 QUASAR >> Data Center performance tool
Stanford engineers Christina Delimitrou (a doctoral student
in electrical engineering) and Christos
Kozyrakis (associate professor of electrical engineering and of computer
science) have created a software tool that can triple the efficiency of
computer server clusters in data centers.
As cloud computing grows, so will the cost of keeping big
data centers load capacity. Two Stanford engineers created this cluster
management tool which will triple server
efficiency while delivering reliable service at all times. This way, data
center operators will be able to serve more customers in a more efficient way.
The cluster management system called Quasar, is based on an
algoritm Just like Netflix or Amazon. Netflix and Amazon use an algorithm to recommend movies to each of its watchers. Same way Quasar, will offer advice to
thousands of server-farm computers on how to efficiently share the workload.
Today data centers are managed by a reservation system.
Application developers estimate what resources they will need, and they reserve
that server capacity. But the developers always overestimate, hence this
results in a lot of excess capacity.
Christos Kozyrakis thought that they should move away from the
reservation system. Instead of asking developers to estimate how much capacity
they are likely to need, the Stanford system would start by asking what sort of
performance their applications require. For instance, if an application
involves queries from users, how quickly must the application respond and to
how many users? So Quasar will get parameters from the developers for their applications to evaluate. Under this approach the cluster manager would have to make
sure there was enough server capacity in the data center to meet all these
requirements.
Christos Kozyrakis said their aim is to switch from a reservation-based cluster
management to a performance-based allocation of data center resources. As a result, he said entire data centers could be made to run at 70 to 75 percent of capacity. If so, there could be significant reductions in the amount of power that big computing systems now consume.
Quasar has been in development for two years and has been looked at by several large technology companies. Mr. Kozyrakis said his team was talking with several companies, which he declined to name, about using the software. He may also form a company to develop it, he said, or turn it over to an open source foundation.
|