When purchasing technology, a Service Level Agreement (SLA) refers to the common understanding about the services, responsibilities, and provides specific IT solution provider commitments to a customer. It’s the part of the contract that outlines exactly what a customer will receive from their solution provider and what the provider promises to do (or pay) if those terms aren’t met during the life of the agreement. SLAs are particularly important when using cloud-based technologies since the “technology” companies are buying is actually delivered as a service. Given that, it’s important for customers to clearly understand the parameters of that service. While the search for the perfect cloud provider can be challenging, finding common ground when evaluating an SLA can be just as difficult, and in many ways just as important. With that in mind, there are some key items to look for in an SLA.
- Uptime Guarantees
A site that is down may as well not exist. In fact, it’s often worse than not existing because service interruptions anger customers and can dramatically hurt a company’s reputation. As a result the word “uptime” has become part of every SLA conversation. But what is considered an adequate or appropriate uptime SLA? And what is included in uptime guarantees? In the current cloud space, an increasing number of IT solutions providers promise 99.5 or 99.9 percent uptime. It is important to take this benchmark with a grain of salt and take the time to understand its meaning. How is uptime calculated and over what period? What is it exactly that will be available 99.5 percent of the time? Agreeing to an ANNUAL 99.5 percent uptime SLA means 7.2 minutes of downtime daily over a one-year period (525,600 minutes in one year x .005= 2,628 minutes a year/365 days= 7.2 minutes a day). Pretend a cloud provider has a 3 hour outage during a given month and remains functional during the remaining months of the year. If the SLA is calculated annually, the provider is keeping their promise and no credits are due to the customer. However, if the SLA is calculated MONTHLY, then a 3-hour outage would entitle the customer to a credit, as it exceeds the maximum allowed monthly downtime.
Once the uptime SLA percentages and timeframes are well defined, it is important to know exactly what will be available. Is it the service, or the interface to the service? Troy Garrison, who leads Cloud Experience for Verizon Terremark believes that SLAs should cover multiple areas of engagement, including the basics, like power and network availability but more critically, the automated coordination of virtual resources and business’s specific needs and objectives, such as provisioning timeframes. Even if you have it all covered, what happens if something stops working? In the event of an outage, the service provider should pay the customer, usually in service credits. - Performance Consistency
Businesses can only realize the benefits of cloud computing if they can predict application performance- meaning acceptable performance levels in the infrastructure, like network speed and availability.While not all applications need high performance, John Considine, Verizon Terremark CTO states “it is fair to say all applications need consistent performance.” Managing performance can be extremely difficult. Cloud providers can, and should, carefully provision resources and offer a solid infrastructure to their customers. Analyzing the provider’s supplying procedures as well as their commitment to performance can be a deciding factor between a provider that can meet your needs and expectations, and one who cannot rise to the occasion. - Shared Responsibilities
In any business partnership, each party is asked to bring something to the table to ensure its success. Think of the service provider as an extension of the IT organization within the enterprise and structure the SLA to create a clear understanding of the provider’s, responsibilities and commitments. Once agreements are signed, vague promises and assumptions may hurt the success of this newly formed relationship. - Multi-Redundancies
The ability to commit to a certain level of uptime and performance comes from the provider’s physical infrastructure -- the environmental elements within the data center, as well as the people who manage the infrastructure. The more experience the cloud provider has in managing a data center and virtualized environments, the better performance and response the customer will receive. It is important to take the time to learn how the facilities that house the cloud infrastructure are built and how they operate. A lot of cloud providers are very transparent about their data centers but some are not- which should raise a red flag. At Verizon Terremark, prospects are always invited to tour the facilities and see the infrastructure, operations, security, and redundancy systems with their own critical eye.
Cloud infrastructure and technology is only as good as the services behind it and clearly outlining those service commitments is the role of the SLA. As companies go through this process they should actively inquire about uptime, resources, etc. and make sure the SLA covers technologies, services and availability levels that best fit their businesses’ needs. Cloud technology vendors that have the infrastructure and experience to serve even the most demanding client requirements will not shy away from committing to them in writing.
