Promising New Research that Addresses Issues in Cloud Computing
Cloud computing is a relatively recent innovation in computer technology. Given the fact that current semiconductor manufacturing techniques have reached an upper limit in switching speed, the industry has adapted by producing machines with multiple processor cores. This allows software instructions to be run concurrently on a single machine. Concurrent processing allows for virtualized systems to be installed on a machine, which means that one piece of hardware can run many Operating Systems at once. This virtualization is commonly known as infrastructure as a service, and it is one of the most basic forms of cloud computing.
The Advantages and Disadvantages
Infrastructure as a service provides many cost advantages for both clients and cloud computing firms. Virtualized Operating Systems can use as much or as little of the underlying hardware resources they currently demand. This allows for an efficient distribution of hardware resources, which effectively lowers costs and improves the quality of service to end users. However, it introduces problems that can cause an entire physical machine to crash. For example, if one virtual Operating System encounters a serious bug, it may begin to demand more physical resources than it actually requires. In the least severe case, this can lower the quality of service for all users of that physical system. In the most severe case, the physical machine may lock up and cause all virtual Operating Systems to crash.
Attempts to Resolve Issues
North Carolina State University has spent a considerable amount of time and money researching this problem in cloud computing. Their solution uses specialized diagnostic software that preserves the security and integrity of each virtual Operating System. Security is very important for infrastructure as a service clients. Since virtualized systems share hardware with other clients, it is feasible that a hole in a hypervisor application can expose sensitive data to unauthorized users. The software solution developed by North Carolina State University preserves the security of each virtual Operating System by treating them as “black boxes.” In other words, each virtual system is analyzed based on external data alone, which doesn’t require access to the sensitive user data internal to each system. This external data is regularly analyzed and compared to metrics that indicate nominal behavior. Abnormal behavior can be detected before serious performance issues manifest. Steps can then be taken to ensure that the system does not crash.
A Step in the Right Direction
This technology is promising for a number of reasons. First, it can mitigate quality of service issues with virtual Operating System faults. Second, it preserves the security of virtual machines. Security and quality of service are very important to infrastructure as a service clients. These issues must be minimized to ensure the success of cloud computing. Cloud computing is a promising direction for the computer industry, as it addresses the physical limitations of semiconductor switching speeds. By combining multiple systems on a single piece of hardware, many clients can be served at once with reduced energy demands, which will help lower the cost of cloud computing and help reduce global energy demand. In order for these goals to be realized, solutions to common problems in cloud computing must be addressed. The research conducted by North Carolina State University is a promising step in this direction.
This is a guest post by Steven Felsenfeld. Steven is a cloud software engineer who is fascinated by the work of 522 Productions. A production company specializing in high quality corporate promotional videos
