Key Cloud Questions, Part 1: Service Guarantees
When you are considering putting your precious business applications and even more valuable data into the cloud you absolutely must make sure you are asking the right questions during the process of figuring out what to do. You also need to be equipped to know how to understand the answers you get. This is the first article in a series that we aim to release over the coming months that will look at the key questions that need to be asked by businesses that are making decisions related to running their business IT in the cloud. This article is going to focus on perhaps the most important question of all, one that if not considered carefully, will have far-reaching implications that will adversely affect your ability to conduct your business.
First, let’s summarise the key reasons/motivations for moving into the cloud. These are:
- to avoid having to rely on your own ability to support your own IT infrastructure, since your core business is probably not IT infrastructure management;
- to reduce cost or, more accurately, to increase business efficiency; and
- to gain what is effectively a brand new modern state-of-the-art IT infrastructure with very little (if any) capital outlay.
In addition, as a result of the cloud move, you will enter into a relationship with a cloud provider, one that will be a key relationship for your business, and one that needs to provide benefits beyond the motivations noted above.
Now because the provider relationship is so crucial, the process of selecting a provider needs to be robust. During the selection process you will discuss at least three major areas with prospective providers:
- Function (does it do what I need it to do);
- Delivery (how and where do I get it safely, and will it be there when I need it); and
- Consequence (“what if” scenarios, whether there are any operational changes, finance related matters, etc.).
For the sake of this article, let’s assume that all prospective providers are functionally equal (by no means the case, but let’s just assume), and that the questions around consequence are also sufficiently answered by all. That leaves the area of delivery. Delivery covers a number of topics, including continuity, security, connectivity, and more, but our focus here is specifically on service guarantees. Does the provider offer any kind of guarantee that the service will be available for you to use when you need it? In other words, how much does the provider back their service?
Whether a service is available for you to use is obviously important and of equally obvious importance are the technical measures taken to ensure that’s the case. However before the technical merits of a provider can be discussed the most important question to ask is this: Can the cloud provider guarantee their service and to what extent? Put another way, how can you as the service consumer be sure the provider can support your requirement that when you need to work, you can?
It’s no good being sold the dream of a wonderful life in the cloud if the provider isn’t able to provide you any assurance that their service will be available when you need it most. There is a significant difference between being able to provide assurance through the use of technology, and being able to provide it contractually. What you need is both, but crucially you need to know that the provider’s ability to execute both is real.
Here are a few of the main reasons why a provider will find it difficult to provide guarantees for their service:
- They use someone elses infrastructure to provide the service, e.g. they say it’s their service but they host the service on Amazon/Google, over which they have very limited control;
- They lack diversity in supply, for example they use a single carrier for core network connectivity, or a single data centre facility;
- They are a reseller of a service from a provider for which either 1) or 2) or both is true.
If at least one of the above is true, the provider cannot guarantee the delivery of their service. Alternatively, if the provider does say they guarantee it, they have accepted a significant risk on behalf of their customers, i.e. that it’s okay for their customer businesses to be offline if a 3rd party service that is critical to the delivery of their service fails for a while.
Now you might ask at this point, will the 3rd party service fail? Unfortunately the answer is yes, and probably on more than one occasion (see below for examples). That there will be a failure is not the real issue that we’re getting at here though. The real issue is who is in control of the circumstances in recovering from that failure. In the list of the three reasons above, the inability of the provider to guarantee their service is directly related to their inability to control the means of recovering from that failure. During the failure they are relying on someone else to fix their problem. This lack of control on the part of the provider under consideration is extremely important to recognise because:
- your business cannot operate during that time;
- neither you nor your provider have control over recovery timeframes;
- go back to 1.
If the provider had the required control, they could guarantee the service since they own the recovery mechanisms and actions. That’s really important because unplanned outages are by definition unplanned: they could happen at the worst possible time for your business. What you need to know is that when it happens, you aren’t left waiting for some other organisation with which you have no relationship and who therefore doesn’t even know your business is affected.
So before you sign up with a cloud provider ask them the right questions – where is their cloud infrastructure hosted? Who owns it? Is it built on someone else’s equipment? What are the service levels? What are the incident response times? What happens if they have a problem that stops me doing business? How long will I have a problem for? If they don’t provide any guarantees around the service levels that you receive, walk away. If they do provide a guarantee, assess whether that guarantee holds water, because even though you might have recourse through the terms of your contract, the risk is still real, and your business doesn’t need to take that risk.
_____________________________________________________________________________________Here are examples of outages for Google/Amazon. To reiterate, this article is not taking issue with organisations such as Amazon/Google having outages. March 2013 http://www.zdnet.com/amazon-web-services-suffers-outage-takes-down-vine-instagram-flipboard-with-it-7000019842/ Jan 24, 2014 http://mashable.com/2014/01/24/gmail-down-012414/ Mar 17, 2014 http://techcrunch.com/2014/03/17/google-hangouts-and-gmail-chat-go-down/