Skip to navigation
Enterprise

11 golden rules for virtualisation

Posted on 5 Aug 2010 at 16:34

8) Lie about the money

Free virtualisation software can achieve amazing things. Perhaps the most amazing is to drive your board completely round the bend – the way basic hypervisor code is often given away for nothing, and then one or another utility turns out to cost £5,000, is the cause of immense angst inside projects.

The virtualisation vendors think of this “upsell” as a good thing. The reality is, it’s an Appraisal Meltdown Device. So rule eight says: do not admit when you’re about to use a free tool.

Keep the freeness to yourself, for your prototyping period; cost the whole approach based on getting all the bits you want, as a lifecycle view, even in the initial project-scoping conversations.

9) Watch the big guys

So much of what’s offered in the virtualisation marketplace is there to turn this architecture into a must-have purchase decision for large corporates.

Jargon buster

Guest: A virtual machine, incapable of running without a hypervisor on a host of some kind.

Host: A machine with virtual machines inside it.

iSCSI: The storage network standard that uses Ethernet to join together dedicated disk farm machines to equally dedicated compute machines.

So we see tonnes of effort directed at producing mobile guests – not VMs on your phone, guest machines that can move from one host to another without anyone using them realising it’s happened.

This is how we come to have snapshot-based backup (so each backup takes only a few seconds) and central management tools that quote collections of guests up into the many hundreds, all visible from one software interface: tools that don’t really kick in until you have a truly large network.

There tends to be an incredulity gap between large and small LAN deployments – the big boys can’t understand how small networks get by without massive management toolkits, and the small players can’t work out how on earth the big guys spend so much money. To make good use of a virtualisation tool, you have to put some time into understanding the world it’s most commonly sold into.

10) Upgrade the other bits

You may well find your NAS is too slow to get through a VM backup in time, or you need to spend money on a Layer 3 switch and someone to set up the IP config for you in a usable way.

So, in the course of that virtualisation project you end up blowing £10K on a chunky Cisco switch, a few iSCSI disks and a couple of external storage arrays.

server

Notice we’ve slipped in another frequently encountered hurdle by referring to extra storage arrays in a server. If you’re used to a single channel of SCSI with lots of gigabytes hanging off it then prepare for a multiserver virtualisation project to expose the performance shortfall of that architecture. Virtual machines love disk I/O.

This is equally true for smaller businesses. Once you start looking at making a server with six or eight separate storage devices, each one at least mirrored or showing some other type of fault tolerance, the costs soon crank up even based on eSATA controllers and drives.

11) Don’t virtualise file servers

File servers do one simple low-processor task with lots of disk capacity, but the best servers to virtualise need little disk space while performing lots of complex jobs.

Most often when the term “virtualisation” is used referring to file servers, the speaker is talking about having file servers hosted, not “virtualised”.

Big players may say they’ve broken out of this “don’t” and “it’s all virtual” – they’re using a SAN as a giant pool of disks and their version of a “file server” just brokers file requests between PC LAN on one side, and SAN HBA connections on the other.

Author: Steve Cassidy

1 2 3
Subscribe to PC Pro magazine. We'll give you 3 issues for £1 plus a free gift - click here
User comments

The 12th Golden Rule

This was a very interesting article, however I think there is a very important first step that was missed out which is to discover, map and create an inventory of your network resources first.

A layer 2/3 discovery tool, such as WhatsUp WhatsConnected, will identify not only the physical server that you have deployed in your environment, but how everything is connected down to the individual port. This process is actually an eye opener for many organisations, because you can discover pieces of unaccounted hardware, or even inter-device connections that you didn’t even know they were there.

Once you have a hardware inventory in place, you can use it to document your network for auditing purposes or simply to try to reduce costs by repurposing or moving to the cloud underutilised resources. And since networks are “living entities”, don’t forget to re-run your discovery on a periodic basis (more frequently if you are part of a larger organisations, or if you have distributed network locations) and export this information into other tools such as Visio or Excel for continuous use throughout your organisation.

By Ipswitch_WUG on 6 Aug 2010

@Ipswitch_WUG

Thanks for the advert!

By milliganp on 9 Aug 2010

Very Poor Show chaps

If you have a product to advertise then advertise it, or propose that we do a proper comparative review. Don't stick an irrelevant comment on an unrelated piece of coverage: it makes you and your product look like unsolicited Chinese spam artists, and most likely does more damage than benefit. Talking about inventory in response to an article on virtualisation is a monumental non sequitur; I don't need to inventory my whole network before I virtualise my lone Exchange Server. By all means, mail me a link to an eval copy of your inventory product and I'll give it a look-over.

By Steve_Cassidy on 9 Aug 2010

It is recommended practice to inventory all your servers that are to be virtualised. If it's a lone server then not much point (in either performing an inventory or virtualising it) but when you're virtualising 20+ you need to get an idea of resources required, lun sizing, bandwidth, the list goes on... only saying this because if I did a proper inventory to start with then I wouldn't of had to rebuild my luns due to iSCSI disk latency.

However... you don't need software to do an inventory. Just a count of servers, services provided, how many users they serve and factor in room for growth.

Not much but still a valid step.

By JmLing on 9 Aug 2010

That would be why I suggested that people prototype their setup. You also mention "LUNs", which means you are in iSCSI world. That's two orders of magnitude larger than the target I have in mind for these rules. At the small scale, I'd be really quite scared if someone needed to run an inventory - that means they are not familiar enough with the pre-virtualisation machine to evaluate what their prototype delivers.

By Steve_Cassidy on 9 Aug 2010

Alternate 12th Rule

Perhaps a better alternate 12th rule is don't keep puting Virtualisation off!
Alternately virtualise all NEW machines, even if it's at very low (even 1:1) ratios. Virtalised machines are much easier to resize, move or repurpose than physical machines -and you're not increasing the future need for p2v.

By milliganp on 12 Aug 2010

Leave a comment

You need to Login or Register to comment.

(optional)

For more details about purchasing this feature and/or images for editorial usage, please contact Jasmine Samra on pictures@dennis.co.uk

advertisement

Latest News StoriesSubscribe to our RSS Feeds
Latest Blog Posts Subscribe to our RSS Feeds
Latest ReviewsSubscribe to our RSS Feeds
Latest Real World Computing

advertisement

Sponsored Links
 
SEARCH
Loading
WEB ID
SIGN UP

Your email:

Your password:

remember me

advertisement


Hitwise Top 10 Website 2010
 
 

PCPro-Computing in the Real World Printed from www.pcpro.co.uk

Register to receive our regular email newsletter at http://www.pcpro.co.uk/registration.

The newsletter contains links to our latest PC news, product reviews, features and how-to guides, plus special offers and competitions.