Whether you have presided over many outsourcing agreements or undertaking your first endeavor, the requirement for a strong service level agreement is clearly evident. There are several instances where the IT professional negotiates service levels from the simple, such as off site tape storage turnaround, to the complex, such as with a major application integration project. Service level agreements are no less important with IT infrastructure outsourcing deals. In fact, in some cases they are more important.
With diversity in hardware configuration, the days of simple five-9’s on mainframe availability is no longer the only solution. You as an IT professional may have some ideas about service level measurements, but not nearly the depth and breadth of providers. Providers are very much in the diver’s seat with real benchmark data. They have the latest experiential data with their own data center operations. In the case of top tier providers, this data can incorporate multiple hardware platforms with varying degrees of complexity.
The Basics
A basic service level agreement is pretty straight forward, and you can likely find enough ideas to create one on the internet. You will cover the usual suspects: availability of computer hardware, turnaround time for desk-side support and perhaps even service desk answer time. There are several factors in the basic elements of service levels that you will also consider. These are things like number of office locations and geography, hours of operations, critical services, and you might even provide VIP service to import executives.
Beyond the Basics
Perhaps the most misunderstood service level element is that of the relationship between hardware architecture, systems and availability. In today’s data centers, there is a mix of architecture and systems, some proven and some trying to make a case for production work. To the degree the hardware is configured to include redundancy and compute power for a particular task will dictate appropriate service levels. In most cases it is advised you consider multiple levels of availability depending on the criticality of service and the budget for hardware.
Then comes the question, “What’s available, the server or the application?” Providers are very willing to step up and provide solid availability on computer hardware. Availability to the application level is a more precise equation. You will see these incorporated into service level agreements from ASPs or when you are being provided a “service” such as MS Exchange. You will also see this in agreements where the scope of work includes applications support in IT outsourcing agreements.
There are several other areas that require careful consideration. Telecommunications is tricky around availability and response time. Providers are quick to dismiss third party provisioning, but it should not be overlooked. Desk-side should include a number of considerations: geography and locations, technology and refresh, number of images, desktop or laptop, etc. Even IMACs, whether they are around desk-side or data center, the types, critically and frequency will dictate the most effect approach to a proper service level agreement. Service desk should consider more than just time to answer. With each call to the service desk, there is an opportunity to measure client satisfaction by survey, and it should have a service level associated with it as well.
Final Thought
The decisions you make regarding service levels will impact the providers staffing and other allocation of resources, including systems and software tools. This will of course be in direct correlation with the price. The provider comes armed with lots of experiential data regarding service levels and resources necessary to achieve results at varying levels of service. The balance between required or desired service and effect on cost can be a challenge.
