We live in a very dynamic industry, and spend a lot of time throughout the year talking with customers, analysts and other virtualization vendors, trying to get a handle on how the market is shifting, and how market needs are changing over time.
At this time every year this activity intensifies as market leaders put their reputations on the line and make their own predictions for the new year.
Embotics has been doing this now since 2007, and has a pretty good track record when it comes to our predictions. Although any market analyst will tell you that past success is no guarantee when it comes to predicting the future. The market usually finds a way to humble us all.
That being said, we have identified three powerful industry drivers that will significantly impact the market when it comes to virtualization management. These are:
- Increasing complexity in the virtual environment,
- Increased pressures for visibility into the virtual environment, and
- Lack of time within administrative teams
These drivers are not independent of each other and we believe that they will work together to create a challenging vortex or perfect storm for virtual administration teams.
Increased Complexity
Anyone who works in the “virtual datacenter” knows the complexity of virtual environments. They include all the traditional complexities of the physical datacenter with the added dimensions of mobility, increased volume of servers and wider mix of configurations. – But without the benefit of the standardization and management tools that are available for the physical space.
And of course, as virtual environments grow, by their very nature, this complexity increases making it more and more difficult to manage and optimize. There are lot of dimensions to this complexity. Configuration standardization and management are an example of one.
Standardization is the key to effective operation of any IT environment. Standard configurations make it easy to maintain, troubleshoot and predict overall performance. Configuration standardization is relatively easy to do in the physical datacenter, where provisioning can be tightly controlled. But can be more challenging in the virtual space where not only the VM can be configured differently. However, what’s inside the VM (operating system, applications and patches) can move off standard over time.
Most organizations use templates to enforce standard configurations to start with and a variety of tools and manual process to update and maintain them. But, without a means of automatically identifying drift this is a challenging task. For example, we frequently see customers who used standard templates and updating tools that are surprised to discover that a sizable percentage (sometimes well over a third) of their VMs were created manually, and consequently were out of standard to start with, while other VMs that were identical to start with had drifted significantly over time.
Increased Pressures for Visibility
While the virtual environment was a small percentage of the overall datacenter, few people outside the administrative team and their immediate management needed visibility into the virtual environment, and the admin teams could get what they needed through VirtualCenter or vCenter.
But also as virtual environments become a larger percentage of the datacenter, more and more of the traditional management “silos’” become involved, creating interdepartmental coordination and communication issues. This drives an increased need for information and real time visibility into the environment; especially when process change is needed. These non-admin team stakeholders are not trained on VirtualCenter or vCenter, and consequently have to involve the admin team whenever they need information.
Lack of Administrative Time
At the same time, very few IT virtualization teams have been able to make the investment in additional management and automation systems necessary to reduce their workload and improve their management capability, and are consequently already running flat out.
Add in the increasing complexity, and rising workload providing visibility to non-admin stakeholders, the “perfect storm” forms. Admin teams are already overloaded, and yet additional workload from the first two drivers are inevitable. This means that if nothing is done about this, some form of “crash” is inevitable.
How will this impact the need for Virtualization Management?
The number one reason why customers come to us is to help virtualization administrative teams work smarter (not harder), by providing better, and more actionable information in real time, and incorporating policy driven workflow to take either simplify or automate a variety of day to day tasks and activities.
We are already seeing an increased demand for management functionality to manage the increasing complexity and improve the internal viability without the need to add non-administrators to VirtualCenter or vCenter.
These market drivers discussed are making it clear to a lot of organizations that VirtualCenter or vCenter virtualization management systems by themselves are no longer sufficient to ensure control in their growing virtual environments.
In 2010, the market reacted to this by complementing the VMware management platform with point tools to address specific pain points that were starting to impact virtual environments; discovery tools, capacity planning tools, performance management tools etc.
But the increased learning curves, the need for integration and consistency started a movement away from point tools towards integrated VM management systems.
We predict that the market will see a continuing this trend if not accelerate it. To be effective, this has to be true integration, not just a collection of tools that are stitched together to look like they are integrated, as these will ultimately fail as the complexity rises.
We also predict that we will see new features added to best-in-class virtualization management systems. Features such as VM monitoring, configuration and change management, lifecycle management, capacity/performance management and self service will become key components of any VM management system, as will some level of automation and workflow. All these features are needed to manage the increasing complexity within virtual environments.
As virtual environments increase in size and increase in impact to the datacenter, the need for coordination and process change across all the datacenter management silo’s and support groups also increase. Providing these untrained stakeholders with access to the VMware management platform is something that most virtualization groups would prefer not to do. Yet, they need the real-time access to information on the aspects of the virtual environment that involve them, and in some cases need to manipulate the VMs themselves. For example, support groups or development teams frequently need to re-boot VMs.
The answer to this dilemma is some form of service portal that provides access to information, as well as access to VMs themselves without the need for training for vCenter access.
Integration will also be a key requirement in the coming year. Virtualization management solutions that integrate and automate all these features and disciplines within an enterprise architecture will be essential. Standardization and automation is the only way to move from the tactical implementation of virtualization to a more strategic one. It’s the only real path to the private cloud, and the only way to get true efficiencies.
For example, when we looked at the increased complexities of the virtual environment, configuration management was used here as a discussion point. It’s relatively easy to establish a policy that says “all VMs must be created from an up to date and approved template”. In fact a lot of organizations have a policy like this today. The question is, how do you enforce it?
Given the current admin team workload, manual inspection adds more work and is not likely to be consistent. Basic policy-based automation is needed, something that checks new VMs coming into the environment and highlights ones that don’t meet the standard, leading to more consistency and much less work.
That said, management systems will also have to become easier to deploy and use: Administrators simply don’t have the time for the traditional complex implementation and high learning curves associated with IT management systems. And, the shortage of experience and trained administrators is not helping either.
The challenge to virtualization management vendors will be to marry simplicity of use to full featured and powerful integrated systems. But, we believe that the impact of this driver is so great that virtualization management systems will either become intuitive or they will fall by the wayside.
All in all, 2011 is shaping up to be a watershed year for virtualization management…
