HP Refreshes MediaSmart Server line

[thiggins]

If you're looking for a good preview of the product, check out Alex Kuretz':
Review: HP MediaSmart Server EX487

Notable changes from original MSS:
- Intel Celeron vs. AMD Sempron 3400+
- 2 GB vs. 512 MB RAM
- Supports Mac OS Time Machine for backup
- Offsite/online backup to Amazon S3 service

 

[Unregistered]

Can you install a different operating system on these devices? I'm interested in installing OpenSolaris.

 

[thiggins]

Can you install a different operating system on these devices? I'm interested in installing OpenSolaris.

You might be able to if the OS you are considering has the drivers needed.

 

[br14]

How difficult is it to install an OS on this machine without a CD(DVD) drive?

 

[thiggins]

Depends on whether the MediaSmart's BIOS allows booting from USB. I have installed Vista and Ubuntu from a USB flash drive.

 

[br14]

Tim,
for Vista install, did you just make USB bootable with diskpart and copied the install files or was there more to it?

 

[MoralDelima]

Tim,
for Vista install, did you just make USB bootable with diskpart and copied the install files or was there more to it?

http://kurtsh.spaces.live.com/blog/cns!DA410C7F7E038D!1665.entry

 

[Kiwi]

HP EX470/5 do support jumbo frames

Tim,

I have a HP EX470 with the following Ethernet driver;

Silicon Integrated Systems Corp.
3/3/2008
V 2.0.1039.1110

It does support jumbo frames.

I cannot confirm if the new EX 485/7 support jumbo frames but I would suggest some more discovery on your part Tim, to be sure there isn't an updated driver that will support jumbo frames.

Regards,
Robert

 

[thiggins]

It could be that there is a driver that supports jumbo frames. But the product, as delivered and supported by HP, does not support jumbo frames.

 

[Kiwi]

EX485/7 numbers wrong in NAS charts - cache screwing them up?

Tim,

Looks like the numbers got messed up;
238.4 Mbps for 100 Mbps Average Write Performance
339.2 Mbps for 1000 Mbps Average Write Performance

Looks like caching is invalidating your results.

Also the Netgear SC101 number is higher than what is possible at wire speed for the 100 Mbps Average Read Performance - 66.3 Mbps versus 12.5 Mbps maximum

Reagrds,
Robert

 

[thiggins]

Looks like the numbers got messed up;
238.4 Mbps for 100 Mbps Average Write Performance
339.2 Mbps for 1000 Mbps Average Write Performance

Cached results are not "invalid". They are what happens when you have file sizes that are smaller than RAM in the client and NAS.

The reason you are seeing such high numbers is that the NAS testbed hardware has changed to using a Core 2 Duo system with 2 GB of RAM.
The previous system had only 512MB of RAM. This has the effect of making caching in effect for higher file sizes, which skews write results. Just look at the Throughput vs. file size plots for the tests and you will see the detail behind the averages.

This has been discussed numerous times in reviews and articles.
It is also why the NAS Charts all have this disclaimer at the top:

NOTE: As a reference, the maximum raw data rate for 100Mbps Ethernet is 12.5 MBytes/sec and 125 MBytes/sec for gigabit Ethernet. Throughput above these values is due to memory caching effects in the client OS and NAS under test.

Please read Why Cache Matters in NAS Performance