This is first article in a series on Cloud Computing for HPC. This guide describes the challenges that users face and the solutions available to make running cloud based HPC applications a reality.
IT organizations are facing increasing pressure to deliver critical services to their users while their budgets are either reduced or maintained at current levels. New technologies have the potential to deliver industry-changing information to users who need data in real time, but only if the IT infrastructure is designed and implemented to do so. While computing power continues to decline in cost, the management of large data centers, together with the associated costs of running these data centers, increases. The server administration over the life of the computer asset will consume about 75 percent of the total cost.
Scaling an existing data center in many cases is not possible as demands from users increase, and as the real estate or construction costs become too high or unavailable. Cloud computing can offer IT departments the ability to respond to critical needs while costs remain under control and predictable. With more compute power available to organizations than previously possible, business innovation on many fronts is just a click away when working with a cloud computing provider.
The term, “cloud computing” can mean a variety of things to different people. The National Institute of Science and Technology (NIST), which is part of the U.S. federal government, defined the term, which is used in U.S. federal government activities and broadly encompasses a number of important points.
“Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.”
The NIST also defines characteristics that are essential to a cloud computing environment. These include:
- On-demand self-service – The consumer can acquire computing and storage as needed without requiring human interaction at the service provider.
- Broad network access – The services provided can be accessed through standard mechanisms and include a range of client devices.
- Resource pooling – The provider’s computing or storage environment should allow for multiple uses (tenants) at the same time.
- Rapid elasticity – The capabilities can be expanded or contracted depending on the demand of the user.
- Measured service – The billing reflects the resources used in some abstracted fashion and may include compute, storage and network use. These resources can be monitored as used.
Over the next few weeks we will dive in to the following topics on Cloud Computing for HPC:
- Cloud Computing Models
- Economic Saving for Cloud Computing for HPC Applications
- Why HPC in a Cloud Environment
- Critical Cloud Requirements for HPC
- Uses of Cloud Computing for HPC
If you prefer you can download the complete insideHPC Guide to Cloud Computing for HPC courtesy of Penguin Computing