These days it feels like the cloud is everywhere. We use it to store our music and to watch our movies. We host our applications in the cloud, and we expect them to be accessible always and everywhere. Lost in the ubiquity of the cloud is a very simple fact: the cloud depends on connectivity.
The dependence of cloud on connectivity is obvious when you step back and ask the question: What is the cloud? A simple answer is: The cloud is a set of on-demand compute and store resources, usually housed in a data center, that are accessed when needed to provide useful services. In short, compute, store and connect combined for our benefit.
Connect to and From the Cloud
Unless your whole business is in the data center, connectivity services are required to get data in and out of the cloud. Most of the time the internet is the vehicle for that connectivity. However, the internet has become so prevalent that we sometimes forget it exists. As a result, people also forget that internet or other connectivity is needed to reach the cloud. It’s OK for casual users to be oblivious to how they reach the cloud, except of course when they burn all their mobile data watching cat videos on YouTube or House of Cards on Netflix. For business users, cloud connectivity becomes essential for their mission-critical services.
Large cloud providers are recognizing the importance of connectivity by providing for efficient direct handoffs between data centers and certain communications providers. By providing these direct connections, cloud users can access the cloud efficiently at Layer 2. Cloud resources can then be used natively, as if they were directly part of the user’s network.
Discerning cloud users can now optimize their use of clouds by securing private connectivity services to access their chosen data centers. The performance of that connectivity is assured by the service provider, and the cloud facilities are assured by the data center operator. But there is still a gap: end-to-end performance assurance.
Consider an enterprise using a hybrid cloud application built with local servers as well as data center resources, linked over a private connection. Let’s say there’s a performance issue with the applications using this setup. Is the problem in the connection service? At the data center? Maybe it’s in the user’s own network? Another possibility is that there’s a problem at one of the handoffs. In short, you need a way to measure the performance from end to end, and isolate any problems that are detected.
Assuring performance end to end will become more important as mission-critical services move to the cloud, especially when the access is best-effort internet. Another driver will be the move to support more remote users, whether they be working from home or on the road. In either case, they need the same assurance of application performance as their colleagues working in the office.
Connect Is Necessary, but It Isn’t Sufficient
To fully leverage the cloud, users need to identify and work with reputable telco and cloud operators. Ideally, they will use an end-to-end combined offering, with an overall service-level agreement. If not, they will have to apply the available tools to assure the performance themselves. Best effort is not good enough – we need assured performance, from end to end.