Remote Benchmarking

[JPJeffery]

Hi

I've been researching Benchmarking software but everything I've seen seems to be installed locally. Is there anything that can run remotely.

The purpose? We're going to be rolling out some new PCs to our Fund Managers and want to be able to show that

A) their PC is faster than their old one and
B) whether or not their new PC has actually slowed down or not over time.

Spotlight on Windows doesn't quite seem to do the job we need.

JJ
[small][purple]Variables won't. Constants aren't[/purple]
There is no apostrophe in the plural of PC (or PST, or CPU, or HDD, or FDD, or photo, or breakfast...and so on)[/small]

 

[kjv1611]

If you can remote into the PC, then technically any local software then becomes "remote" software to you.

 

[JPJeffery]

Not sure how that helps. I need to run benchmarking against an array of PCs, without interrupting the end users, and by so doing monitor the performance of their PCs. Remoting to their PCs just isn't going to cut it! Remotely installing an agent should though...

Currently looking at something called LANSurveyor which looks hopeful.

JJ
[small][purple]Variables won't. Constants aren't[/purple]
There is no apostrophe in the plural of PC (or PST, or CPU, or HDD, or FDD, or photo, or breakfast...and so on)[/small]

 

[kjv1611]

without interrupting the end users,

Honestly, not trying to seem rude or anything, but how would this be accurate? If the user is running anything that uses any resources, it will not give an accurate reading.

What's wrong with having something or someone remote in and run a test after business hours, or whenever the end user is not in the middle of something?

As for LANSurveyor[/LINK], it sounds like it may be a nice program, but it doesn't seem to actually test the machines. It only finds the resource usage - tells you what's being used... if I'm reading it right.

LANsurveyor is proven, network and desktop management software. LANsurveyor makes it easy to map, manage, and report on your entire network. LANsurveyor is unique because it provides four essential functions in one cost-effective application: [HIGHLIGHT]automatic network maps, asset management reports, network monitor, and remote administration and distribution.[/HIGHLIGHT] LANsurveyor generates a map of your network automatically using several different methods, including ICMP (ping), NetBIOS, and SNMP. Maps can be printed or exported for display or editing in Visio or other editing programs.

Version 9.5 supports a Repository to collect information from workstations and servers running the Neon Responder client software and store the data on a regular basis in a database.

Well, hopefully you'll end up finding something that'll work for your needs. I'd be curious, though about using a locally installed testing app that can be run from a command line. Then that command line could be run over LAN via your choice of many different methods for sure... You'd just have to configure the settings for login info, IP address, machine name or whatever...

 

[JPJeffery]

Ah, I think I see the problem here. I've inappropriately used the word Benchmarking.

So let's pretend I never used it!

This is all about measuring performance while the PC is in use, and keeping historical data so we can compare new to old and prove (hopefully!) that their new PC is better, and hasn't got significantly worse over time (or if it has we pick it up before the user starts complaining).

So, running tests out of business hours wouldn't be useful.

JJ
[small][purple]Variables won't. Constants aren't[/purple]
There is no apostrophe in the plural of PC (or PST, or CPU, or HDD, or FDD, or photo, or breakfast...and so on)[/small]

 

[kjv1611]

Gotcha... So am I correct to assume that you're not JUST testing hardware, but the correlation of the software/hardware combination?

Either way, I'll be on the look out for other options.. Post back with whatever you end up doing..

 

[JPJeffery]

Yes, it's all about the end-user's perception of their PCs' responsiveness.

So they might complain about their PC being slow but the delay might actually be at the server (here in the office or somewhere in the internet).

JJ
[small][purple]Variables won't. Constants aren't[/purple]
There is no apostrophe in the plural of PC (or PST, or CPU, or HDD, or FDD, or photo, or breakfast...and so on)[/small]

 

[goombawaho]

You can't really compare because you can't know what the user is doing at any given moment.

Also, newer PCs can do more than older PCs so users may open 14 application windows now instead of 8, so performance would SEEM less improved on the new PCs because of that alone.

It's too relative to get good data. The only way to compare is running the same benchmarking tests (video, hard drive, CPU, etc.) with nothing else running on an old and a new PC.

 

[JPJeffery]

Reasonable points but it's unlikely the users in question will do anything different. They're a pretty focussed bunch.

And even if they do, I'd hope that we'd be able to demonstrate that they are actually using more apps.

JJ
[small][purple]Variables won't. Constants aren't[/purple]
There is no apostrophe in the plural of PC (or PST, or CPU, or HDD, or FDD, or photo, or breakfast...and so on)[/small]

 

[cdogg]

SiSoftware Sandra 2010 should work. It allows you to run tests remotely from another workstation or server:

http://download.cnet.com/SiSoftware-Sandra/3000-2086_4-10556571.html?tag=mncol

~cdogg
"Insanity: doing the same thing over and over again and expecting different results." - Einstein
[tab][navy]For posting policies, click [/navy]here.