Document Sample
Virtualization Powered By Docstoc

Introducing virtualization to our

            Virtualization benefits
• Virtualization of the legacy servers
   –   Reducing hardware support costs
   –   Reducing the range of platforms supported
   –   Reduce our power and space requirements
   –   Standardize configurations during the virtualization activities
• Abstract the application from the hardware
   – Changes to configuration less intrusive and time consuming
   – Dynamically change configurations to meet business need
• Improve delivery through faster build times
• Snapshot and roll back opportunities
• Achieve more with less

   Benefits of virtualization of the
           legacy servers
• Reduce the derivatives of server models
  and configurations
• Reduce the range of components to be
  kept on site in case of failure
• Reduce the complexity of the server estate
• Improve the performance per watt ratio of
  the estate

Abstract user and application from
• Isolate the applications from hardware
  – Reduce the time it takes to upgrade or
    change elements of the infrastructure
  – New disks/memory/network and cpu need not
    take weeks to action
  – Avoid the hardware limitations – your server
    only supports 4GB RAM or one network card
  – Rebuild/re-load the operating system in
    minutes not days
 Introducing dynamic infrastructure
• Achieve more with less
• Remove the barriers to delivery
• Dynamically allocate resources in line with
  the business need
• Move workload around the infrastructure
  to improve server utilization
• Meet concepts of ‘Chinese Walls’ without
  separate infrastructure
           Key considerations
•   Project issues
•   Technical issues
•   Non technical issues
•   Avoiding the barriers to glory
•   Port the technology to your business

                   Project issues
• What are we trying to achieve with virtualization?
  – Server consolidation or virtualization
  – Service improvement through virtualization
  – Or is it all three?
• How we identify what is in scope?
  – By application or business unit
  – By server model type, size or age
• How do we do it?
  – Phased approach which evolves
  – Immediate next generation infrastructure from the start

              Project issues 2
• Understand the need to perform an application
  and infrastructure inventory
  –   What upstream/downstream application feeds
  –   What network requirements there are
  –   Databases involved
  –   Backups and high availability requirements
• Understand which systems are in scope
• How we go about the project
  – Virtualize what we have or provide new machines
    which we migrate to
  – Virtualize everything or focus on maximum number of
    servers virtualized per physical host server
          Project issues 3
• Define what we mean by a successful
• What would be deemed as a failure and
  what operational risks are there?

           Technical issues
• What specification is a standard virtual
  – 2GB RAM / 1 cpu / 20GB disk space
  – Quantify true system requirements
• Are the applications validated and
  supported for use on a virtual machine
• System configuration issues
  – Non standard requirements
  – Specific networks or multi casting
                      Non technical
• Who owns and funds the virtual infrastructure is it IT or the end user
    – Do we have a maximum virtual machine ratio to physical server?
    – Who determines when a virtual host server is ‘at capacity’
    – How do we fund investment to meet requirements?
• Who supports the virtual layer and to what level
• How do we manage virtual environment
    – Individual virtual machine ownership
    – Roll back and security
    – Support costs and charge back
    – What’s the per unit marginal cost of the basics, extra space, memory or
    – What constitutes a change in the virtual world?
    – What helpdesk or process updates are made to meet this new ‘on
      demand infrastructure?’
    – Issues of capex/opex and what happens to remaining servers – stop re-

    Avoiding the barriers to glory
•   Poor perception
•   Production ready
•   Inadequate planning
•   Ineffective consolidation
•   Delivery
•   Cost
•   Adoption

           Poor perception
• That one user/business unit can rubbish
  the concept throughout your business
• The need to identify, own and resolve
• The need to involve the user community
  with the technology
• The need to deliver what was promised

           Production ready
• Ensure the technology can deliver the
  minimum requirements for ‘production’
  – Resolve issues of ownership and support
  – Monitoring and backups are in place
  – Help desk calls/requests are handled and
    routed to the correct team
  – Issues identified and resolved with community
    feedback and reporting

        Inadequate planning
• What are the future plans and standards
  – How will we scale up what we have
  – What do we need in place to virtualize more
    of the estate or start on production
  – Ensure we are able to prove the concepts
    before offering it out both technically and in
    terms of process
  – How do we respond to and charge (or not) for
    specific user requests
         Consolidation issues
• Create a unified voice for virtualization
• Identifying, resolving and reporting issues
• Agree set virtualization processes and
  standards to improve delivery
     • If the server has 256MB RAM do we increase to 512MB
     • If the server has a 4GB C drive, do we increase to 8
     • Are we just virtualizing what we have or combining it with the
       opportunity to improve the configurations we have (where

• Do we roll back or just ‘get it working’
• A brand identity for your virtual infrastructure
• A team that we can call who’ll handle everything – not
  bounce issues around the universe
   – Pre-answers dealt with
   – How do I upgrade memory, add a processor/get more space and
     what’s the cost
   – What constitutes a change to the environment and who needs
• Ensure build process/upgrade processes have been
  updated for the virtual world
   – I’m not waiting 2 weeks for more memory on a virtual machine
     because your help desk process is wrong
   – How long does it take to rebuild or deploy a new virtual

• Consider all the issues around cost
  – Are we seeking to make a profit or deliver a more
    efficient platform at the least operational cost
  – If the server is to be virtualized and there is
    depreciation left – who pays?
  – What is the per unit cost of a virtual machine?
     • How do we charge for a virtual machine (capex/opex)
     • How do we keep it less than the cost of a physical server?
• What are the marginal costs of upgrades?
  – How much for more memory or storage and what
    would not be supported
  – How do we avoid the scenario where virtualizing one
    server might use 80% of our storage
• Brand identity – own the infrastructure, own the
  problems and the successes
• Establish the pre-answers
   – Establish the helpdesk and workflows
   – How do I request extra memory for my virtual machine
   – If the virtual host is full, what’s the lead time to more capacity
• What barriers real (or perceived) are there
       • How we can on board new users and business lines
       • How to maintain existing users perception of the platform
       • What improvements need to be made to the service
       • Understand new business and technical requirements
       • What backup facilities do we have and how do they deliver in the
         virtual world
       • What roll back options do we have and have we delivered on our
         promises in terms of dynamic and flexible infrastructure

Port virtualization to your business
• Update your processes in line with the
• What can we achieve with this new
  – Where can we take it from here
  – What are the possibilities
  – What do we want in virtualization 3.0?