Written by Chris O’Malley, CEO, Nimsoft
Chris O’Malley, CEO, Nimsoft outlines the considerations that should be made when considering a move to the Cloud.
Formalise communications, roles, service definitions and processes
The saying “Good fences make good neighbours” is very applicable to relationships with Cloud service providers.
Clear demarcation of services, roles and responsibilities is vital to these relationships. The more that’s left to subjective opinion, conjecture and guesswork, the more likely it is that misunderstandings, mistakes and missed opportunities will result.
Good “fences” for the Cloud are built by asking the right questions: Which services are ‘core’ and which are ‘chore’? Which will be off-loaded to a Cloud partner and which will remain in-house? What has front-line responsibility in the event of a problem? Who is the next contact in line if that person isn’t available?
Detailed process documentation, well-defined roles and clear accountability are all essential for managing an increasingly disparate IT infrastructure; one that includes partnerships with Cloud services providers.
This, by the way, is one factor that differentiates today’s high-value Cloud engagements from the less successful outsourcing engagement of the past. With the Cloud, IT can tightly focus what it off-loads and what its service-levels expectations are. So it can actually gain control, rather than losing it.
Start with concrete IT service requirements and definitions
Service level agreements cannot be standardised. Performance that is acceptable for one service will not be for another. And there cannot be grey areas. Everyone must understand the performance target.
This is why it is so important to have watertight service contracts in place―ones that detail deliverables and potential penalties―thus ensuring there can be no questions or misunderstandings in the event of service disruption. It is essential to detail what constitutes acceptable up-time. And if an outage is reported, it must be clear how quickly it is to be addressed.
One good way to set thresholds for Cloud service providers is to use benchmarks based on service levels historically achieved by the internal IT organisation.
Other considerations when creating SLAs include:
- Penalties – What service credits or compensation will be honoured if SLAs are missed?
- Realism – Can 99.999% reliability actually be guaranteed by the service provider? Does the business really need to bear the cost of premium reliability for every service?
- Accountability – Service Providers’ claims about service levels should be validated against reports of problems from end-users.
- Alignment with business requirements – Different departments, for example may require different service levels.
Stipulate on-demand service status updates
Irrespective of the SLAs, or any other service agreements that are in place, it is vital that the business has 24 by 7 visibility into the performance status of its Cloud services. The sooner the business is aware of an issue, the better it will be able to mitigate its impact on end-users and customers.
Many established Cloud service providers offer their clients online portals that provide access to real-time data on the performance and status of their managed infrastructure and services. Some also provide automated reporting and alerts. This transparency gives customers a higher level of confidence that business needs and SLAs are being met.
How should an organisation manage its Cloud services?
The use of Cloud service providers does not eliminate the need for businesses to monitor the performance and availability of critical services. On the contrary, it is still important to identify problems and hold service providers accountable.
So rather than relying exclusively on the monitoring and reporting offered by each Cloud service provider, it is advisable to implement a comprehensive solution for monitoring any and all externally provided services. Such a solution will help ensure the success of Cloud computing initiatives by measuring all key services and applications in real-time.
Ideally, this monitoring solution should unify the entire operational IT infrastructure, bringing both internally and externally provisioned services under the umbrella of a single IT monitoring and management process. This unified approach is the most cost-effective way to ensure that all IT services are meeting the needs of the business in terms of performance and availability.
