|
By: Jason
Cowie
As organizations quickly adopt
virtualization and begin the preparation for the cloud, IT faces increased
pressures to more effectively manage existing infrastructures while
continuously improving the speed and quality of service being delivered. Without operational changes and improvements
in process discipline, automation and integration of management tools,
complexity will continue to overshadow cloud adoptions. With cloud-based offerings being highly
dependent on virtual data center agility, traditional management disciplines
like capacity management are experiencing unusual turmoil as complexity is
outpacing capability.
At the forefront of this turmoil is
the notion that like electricity, it is nearly impossible for an individual
on-demand IT service to exhaust the supply of infrastructure. It is therefore not surprising that as more
IT services are delivered through the virtual data center, the traditional
definition of capacity management has been challenged. Capacity management, in its most simplistic
form, is balancing the demand for IT resources with the supply. Even if Moore’s
Law prevails and we are given exponential compute capacity with infinite
storage, capacity management must factor into organizational capacity (people),
processes and technology.
While IT has been focused on meeting
demand through the scale-out of virtualization, organizational awareness to the
underlying costs, risks and constraints of the virtual data center have
remained out of sight and out of mind.
In particular, the lack of conservation and resource optimization has
had a profound effect on operational costs and virtual sprawl and has led to the
proliferation of non-standardized processes and procedures. With some organizations now acknowledging
that legacy capacity management processes were ill equipped for the virtual
data centers, they are starting to embrace a more sophisticated approach to
balancing IT supply and demand.
Overshadowing the cultural aspects
of capacity management is the interdependency of IT services that span
clusters, hosts, datastores and virtual machines. Combine that with the
necessity to manage multiple stacks (storage, network, application, server) and
the lack of standardized processes and procedures, and it is not surprising
that virtual stall has become a widespread symptom of a demand-driven data
center. Given the complicated tasks of
balancing supply with demand, IT must implement new and innovative ways to
continuously monitor and optimize virtual assets. To effectively and efficiently manage
capacity, it is imperative that IT:
1. Implement a virtualization management solution: That encompasses
cradle-to-grave capabilities across multiple hypervisors and integrates with
existing processes and frameworks.
Performance, configuration and change management capabilities help drive
effective capacity management similar to the way in which automated policies
help drive lifecycle management.
2. Embrace
processes and procedures: From
the initial provisioning of the asset to its eventual decommissioning, put
controls in place to control what happens to your virtual assets through their
entire lifecycles. Define objectives,
measure how you are doing, analyze what is wrong, and improve as needed.
3. Scale out
IT operations though automation: Scale out virtualization management bandwidth
through automated, repeatable processes and procedures. Automate routine VI administrative tasks to
eliminate errors and increase predictability.
4. Optimize
virtual resources: Proactively identify and eliminate
constraining resources and right size virtual assets by monitoring utilization
(performance) values. Capacity
bottlenecks can be addressed by adding resources or by reconfiguring existing
resources to maximize asset utilization.
Adopt policies that help classify, manage and optimize virtual assets
according to criticality and SLAs (needs, demands and wants vary across the
organization).
5. Reclaim
and recover wasted capacity: Conserve virtual resources by disseminating cost metrics
back to organization stakeholders. Continuously
recover wasted capacity created from sprawl, idle/zombie VMs, orphaned VMDK
files and over-provisioned disk space.
6. Plan: Implement processes and procedures today that will continue
to support the business needs of tomorrow.
Capacity planning requires accurate prediction and forecasting of growth
trends and capacity requirements based on assumptions you have today. Verify assumptions and continually model
capacity; it can and will change frequently.
While
virtual data centers and cloud-based offerings have increased the complexity of
IT management, standardization of processes combined with a holistic approach
to managing virtual assets will help optimize infrastructure supply. Process automation will increase management
capacity (people), while cost visibility and chargeback will help curb
unjustified or unnecessary infrastructure demands. Perhaps Leonardo da Vinci was correct when he
said, “Simplicity is the ultimate sophistication.”
Jason Cowie is VP, Product Management for Embotics (www.embotics.com)
Only registered users can write comments.
Please login or register. |
