Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: Mary Cluney who wrote (140439)	7/31/2001 3:40:21 AM
From: pgerassi	Read Replies (1) \| Respond to of 186894

Dear Mary: Actually in ERP, database queries are even more frequent and require even higher I/O bandwidth. I have designed, built and maintained such systems. The number one bottleneck for most database systems is I/O bandwidth and access times. Typical enhancements are more memory (in many situations, more memory can speed past a faster CPU system with less memory), SSD (Solid State Disk), low latency disk (high RPM and ultra low seek times) and more disk (more spindles and more channels). Ditto for Decision Support Systems, Data Mining, and OLTP. In telephony apps, like those that Amtelco builds, downtime is both costly to have and costly to repair. Some of these server types are sitting on a mountain top somewhere very remote running a microwave relay station. They have to fail in small increments and still be usable even with many failures. You have to be able to turn off failed components or bypass them. You need to be able to replace failed components while the system is running and bring it back to full performance. It must be able to be managed remotely. If a power failure occurs, it should be able to start a backup power generator and worst case, restart when power is restored. Reboots are to be greatly deprecated. The boards and chassis in such systems have a ten year MTTF and MTTR in minutes. Clustering in such systems in not a nicety, it is a must have. Hot swap is another such requirement. These systems are $30K and up per enclosure and are run by a pair of Celerons, bottom end P2s or P3s. In some places the only real way in is by helicopter and only when the weather is cooperating. They are designed to operate in conditions that a typical large server would throw up its cables and surrender. They are robust and very tough. Either of these types are very expensive yet modest in computing power for their size. Pete