CLOUD

With the advent of cloud supercomputing, users worldwide are provided huge data storage that can be accessed and shared from anywhere using laptops, smart phones, tablets, and other devices. Such global data access and sharing allows greatly-accelerated progress in scientific and technological achievements, not to mention innovative social media uses.

However, with the data available worldwide, challenges remain in the ability to handle the volume of users attempting to access information simultaneously, as well as in the capability of a data center to respond to the varying request times that might limit performance.

To help manage this storage of data in an efficient manner, Ali R. Butt (http://people.cs.vt.edu/butta), associate professor of computer science (http://www.cs.vt.edu/) at Virginia Tech, has received a NetApp Faculty Fellowship.

The storage of shared data via cloud computing is relatively new, although the original concept dates back to the 1950s when large-scale mainframe computers became more available. 

However, distortions may still affect the retrieval of information. Butt's plan to address this difficulty that often comes with a high volume of requests as well as performance problems from varying request patterns is to further develop Mizan, a high-performance and load balanced in-memory object caching storage system. In computing, a cache is a component that stores data in memory so that future requests for that data can be served faster from the memory instead of retrieving it from a slower device such as a disk.

"The goal is to make the load-balancer track hotspots and apply appropriate strategies," Butt explained. This tracking might reveal key repetitious inquiries or provide a better idea of where data is migrating. With this information, "we may be able to dynamically mitigate imbalances, while supporting high-performance and efficient use of resources." 

Butt already knows that Mizan can quickly detect hotspots in workloads, and is able to lessen any load imbalance that is critical in cloud computing.

"The choice of this design will allow Mizan to scale-up to a large number of cache nodes and scale up its performance by fully exploiting the parallelism offered by emerging massively multi-core computer architectures," Butt said.

Butt is a previous recipient of a National Science Foundation CAREER award for his research in high performance supercomputing power.