Home > Scale > Designing a Virtual Infrastructure that Scales: Part 1, Taking Stock

Designing a Virtual Infrastructure that Scales: Part 1, Taking Stock

December 7th, 2010

I recently had the privilege of presenting at the London VMware User Group meeting, #LonVMUG where I talked about some of the things you should be thinking about when scaling your virtual infrastructure.

I’ve turned part of the presentation into a series of posts, going through some of the aspects you should be considering when your virtual environment demands bigger, better, faster, more!

Part 1, Taking Stock
Part 2, Speak to the People

Part 3, Scaling for VDI

Part 4, Designing Thinking
Part 5, So, after all that…

Designing a Virtual Infrastructure that Scales: Part 1, Taking Stock

No-one is downsizing their VM infrastructure!

Your virtual hosting environment is the one place in your IT department where you are not looking at making anything smaller. The benefits of virtualisation in managing costs by making better use of resources, improving availability by increased uptime and simplifying and standardising on your IT processes means you need to virtualise more and more. (This could be the marketing slide for any number of virtualisation related companies / products!).

Many companies now have a virtualisation first policy for new servers and have already probably virtualised any physical kit they can. As hardware capability increases and virtualisation software matures, IT departments are now starting to virtualise the next tier of heavy resource utilisation servers that only a year ago would never have been virtualisation candidates. Cheaper and faster bandwidth means companies can also centralise servers which previously had to sit at remote offices and make them VMs.

Metrics such as % virtual vs. physical servers are now being gathered and IT Managers are being measured on how many servers they have virtualised with percentage targets increasing year on year.

That’s just servers, what about virtual workstations? VDI is just starting to gain traction in corporate datacenters. Being able to provision desktops on the fly allows for greater agility for IT and Windows 7 has come at just the right time as IT departments ignored Vista and are now upgrading from XP and looking at new ways to deliver applications to users.

Gartner itself says “Virtualization to Be Highest-Impact Issue Challenging Infrastructure and Operations Through 2015”

So, I think I’ve made the point that your VM environment is growing very quickly. I call it the march to host more VMs.

Unfortunately the pace of growth has meant that the virtual hosting infrastructure you built a few years ago may be starting to get a little unwieldy to manage.
Think back to where your virtualisation journey started, from a test VM on VMware workstation and thinking how amazing the technology was to maybe GSX and then onto ESX with local storage. You then added SAN/NAS connectivity and now had truly flexible VMs. Since then you’ve been adding and upgrading more and more hosts. Your journey may not even have been as long,  you may have started more recently with Hyper-V or XenServer.

Now is the time to pause if you can and take stock. Have a good look at your current environment and then zoom out and look at the big picture to plan the next stage of your virtual infrastructure because if you don’t you may find it running away from you.

Start with a simple numbers exercise. How many VMs have you historically been hosting, a year ago, 2 years ago, 5 years ago. Draw a rough graph and see what it looks like. Then look ahead (you may need a crystal ball for some of this but do the best you can). How many VMs are you looking to host in 12 months, 2 years 5 years?

Do you have a percentage target figure in mind for how many of your servers will be virtual? Have a closer look at what you don’t currently virtualise. Will this still be true in 12 months, 2 years, 5 years? Why are you not virtualising these VMs? Capacity, security, availability, performance? Will these still be issues in 12 months, 2 years, 5 years? If not think ahead and add them to your list. Are there some VMs you won’t need at some time in the future. Do you foresee yourself outsourcing any IT functions? Maybe you will move some apps into the cloud. What will you consolidate? What upgrades are coming up, how will you handle these differently?

What are your virtual workstation plans? Are you looking at hosted desktops at all? Do you have a percentage target figure in mind for how many of your workstations will be virtual? How will your staff numbers change in the future?

Do you have any major company changes happening: mergers, acquisitions, outsourcing, selling off etc?

Phew, that’s a lot of questions, but the idea is to get you thinking about it now and also thinking big.

Have a look at your historical graph and trend it against your future graph and try to get some understanding of what you may need from a virtual infrastructure in 12 months, 2 years, 5 years.

Then have a look at how you do things currently. Say for example you have rack mounted servers, 1 Gb ethernet, fiber channel, SATA disks, no DMZ?

Would this be able to scale to meet your plans?

How long can you progress along your current path before something runs out? How many Gigabit ports do you have left, how much disk space do you have available, how many IOPS, how much rack space, power etc. Will you need a DMZ, an IP load balancing system, more servers / workstations in BR?

You may find that by continuing to scale your virtual infrastructure the way you have been is just not sustainable or even if it is, think carefully about whether it is the smart way to do it.

Can you and should you be doing it better?

In Part 2, we’ll make the job easier by asking for some help.

Categories: Scale Tags: ,
Comments are closed.