Top Five Perils Facing Cloud Computing
As cloud computing becomes more favorable among companies, they are forcing their applications out of the internal network into the cloud, causing them to be vulnerable to Web threats.
Cloud providers can guarantee that they can deliver a working environment not just for bulk data and B&R functions, but for enterprise applications and other higher order functions.
Part of the deal is that the provider promises to maintain an agreed-upon service level, providing a level of assurance that operations won't be disrupted to any greater or lesser extent than those housed on traditional owned-and-operated infrastructure. The question, though, isn't whether they can support your applications, but whether they can provide the same performance that you and your users have grown accustomed to.
This isn't a trivial matter, for providers and users alike, as the cloud will be forever relegated to second-class status unless it can provide functionality at or near today's data environments. And the simple fact is that the more distributed architectures become, the greater the challenge to ensure robust and responsive operations.
According to Quocirca, more than half of top enterprise officials name application performance management as the number one issue for the coming year. Indeed, if IT sheds more responsibility over the development and maintenance of actual infrastructure, measuring application performance is seen as the only way to ensure that IT budgets are going to good use. And with nearly three-quarters of IT executives expecting application demands to increase over the next few years, interest in shoring up both the application software layer and the network/user access environment is on the rise. Key requirements in all of this are the proper app performance metrics and the ability to map them to business goals, as well as more advanced, proactive monitoring capabilities.
The cloud, though, is nothing if not distributed. Which is why app performance will depend very much on your ability to track applications across increasingly diverse network infrastructure and to make adjustments to optimize the relationship with dynamically shifting resources. As The Aberdeen Group noted in a recent report, application infrastructure will not only have to deal with increased volumes and a wider variety of data types, but with multiple types of WAN links like traditional corporate Internet, broadband, wireless and MPLS. This will drive a need for greater application awareness and intelligent, adaptive networking.
The fundamental problem with app performance in the cloud is that it relies upon infrastructure that the enterprise does not control and cannot monitor as effectively as it should. As Visual Networks Systems' Belinda Yung-Rubke told Network Computing recently, a crucial feature of any cloud deployment is the separation of data in multitenant environments. This is the best way to ensure that the data coming from your monitoring system, like response times and traffic flows, is germane to your application environment. Ideally, both internal and external environments will be managed as a cohesive whole.
For the cloud to be truly effective, enterprises will need to seamlessly transition workloads, applications and resources across internal, external and hybrid infrastructure. But while past solutions focused on the infrastructure itself - how it is deployed, what components are over- or under-utilized, and where the trouble-spots are - the new world will depend largely on the applications themselves. As long as you maintain visibility into the application layer, you should have all the information you need to keep things running smoothly.