To: Technocrat who wrote (18657 ) 10/17/1998 9:46:00 AM From: KJ. Moy Read Replies (1) | Respond to of 29386
Kurt, <<<To get a high-performance arrangement meant you had to lay out a pile of money on super-fast machines to act as the master so as to compensate for the bottleneck. I do not consider this to be SAN, but just client-server computing like it has always been done. More often than not, you merely replicated the data with new disks when you added servers.>>> You are right. This configuration is not a true SAN. It is IMO will be the big bulk of implementations of FC in 1999. Disk makers call their new products SANs while adopting FC technologies. My opinion on your version of SAN(the true one<g>) will be adopted by CPQ and DELL shortly. I happen to believe that both scenarios(server-attached and share disks) will serve its purpose. In other words, there will be applications that will fit better with the true'SAN' and there will be applications that will fit better with the superfast server type. For someone or business who is not in the same proximity of the SAN will not be able to use the SAN the same way because of inherent problems in the network. I believe that's why FC was designed the way it is, channel oriented(little or no tolorent of errors), distance limit, some basic physical limitations of electronics, propagation delays, etc. FC can guarantee deliveries, allow flow control while ATM cannot(ATM makers chose not to do it). The scenario of business having their disk storage somewhere else won't happen for a long time. I haven't heard anyone even start talking about the network problem yet, let alone solving it. The buzz word 'SAN' is being used everywhere. It is readers beware and interpret its meaning on your own. Bottom line though, FC will be used either it is a 'TRUE SAN' or a 'HYBRID SAN'. <<I am not at all surprised that Sun has avoided announcing a FC switch deal.>> Agreed. I think they may have to soon. Do you think Ancor may have an outside chance with SUN with MKII 8 ? KJ