Jump to content
RESET Forums (homeservershow.com)

Some Performance Data with my ML10v2


Joe_Miner
 Share

Recommended Posts

I posted some performance numbers on my ML10v2 at Setting up my ML10v2 for the first time, Part 6: Performance http://homeservershow.com/setting-up-my-6-performance.html 

 

 

and also tried to do a comparison between CPU's of the Stock ML10v2 and previous home servers

 

01-20151216_PassMarkComparison.jpg

 

 

Here are some highlights.................

 

02-20151215_DriveC.jpg

 

03-20151215_DriveD.jpg

 

07-20151216_NASPT_ThroughputComparison.j

Edited by Joe_Miner
Link to comment
Share on other sites

Fantastic:

 

1. Maximises the value in the ML10v2.

2. Clear, comprehensive instructions and benchmarks.

3. Useful long-term comparisons.

 

I'm recommending this blog and forum to anyone who asks ... and some who don't ask!

 

... HOWEVER ...

  • Like 1
Link to comment
Share on other sites

While I do a lot of theorising, as opposed to Joe doing a lot of implementation ...

 

1. The cost of XEON processors and server operating systems in a home system worries me.

2. Combining storage server and VM test bed functionality has both compromises and inefficiencies.

 

The costs of a base level HP Microserver and ML10v2 in the UK are £110 and £289 after cashback.

The costs of Windows 10 Pro and Windows Server 2012 are £152 and £400.  

A storage server needs very little processing and electrical power and to overcome the GBE single channel transfer rate of 125MBs ... 2 or more network connections.

VM's need fast disks and processors.

 

It seems to me that TWO servers would be a better choice:

a. A storage server: a base level Microserver with both network channels cabled up (the power differential is 35 v 80 watts, which will mount up over 5 years)

b. A VM testbed: say an overclocked Skylake consumer CPU (faster than the XEON) with an NVME U2 SSD (much faster than a pair of RAID 0 SATA SSD's).

(In the home context the VM test bed and the main workstation may be one and the same device.)

 

The loss of course is ReFS on the storage server, which might be an issue for some.

 

Hey its XMAS, somebody had to play Scrooge ;-)

Have a good one!

Edited by JackoUK
Link to comment
Share on other sites

The other things that jump out at me from the comparison tables:

 

1. How bad the Mediasmart servers were in comparison to the N36L.

2. Having nailed the balance of CPU power and electrical power with the N36L ... everything seems to be heading back to more unnecessary power for a storage server since.

 

My design for a Gen 9 Microserver would be:

- slot for SSD instead of the redundant DVD feature

- 15W Skylake CPU (the U series)

- 8-bay variant

Edited by JackoUK
Link to comment
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
 Share

×
×
  • Create New...