You may be ready to implement server virtualisation, but ensuring that each stage of the process is successfully carried out can be easier said than done. This 10 step easy-to-follow guide provides ten expert tips to ensure that server virtualisation is an exercise that is undertaken only once and provides a platform for future application containment.
With the assistance of purpose-built software tools they have engaged consultants to undertake the obligatory capacity planning exercise and armed with that information new hardware has been purchased and the first applications migrated into production. But is that really any way to implement what is truly a paradigm shift in the way applications are delivered to the business? Many organisation’s that have rushed to embrace this technology have found to their cost that further due diligence is required; that the third-party ‘Capacity Planning’ report tell them little more than they already know, certainly not enough to deploy the technology in anger. To ensure server virtualisation is an exercise that is undertaken only once and provides a platform for future application containment, the following 10 tips are provided:
1. Adopt a Strategic Approach
To avoid these pitfalls it is important that organisations take a more strategic approach: align project milestones with the hardware replacement program and proactively plan scalability of the server farm accordingly; scope and define a standard platform; implement a Proof of Concept but replenish the test environment; react to short-term business pressures by implementing tactical requirements strategically.
2. Align with the Business
Aligning with the business therefore means working ‘with the business’, either to overcome these constraints or where this is not possible to recognise them in the design process. It also means ensuring that the organisation’s business strategy and project success criteria form part of the design process:
o What targets for cost savings has it set?
o Does it have intentions to move to a 24 x 7 operation?
o Is the organisation looking to improve its Disaster Recovery capabilities or perhaps intent on implementing a ‘charge-back’ model?
o What Service Levels has the business associated with the applications that are to be virtualised and how can this be achieved?
3. Consider the Underlying infrastructure
A number of questions concerning the underlying infrastructure needs to be considered:
o What are the additional storage requirements?
o Is a Storage Area Network required: – fibre channel or ISCSI?
o How will the storage be presented: CIFS, NFS, Lun or raw device?
o What are the implications for: backup/recovery, business continuity, local & wide area networking?
o How will the training budget be affected?
4. Design a Platform for Growth
To realise these objectives, break the cycle of purchasing hardware to fulfil short-term tactical requirements and lay the basis for utility-based capacity-on-demand, organisations should resist the temptation of only provisioning sufficient capacity to meet short-term needs. Instead they should provide a ‘platform for growth’ – a platform that is scalable – with sufficient capacity to meet longer term needs and that can react quickly to the changing demands of the business.
5. Rationalise your Application Set
Having undertaken a capacity planning exercise the software inventory that generally accompanies it is typically overlooked in a rush to scope the server farm. However unless the organisation already has this information it is a good idea to review the inventory as a pre-requisite to defining a set of standard operating system builds and reducing both the set of applications / versions the organisation is supporting
6. Use Migration Tools with Discernment
In migrating applications to a virtual state, resist the temptation to jump in and make indiscriminate use of Physical-Virtual (P-V) migration tools, which may speed application deployment, but migrate with it a lifetime of patches, fixes and unknown registry changes – ‘rubbish in equals rubbish out’.
7. Understand the impact on Operations
The operational management of the new environment is another area that can typically get overlooked, surprisingly when the loss of an individual server has significantly greater potential business impact. A number of questions need to be addressed:
o How will the new environment be managed; what will be the impact on people and processes; will it fit within the existing management framework or will additional management tools be required?
o How will applications be protected; what service level agreements have been defined?
o What are the backup/recovery options and what disaster recovery plans are required?
o How will new applications be deployed and growth of the environment managed?
o What are the security implications; where are the risks?
o What additional skills / training will be required?
8. Security, Security, Security!
Security as always should be an area of particular consideration during the design and planning process; Virtual Servers and guest operating systems need to be secured like any other operating system.
9. Review the Options
Microsoft and VMware are clear leaders in this space with many subtle, and not so subtle, differences in their features and overall product set. Similar product offerings are also available from a group of smaller companies such as VirtualIron; XenSource and SWsoft. Organisations looking to adopt server virtualisation technology should make themselves aware of the various merits of each vendor before making their platform decision.
10. Engage with a Partner
In order to ensure consolidation is an exercise that is undertaken only once and provides a platform for future application containment, organisations should engage with a partner who will take the time not only to undertake a capacity based assessment of their server estate, but will understand their complete environment