Scientists and Cloud Computing - part 1
Scientists and Cloud Computing - part 2
The author ( Krishnan Subramanian) outlines two advantages of Cloud Computing for the scientific community:
Cost Savings: Unlike many enterprise level customers, the requirements of individual scientific groups, in most cases, comes in bursts. It doesn’t make sense to invest heavily in infrastructure. In fact, in almost all the cases, they cannot afford to invest in huge infrastructure. This is one of the reasons why high end scientific computing infrastructure are usually shared by several hundred research groups from all over the world. There is another factor in play too. Most of the scientific funding is done for short term. The money available in such short term funding is not enough to consider building high end infrastructure. This offers a great opportunity for Cloud Computing to jump in and fill the void. Scientists can just “rent” the infrastructure for a short period when they need to use the computing resources and pay just for the usage. This offers them tremendous amount of cost savings and, also, gives them an opportunity to conduct some experiments which they would have left out due to funding considerations.
Time Savings: One of the biggest problems facing the scientists needing high computing resources is the issue of time. Building an infrastructure for computing costs lots of money but, more importantly, it takes quite a lot of time too. Even if a scientific group is part of the hundreds of groups that share computing resources from places like CERN or other national laboratories, they will have to wait for their turn to come. [...] In this fast paced world where different scientific groups are vying for the same piece of the scientific puzzle, time is the most crucial factor. [...] This is where Cloud Computing will come handy for scientists. They don’t have to wait for anything. They can just launch the necessary computing instances on the cloud and start their computation. In fact, they can scale up without much or any extra investments (for eg. instead of running 1 instance for 100 days, if they can “parallelize” their program, they could run 100 instances for 1 day and get the results). Such scaling options gives them the necessary agility in the competition to churn out scientific results fast.
Kevin
ZeroTouch IT Ltd
