Over the last few weeks a discussion at work has revolved around the use of ‘name brand’ [read: Dell, HP, Compaq, etc] vs. ‘Nameless’ [read: custom configurations without a logo] for servers. These servers will be used for a variety of tasks, but mostly web and database work.
For the last 4 years or so we have using the Nameless types - and in general they have preformed quite well. However, lately we have been having a succession of failures. Power supplies, unexplained shut downs, poor performance under stress and generally poor throughput. Now this could be due to application changes, the fact that we have had to expand greatly in capacity and our architecture does not scale as hoped - or our hardware is crap.
Of late - the thinking is we should go out and spend the extra money on some name brand servers from companies with quality reputations [I will skip over the capacitor issues of Dell that cost us weeks of downtime in the past]. Now in this economic climate - this is a rather big decision. I am sure we could waddle through our needs with our past methods of purchase, but maybe we could do more with less - better hardware = better capacity?
I am wondering if anyone out there has a strong opinion - either way - on this subject. What have you done in the past - what are you doing now? Was the money worth the performance? Was there a performance difference?