This article originally appeared in the March issue of INTERNET TELEPHONY magazine.
Last issue, I wrote about solutions that are available today to interconnect an off-premises cloud deployment with an on-premises data center. For the most part, every cloud provider offers some type of connectivity solution, either in the form of a long-term or permanent tunneled connection (possibly using one of the platform solutions I mentioned last time such as VMware’s vCloud Connector) or as temporary links to copy and/or upload data required for applications in the cloud – think of your most basic HTTP/FTP uploader here, and many cloud providers are still offering these basic solutions.
Cloud providers have to offer some way to accept customer data over an interconnect solution, but the level of sophistication can, and does, vary widely between providers. This month, I’d like to dig into how moving to more advanced cloud gateways can help bring out the true power of cloud computing, especially in a hybrid world.
Hybrid clouds are a combination of services spread between an on-premises cloud environment and an off-premises cloud. The services can be split by function – an isolated part of the application runs off premises but the rest of the app runs on premises – or by architecture – computation or analytics are done on the data in the cloud but the user interaction is managed on-premises – or even dynamically scaled to the off-premises environment as needed. Dynamic scale is usually the example we use when talking about hybrid cloud computing because it’s the most interesting to the business: using the cloud only when needed but maintaining control on premises is the best of both worlds.
A dynamically scalable hybrid cloud can be further broken down into two additional categories: cloud bursting, the really juicy cloud use case, and disaster avoidance, a combination of disaster recover, business continuity, and dynamic scale enabled by virtualization and cloud computing. Although they’re both hybrid models that involve dynamically redirecting users and services between on and off premises, cloud bursting and disaster avoidance are different in their design. Cloud bursting is about bringing up services in the cloud as needed, bringing them back down when not needed, and managing user access between the two environments. Disaster avoidance is typically a longer-term solution that involves keeping multiple cloud environments cloned and ready to go when one becomes unavailable. The distinction between these two types of hybrid cloud computing comes into play when you start looking at cloud gateway devices.
Historically, cloud gateways were known as cloud bridges because they bridged – in both general and IT-architectural terms – the on- and off-premises data centers. This term was a bit of a red herring, however; not every cloud bridge was an actual layer 2 network bridge. But some were, and those real bridges tended to be highly specialized and came with limitations, such as single vendor solutions and distance restrictions – a layer 2 bridge can only travel so far. The technical limitations eventually affected the terminology limitations and newer, more sophisticated cloud bridges rapidly became known as cloud gateways, a very apt term given all that can be done with a cloud gateway.
Cloud gateways go beyond simple network connectivity and can provide many network-related features that are missing from most off-premises cloud deployments. Cloud gateways can optimize WAN links, exchange information with the cloud provider such as resource metrics, and can even manage user security such as on-premises authentication and authorization. The two different hybrid computing models – cloud bursting and disaster avoidance – benefit from cloud gateways in different ways. In order for cloud bursting to be truly successful, the off-premises environment needs to be treated as a logical extension of the local data center.
Tools that manage cloud bursting events – such as application delivery controllers – need to see the cloud as an available resources pool for the application that’s bursting. A permanent cloud gateway that maintains a secure and optimized WAN connection is ideal for cloud bursting because it basically becomes a long-distance LAN, and applications in the cloud become true extensions of the applications in the local data center. Cloud gateways can also help create and manage cloud environments for disaster avoidance by providing optimized connectivity to the cloud provider for moving virtual machines, large data sets, and for keeping data synchronized between the two locations.
Down the road there are some really interesting things that we’ll be able to do with cloud gateways, such as dynamically creating WAN connections between providers, but for now we can be happy sticking with what cloud gateways can do for us today. These advanced connectivity and network management solutions aren’t available with traditional cloud bridges; you need a cloud gateway to truly extend the data center reach into the cloud.
Edited by Jennifer Russell