“It’s the end of the world as we know it, and I feel fine!” – R.E.M.
As I mentioned in my previous blog post, Windows Server 2003 was a hugely successful operating system, and even after Microsoft ended support for it on July 14th, there are still millions of servers running in production environments that are yet to be migrated.
Extended support is available, but incredibly expensive, just 100 servers can cost you up to $3.1 million in licensing for the first year, and that price will double in the second year. In the third year, the second year price doubles. It all starts to add up.
With this financial incentive for Windows Server 2003 migration, organisations have to face up to a number of challenges:
- For production systems, you do not want to risk extended downtime, performance degradation or worse. The project has to be planned and executed very carefully.
- IT Admins will have to complete this project, while making sure it doesn’t affect their day to day job – keeping the environment healthy and assuring application performance.
- Many of these workloads are physical servers, and are being virtualised for the first time. This means adding on several new virtual machines into your existing estate, and has to be done in a way that doesn’t degrade performance of the rest of the environment. Also, how do you determine how many new physical servers you are going to need? Are your applications going to behave themselves on virtual hosts?
- Licensing – If software is licensed per physical socket, how do you limit your workloads to specific host(s)?
- Performance – Deploying several servers at the same time on shared infrastructure can cause performance issues in the short term, impacting other workloads, as the entire environment changes very quickly in a short space of time.
Let’s do this! – Planning our migration
Microsoft’s Migration guidelines breaks down the Windows Server 2003 migration process into 4 phases:
- Discover –
- Define migration Scope.
- Discover any datacentre dependencies.
- Take inventory of your existing production environment and IT infrastructure, understand current utilization.
- Assess –
- What are the most critical applications that need your focus?
- Are you considering migration to a public cloud?
- Target –
- Upgrade your hardware where necessary.
- Decide which platform you will be using – server virtualisation or public cloud.
- Migrate –
- Deploy the virtual machines or applications in your environment or in the cloud.
So how can VMTurbo help?
VMTurbo can add value at every stage of the Migration Process
Once you drop in VMTurbo into the environment, in less than thirty minutes, the software has a complete picture of your entire virtual estate. The patented Economic Scheduling Engine is able to build out an accurate picture of Utilization across the virtual infrastructure. All of this is done with minimal configuration.
VMTurbo can help to prioritize your most important applications, and assign them a higher Quality of Service.
The Desired State is where you are assuring application performance while maximizing efficiency. With Software defined-control, you can safely drive up your VM-to-Host density without compromising on performance. By freeing up capacity, you can limit the money you’ll need to spend on new physical hosts (and the associated operational costs), licensing and support fees.
Taking your existing utilization into account, VMTurbo’s planning capability can determine if your existing infrastructure is sufficient, and if it isn’t, accurately establish EXACTLY what you need to purchase.
As you start to deploy your new virtual servers in the environment, you have to decide where to place them. VMTurbo can handle the initial placement using our Deploy capability, why disrupt your already healthy environment? Reserve capacity for deployments ahead of time, which also helps you to keep an eye on capacity.
Looking to make use of Azure or AWS? VMTurbo can identify the best workloads to migrate to the cloud.
With VMTurbo’s workload placement policies, you can take any licensing constraints into account, and limit certain workloads to a set number of physical hosts, and keep your license budget down.
Happy Datacenter = Happy Users
In conclusion, as consultants and System Administrators around the world are currently planning or in the middle of these migration projects.
Consider VMTurbo – before, during and after migration – with demand-driven control, you can stop monitoring and analysing alerts, and focus on more strategic projects (like this one!)