Date: Fri, 12 Oct 2001 22:23:55 -0700 From: "Ted Mittelstaedt" <tedm@toybox.placo.com> To: <Bsdguru@aol.com> Cc: <hackers@freebsd.org> Subject: RE: Imagestream WanIC-520 interface cards Message-ID: <000401c153a7$409d9ec0$1401a8c0@tedm.placo.com> In-Reply-To: <103.a728861.28f84948@aol.com>
next in thread | previous in thread | raw e-mail | index | archive | help
>-----Original Message----- >From: Bsdguru@aol.com [mailto:Bsdguru@aol.com] >Sent: Friday, October 12, 2001 6:25 AM >To: tedm@toybox.placo.com >Cc: hackers@freebsd.org >Subject: Re: Imagestream WanIC-520 interface cards > > >In a message dated 10/12/01 1:08:44 AM Eastern Daylight Time, >tedm@toybox.placo.com writes: > >> Consider that the Hitachi controller chip used on the WANic 405 is the >> SAME chip that Cisco uses in it's 25xx series of routers, and the Cisco >> 2501 is the most used router in the world and has the most installed >> units. >> >> SBS Communications is STILL selling the RISCom series of cards which is >> the predecessor to the WANic, uses the same controller, these are >ISA cards >> that have a design that's over 10 years old. >> >> You tell your supplier that since the 405 is being "phased out" that he >> should sell you a bunch of them at closeout prices. >> > >Perhaps the fact that the (incredibly slow) 2501 has been >phased out has caused Hitachi to "phase out" their part? You know I really think your baiting me. You know perfectly well that Hitachi and Cisco are not the same company. Anyway, Cisco is still selling the 2501 router, (although list on it is $2600 and nobody buying it for a single connection to the Internet would buy it new) I just logged into their reseller's website and the 2501 is still readily available. And as far as speed of the router, yes the 2501 is incredibly slow. So what? This has absolutely nothing whatsover to do with the HD64570 controller chip which is rated to 7Mbts. The 2501 is slow because it has a 20Mhz Motorola 68000 series CPU, not because the serial controller chip is slow. >You know, big chip >vendors have bigger fish to fry than the WanIC series. I don't understand why your spreading FUD when a few minutes work on the Hitachi website could answer your question. Hitachi Semiconductors manufacturers thousands of semiconductors and many have much smaller markets than this. A check on their website shows this contoller still available. In fact the most recent documents are dated 1998. Trust me they still know that this chip exists. And anyway, as long as there are customers for the controller chip then someone out there will buy the design and continue to manufacture the chip. After all AMD is still selling 8088 CPU variants for the embedded market. >HDLC controllers that >can barely run 2 T1s are not mainstream these days, so they may just >not want >to make it anymore. First of all, it takes one HD64570 per port, not two. The HD64570 is not a dual ported chip. And secondly the chip is perfectly capabable of running at a T1 speed and in fact much faster. It's all a pricing issue. SBS Communications (manufacturer of the WANic) is going to manufacture these cards until they can't sell enough of them to make a profit anymore - since they haven't said they are stopping making them, that's proof they are still profitable. If they start having trouble sourcing the controller from Hitachi and there's still sufficient interest in the market then they will just approach another chip manufacturer and that one will buy the controller design from Hitachi. >If you think that Hitachi cares that there are "open >source" drivers in use in what they would consider a handful of >applications, >you dont understand the chip business. When cisco goes away, they find a >better use for their foundry. > Everything your saying seems to indicate that your the one that doesen't understand the chip business. For starters, a chip foundry that is set up for manufacture of this controller could never manufacture more modern chips without being completely razed and rebuilt. Secondly, chip foundaries don't operate like this. What happens is that Hitachi has made this chip before, so they know that their cheapest price point to make the chip is to set up a production run of, say, 15,000 devices that takes 2 weeks to run. So they simply monitor inventory and as long as the product is moving out of the warehouse, they simply wait until they are down to, say 500 devices, then at that time they schedule another run of the controller chip in the foundry. That run completes, the warehouse is replentished, and they then turn the foundry over to work on a different chip design for some other product. All Hitachi cares about is that there's customers still buying the controller, and as long as there are they will make the chip. Consider also that this is a general purpose synchronous serial controller and can be used in a lot of applications other than just T1's. What all this discussion is boiling down to IMHO is Imagestream - because they sell the WANic+driver package to ISP's that are attempting to build routers out of PC's. They have Linux drivers for the more expensive WANic cards so of course they have a vested interest in seeing the WANic 405 go away, because SBS allows any reseller to register with them to sell the WANic cards and when someone like Joe Blow Computers fills out the paperwork and start selling WANic405 cards on FreeBSD boxes as routers, Imagestream sees it cutting into their Linux/expensive WANic router business. The rumor that the WANic405 is going away is totally unsubstantiated by the facts and if it ever did then Linux would see immediate benefit because the FreeBSD OS would end up losing one more of the few cards that it has that supports a T1. So consider who it is that has a vested interest in spreading this kind of FUD. Ted Mittelstaedt tedm@toybox.placo.com Author of: The FreeBSD Corporate Networker's Guide Book website: http://www.freebsd-corp-net-guide.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?000401c153a7$409d9ec0$1401a8c0>