The rise of cloud computing is a perfect example that it is more economical to scale horizontally and let more machines work than vertically where we have one machine which becomes more and more powerful. But the question we have now is, how can we harness the power of horizontal scaling effectively? As we add more machines to do our bidding, we need a way to coordinate the activities among them and bring those discreet working entities to work concurrently, load balance the work among them, increase their throughput and make them achieve a common goal. We must also be able to take care of exceptional conditions when entities start going crazy or failing. Also important in Cloud’s context is tunable processing capacity. We must be able to scale up or scale down dynamically according to our needs.

This talk explores at a technique called Space based programming. We will discuss about how Spaces fits the domain of the Cloud to the challenges described. Also we will look at working code and live example to understand the method of using tuple spaces, their advantages and their constraints. We also would contrast tuple spaces with other methods like using Message Queues and illustrate how Spaces help us solve the problem in a simpler and easier way than the conventional techniques we choose at the Cloud.

Takeaways from the session
We will take a look at new and lesser known approaches/tools to achieving high levels of scalability on the cloud. The talk will consist of topics such as space based programming, tuple spaces and their relevance to cloud computing. The audience will also get to see real world examples as well as working code.

Speaker: Hari Krishnan & Venkatesh Sampath R
This session on “Scaling In The Cloud Using Spaces” will be presented at the 1st IndicThreads.com Conference On Cloud Computing to be held in Pune, India on 20,21 August 2010.Click here for a list of other Sessions @ The Conference

Comments are closed.

preload preload preload