High Performance Computing: Considerations When Deciding To Rent Or Buy
نویسندگان
چکیده
The commercial cloud computing (CCC) industry has reached a level of maturity to make it a truly viable alternative to the traditional, in-house data center. Although there are many notable examples of CCC platforms and technologies being piloted for high performance computing (HPC) tasks, it has yet to enter the mainstream. A variety of obstacles exist which have slowed or hindered adoption of CCC platforms, including implementation complexity, cost confusion, and security concerns. This paper describes the author’s experiences in using CCC for various HPC tasks and compares the results to the same tasks being executed on in-house computing resources. The breakpoint at which CCC becomes more costly than in-house equipment is identified. Lastly, a series of “lessons learned” are presented to assist researchers in effectively interacting with CCC vendors and platforms.
منابع مشابه
Learning to Make Rent-to-Buy Decisions with Systems Applications
In the single rent-to-buy decision problem, without a priori knowledge of the amount of time a resource will be used we need to decide when to buy the resource, given that we can rent the resource for $1 per unit time or buy it once and for all for $c. In this paper we study algorithms that make a sequence of single rent-to-buy decisions, using the assumption that the resource use times are ind...
متن کاملCompetitive Analysis
The ski rental problem arises from the following scenario. You want to start skiing, but you don’t own any skis. On each day you go skiing, you can either choose to rent skis for cost 1, or you can buy skis for cost s. Since you don’t know when you might get sick of skiing, (or when you might break your leg and be unable to ski again), the number of days d that you will end up skiing is unknown...
متن کاملA Constant-Factor Approximation Algorithm for the Multicommodity Rent-or-Buy Problem
We present the first constant-factor approximation algorithm for network design with multiple commodities andeconomies of scale. We consider the rent-or-buy problem, a type of multicommodity buy-at-bulk network design in which there are two ways to install capacity on any givenedge. Capacity can be rented, with cost incurred on a per-unit of capacity basis, or bought, which allows u...
متن کاملStrict Cost Sharing Schemes for Steiner Forest
Gupta et al. [J. ACM, 54 (2007), article 11] and Gupta, Kumar, and Roughgarden [in Proceedings of the ACM Symposium on Theory of Computing, ACM, New York, 2003, pp. 365– 372] recently developed an elegant framework for the development of randomized approximation algorithms for rent-or-buy network design problems. The essential building block of this framework is an approximation algorithm for t...
متن کاملOptimizing Cloud Use Under Interval Uncertainty
One of the main advantages of cloud computing is that it helps the users to save money: instead of buying a lot of computers to cover all their computations, the user can rent the computation time on the cloud to cover the rare peak spikes of computer need. From this viewpoint, it is important to find the optimal division between in-house and in-the-cloud computations. In this paper, we solve t...
متن کامل