AWS provides a broad range of compute nodes and clusters, database engines and web development tools. U-M has an enterprise agreement with AWS providing certain benefits to U-M users, including waiving of data retrieval or “egress” fees.

AWS provides many levels of virtual compute server instances (EC2) running various flavors of Windows Server and Linux Server OS’s (Windows OS costs 50-70% more than Linux).  It also offer various relational database and web application development/deployment, among other services.

General purpose (T2 & M3) EC2 instances have low CPU performance with modest memory. Compute intensive tasks should utilize C# & R# instances, depending on how much memory per core is required for your applications.

In all of these cases, the number of vCPUs reported are twice the real number of compute cores. For example, 4 vCPU equals a 2 core node with 2x Hyper Threading enabled. For general purpose computing this is fine; four light level tasks will all run nicely in this environment. However, intensive tasks will negatively impact the runtime. If your task needs 4 cores, choose an instance offering 8 vCPU.

On-demand resource costs are higher than reserved resources.  Flux offers monthly commitments while AWS offers 1- or 3-year commitments. AWS also offers an auction market for what they call “Spot” instances (non-used reserved) which can be as low as ⅕ on-demand prices.

AWS also offers an extensive infrastructure for web and mobile application development, deployment and hosting.

ARC-TS can help you design your solution and determine which Cloud provider best suits your needs. We can also fully support your environment once deployed. Contact us at hpc-support@umich.edu

Order Service

To take advantage of U-M’s enterprise agreement with AWS, which provides additional features including the waiving of data retrieval or “egress” fees, see these directions.

For more information:

Email hpc-support@umich.edu

Related Event

December 9 @ 3:00 pm - 4:00 pm

MICDE Seminar: Ann Almgren, Lawrence Berkeley National Lab

Bio: Ann S. Almgren is an applied mathematician who works as a staff scientist and acting group leader of the Center for Computational Sciences and Engineering at the Lawrence Berkeley National…

December 12 @ 1:00 pm - 5:00 pm

HPC User Meetup

Users of high performance computing resources are invited to meet ARC-TS HPC operators and support staff in person. There is not a set agenda; come at anytime and stay as long as…