems, *limited* to V.34. Many modem DSP vendors, right up to when V.34 was most current, wrote their pump and even controller code in assembler. But like most, MICA probably reimplemented that in something like C on their 56K devices, using memory protection and exception handling not available on their older hardware. I've even found pump and controller code written in assembler for TI V.90 modem chipsets dated early '99! Believe or not, a good many central-site and client modem controller/DSP controllers are based on the venerable, heavily licensed MOS tech. 6502. Whenever you rewrite something from scratch, there is risk that you will miss something. Further - depending on how you limit to V.34 on the MICA - there may've been other contributors (like V.34 only with V.8bis left on). Also being fair to Cisco - both 'Flex and V.90 took a long time to mature - even if Cisco's code were perfect, there were plenty of imperfect clients it would've had to deal with. -- Josh Bailey (mailto:joshbailey(a)lucent.com) lucent->ins->software->alameda[CA] /* 1601 Harbor Bay Parkway, Alameda, CA 94502 (room 1601/1108C) voice: +1-510-747-3367 skytel: 1-800-skytel2/mailto:1198428(a)skytel.com */ --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
participants (2)
-
Josh Bailey -
Steve Lang