Servers on a budget: 4 Servers tested

How We Tested

  Servers on a budget

  Acer Altos R300
  Altech Apollo 64-R
  HP Proliant DL 320
  Xenon Radon Duo ASR

 Specifications
 How We Tested
 Sample Scenario
 Editor's choice
 About RMIT Test Labs

Interoperability
Which operating systems does the server support? What port and network interfaces dose it have

Futureproofing
How much expansion capability is available such as disk, memory, peripherals, etc?

ROI
Do you get your money's worth or are there better deals available?

Service
What warranties and service contracts are available? Can you get prompt service at a reasonable price?

We used WebBench v4.1 to test the server performance of each machine. For the Windows-based machines, we ran Business Winstone 2004 v1, Multimedia Content Creation Winstone 2004, and WinBench 99.

  • WebBench shows the relative performance as a Web server. We ran requests up to 70 client load with simultaneous transactions. The servers averaged between 5000 to 8000 requests per second from a 40-client load upwards which is quite acceptable for a server at this price point.

  • Business Winstone 2004 shows the relative performance when running Windows-based business applications such as Microsoft Office.

  • Multimedia Content Creation Winstone 2004 shows the relative performance when running Windows-based multimedia applications, such as Lightwave and Adobe Premiere.

  • Winbench 99 provides a facility to test subsystems such as disk and video; we ran the disk subsystem tests on these servers.

    • While most servers are not generally expected to be used for business applications, we thought running these tests on these hybrid machines -- that is a cross between a dedicated server platform and a decent desktop -- would show their relative performance at the very least. Especially look at the Altech: it is an AMD Athlon 64 desktop performance machine in a server chassis. None of these servers had SCSI disk drive units, all were either SATA or IDE type drives; some had a level of RAID functionality.

    We could not run the Windows-based tests on the Xenon platform as it was running Linux. Multimedia Content Creation Winstone 2004 continually failed when run on the Altech server, therefore we could not record a score for that machine, likewise Business Winstone 2004 failed each time on the HP and we could not log a score.

    Overall the test results were quite good. Or particular note are the Altech results. While not really a server, it goes to show what a decent performance-oriented machine can do under the WebBench server type loads. The Altech achieved almost the same score as the Acer server, certainly never falling more than 500 or so requests per second across the whole test. The Altech also performed very well in the disk and Winstone tests. It was also the only AMD processor-based machine in the review and was running Microsoft Windows 2000 Advanced Server.

    Interestingly the major differences in performance between the Intel-based machines were perhaps due to their operating systems. The Xenon was running Red Hat Linux, the Acer was running Microsoft Windows 2000 Server, and the HP running Microsoft Windows Server 2003. The Xenon server's WebBench results, while quite consistent, were a little on the low side. This should not count against it necessarily; with a little bit of Linux and Apache tweaking, the user is bound to get results in the same ballpark as the other machines.

    The Acer results running WebBench were both high and consistent. The HP's WebBench results looked promising initially but once it hit a 40 or 60 client load it dropped off dramatically, perhaps this was due to the operating system more than the server itself.

    < Prev 1 2 3 4 5 6 7 8 9 10 Next >
    Advertisement

    Print feature brought to you by Adobe

    Talkback 0 comments

    Back to top

    Featured