Panelists: Shivani Sud (Intel), Kumud Srinivasan (Intel), Raejeanne Skillern (Intel), Anugeetha Kunjithapatham (Samsung).
The session started with an introduction to Cloud Computing. Cloud Computing is a computation paradigm that makes computation resources available as services. This kind of paradigm is dynamic and virtualized with centralized compute and storage facilities. The model is on-demand self-service one. The storage and computing resources are elastic and scalable. The pricing is consumption-based (pay-per-service). The computing power is available anytime, anywhere, internet is required to connect though.
There are two types of clouds: public cloud and private cloud. One example of a public cloud is Amazon EC2. Public clouds are provided to the end-users over the internet, and using them one can grow as much as one wants. On the other hand the private clouds are offered to a restricted group of users. There is more isolation (control) and are costlier than the public clouds.
What is required to enable clouds?
- Ubiquitous connectivity everywhere
- Reliability of these services is going to be important
- Interoperability between service providers is a key to this kind of paradigm
- Security and privacy is not of much concern here
- Economic value are the key driving factors
Why and When not to use cloud?
- Not a blanket approach
- Connectivity is still spotty
- Financial benefits are not yet been proven
- Security is one of the biggest concerns
- The lack of interoperability so that can make one switch from one vendor to another might deter a buyer
A case-study from Intel was presented. The benefits they felt by using cloud computing were:
- Agility to respond to the users has increased
- Ability to accelerate the projects, especially if the expertise is not available in house
- On-demand self-service approach goes well with the end user
However, the cost picture is not clear. If not managed correctly, it could lead to an increase in cost. Security is a big concern also.