RE: [ga] According to Mike Roberts....

[Roeland Meyer <rmeyer@mhsc.com>]

Mike Roberts betrays his ignorance. A PC in a closet is exactly what it is.
Well, maybe a few PC's. It's called clustering or RAIC (redundant array of
inexpensive computers), the same idea as RAID, capacity AND
redundancy/reliability. The closet may indeed be appropriate as well, as
long as it has a couple of DS1s. Of course, each node in the cluster is the
capacity of greater than 150 VAX 11/780s (For those not familiar with hist.
of comp., 1 VAX = 1 MIPS = 1 VUP). 

Except for redundancy issues, one PC has more than enough umph to run root
service for twice the entire Internet that we have today, with capacity to
spare. I've done 10 million unique page hits per day with just one such box
and g.root-servers.net ran on a single T1 for years. These are very
conservative numbers. We've had this discussion over four years ago.
zs01.root-service.net never exceeds 15% CPU utilization and it maintains a
steady 1.5Mbps throughput. It's a pair of 500Mhz Celeron powered BookPCs
(PCchips), using Linux clustering, for reliability.

You don't need big iron to do this stuff. The web-servers and database
servers place much higher demands on the systems. However, you do need
intelligent operational procedures and someone other than Forest Gump on the
switch.

|> -----Original Message-----
|> From: Jim Fleming [mailto:jfleming@anet.com]
|> Sent: Saturday, October 27, 2001 12:52 AM
|> To: Jim Fleming; jun@wide.ad.jp; liman@autonomica.se;
|> cgomes@verisign.com; demi@registro.br; enoss@tucows.com; 
|> plzak@arin.net;
|> klensin@jck.com; klaas.bouma@dgtp.minvenw.nl; mmr@darwin.ptv.ca.us;
|> karl@CaveBear.com; lyman@nexthop.com; andy@ccc.de; vcerf@MCI.NET;
|> apisan@servidor.unam.mx; lynn@icann.org
|> Cc: ga@dnso.org
|> Subject: [ga] According to Mike Roberts....
|> 
|> 
|> According to Mike Roberts....
|> http://atlargestudy.org/forum_archive/msg01113.shtml
|> "If you're the manager of a top level domain name registry, it's not
|> a pc in a closet time anymore."
|> "The Japanese government and the United States government are sending
|> cabinet level officers to speak at the November ICANN 
|> meeting about how
|> serious this really is."
|> -----
|> 
|> If ICANN is going to get serious, can someone tell Mike that his
|> e-mail address on the ICANN web-site is apparently wrong or
|> not working.
|>    ----- The following addresses had permanent fatal errors -----
|> <mmr@darwin.ptv.ca.us>
|> 
|>    ----- Transcript of session follows -----
|> 550 <mmr@darwin.ptv.ca.us>... Host unknown (Name server: 
|> darwin.ptv.ca.us:
|> host not found)
|> 
|> 
|> By the way, I was not aware that ICANN was running TLDs on PCs
|> in closets, but that is good to know. That will certainly have to be
|> upgraded, as people move from the "toy" IPv4 Internet to 
|> serious systems.
|> 
|> 
|> Do you use a 2002:<IPv4>:0000 prefix ?
|> 
|> http://www.dot-arizona.com/IPv8/IPv4/
|> http://www.ietf.org/mail-archive/ietf/Current/msg12213.html
|> 
|> JimFleming@Unir.com
|> http://www.unir.com
|> http://www.unir.com/images/architech.gif
|> http://www.unir.com/images/headers.gif
|> http://www.unir.com/images/address.gif
|> http://www.ntia.doc.gov/ntiahome/domainname/130dftmail/unir.txt
|> http://msdn.microsoft.com/downloads/sdks/platform/tpipv6/start.asp
|> http://www.ietf.org/mail-archive/ietf/Current/msg12213.html
|> http://www.ietf.org/mail-archive/ietf/Current/msg12223.html
|> 
|> 
|> 
|> 
|> --
|> This message was passed to you via the ga@dnso.org list.
|> Send mail to majordomo@dnso.org to unsubscribe
|> ("unsubscribe ga" in the body of the message).
|> Archives at http://www.dnso.org/archives.html
|> 
--
This message was passed to you via the ga-full@dnso.org list.
Send mail to majordomo@dnso.org to unsubscribe
("unsubscribe ga-full" in the body of the message).
Archives at http://www.dnso.org/archives.html