In the long run, virtualisation may result in physical infrastructure consolidation and efficiency, but along the way engineers are running into their share of challenges.
Those challenges may explain why 98% of businesses are using some form of virtualisation, but most don't go past virtualising 30% of their servers. Overcoming these issues will be key to implementing large-scale virtualisation.
Josh Stephens, head geek at SolarWinds and a SearchNetworking.com Fast Packet blogger, lays out the top five virtualisation problems and what you can do about them. Among these issues are virtualisation backup and recovery, VM sprawl, virtualisation capacity planning, VM stall and becoming a private cloud.
Virtualisation problem No.1: Virtualisation backup and recovery
We often hear the term “safety first,” but when it comes to virtualisation, we frequently put backups and disaster recovery on the back burner. That's got to change.
Backups in the world of server virtualisation and virtualised infrastructure are quite a bit different than in the traditional world of physical server technologies. While most backup vendors now offer VMware Consolidated Backup (VCB) and Virtual Disk Development Kit (VDDK) solutions, you may find that they’re either a paid add-on or require an upgrade to their standard package. Because this technology is a lot different from a backup perspective, several new companies have evolved to provide backup solutions for server virtualisation. Check out companies like Veeam, CommVault and PHD Virtual for next-generation solutions in this area.
Read more about how to address virtualisation backup and recovery in this Fast Packet blog.
Virtualisation problem No.2: VM sprawl
Virtualisation technology can be easy, cheap and effective -- three features you rarely find together. But because it's so easy to provision Virtual Machines (VMs) with the press of a button, a virtualised environment can get out of control with VMs popping up all over the data centre and very little overall management. This is what we call VM sprawl. This doesn't occur in a physical environment where one must order a physical server, wait for it to show up, and then find power and space for it. There are three signs to look for when it comes to VM sprawl: VM orphans, zombie VMDKs and Orc VMs. Wonder what these are?
Read Josh's Fast Packet blog that lays out the signs of VM sprawl and strategies to solve the problem.
Virtualisation problem No.3: Virtual capacity planning
Network capacity planning is one of those activities most engineers know they should do, but seldom find the time or the resources. Yet when it comes to virtualisation capacity planning or planning for SAN management, there is no time for procrastination.
Capacity planning is the art of looking at trends in performance data to predict future capacity needs and then building budget, acquisition and deployment strategies around that data. It’s super critical when it comes to virtualisation because the resources are so dynamic. In fact, a lack of capacity planning can lead to VM sprawl. As part of virtualisation capacity planning, one must understand CPU, memory, disk space, disk I/O and network bandwidth.
Read more about virtualisation capacity planning in Josh's Fast Packet blog.
Virtualisation problem No.4: VM stall
Many companies are hip deep into virtualisation, and 30% or so of their servers are virtualised. Now it’s time for them to start thinking about moving to the next level and placing core applications onto VMs. The problem is that administrators of those applications and databases may drag their feet if they don't trust the virtual environment. Welcome to VM stall. Often VM stall is due to a lack of faith in performance, so engineers will have to go about proving out their virtual environments.
Learn about how to avoid VM stall by proving performance in virtual environments.
Virtualisation problem No.5: Building a private cloud
Once engineers move comfortably into the world of virtualisation, their bosses want to know when they'll be rolling out that private cloud. The problem is many engineers -- especially those who are still living at the 30% virtualisation mark -- don't even know what it means to transition from a virtualised environment to a private cloud. In fact, the definition of a private cloud changes daily, but there are three elements that must exist in any cloud: elasticity of resources, ease of provisioning and accounting of use.
Read more about becoming a private cloud.