WCCS
Microsoft Windows Compute Cluster Server 2003 (WCCS) brings the computational power you expect from an HPC cluster, combined with the familiar Windows environment for building HPC applications.

• Simplified WCCS Deployment: Building a new WCCS cluster is a lot like installing Windows Server 2003. The WCCS setup tools automate the installation. You are asked a few questions along the way but are not required to perform a multitude of configuration tasks. The installation program will automatically install and configure the components required to run the cluster (i.e. cluster networking, node management, cluster security, etc.)
• Powerful Management Tools: WCCS provides you with familiar tools for managing an HPC cluster. Since it is based on Windows you can utilize well-established Windows server tools like Active Directory for authentication and security management, Microsoft Operations Manager for system and job management, Remote Installation Server for installations to remote network nodes, and Systems Management Server for automating updates to cluster nodes.
• Tight Integration with the Microsoft Technology Stack: As you might expect from a Microsoft product, WCCS is tightly integrated with the Microsoft technology stack (see the previous point as an example). This is a solid strategy on Microsoft's part as it allows you to work with tools you already know. The learning curve remains low as do the timelines for implementation.
  
  Something that might be missed here is that Microsoft's developer tools are integrated with the WCCS feature set. Developers can use their .NET coding skills to develop applications that target a WCCS cluster. In addition, Visual Studio provides a parallel debugger that you can use to debug applications that span multiple nodes.


• A "Plays Well with Others" Strategy: No doubt Microsoft would like to lead in the HPC market, but they realize that businesses utilize a mixture of technology platforms. Given this fact, Microsoft built WCCS to work well in mixed environments that include other platforms (e.g. Linux). WCCS can significantly enhance non-Microsoft-based clusters by providing better management tools. For example, Linux clusters typically require developers to download an assortment of open-source tools to complete the same tasks as tools provided out-of-the box with WCCS. In this scenario, deploying WCCS will allow you to manage all cluster nodes using Microsoft tools.

How Your ISV Can Benefit from WCCS
If you are interested in incorporating the centralized computing power provided by WCCS, I recommend that you visit the WCCS web site. In addition, I have provided short checklist of questions to get you started when considering if WCCS is a potential fit.

1. Are you looking to scale your application over several server nodes?
2. Does your application perform a high number of computations and/or transactions?
3. Does your application require large amount of disk space (i.e. greater than a terabyte)?
4. Does your application stream media?

With WCCS, Microsoft is following their tried-and-true strategy of simplifying complex technologies. I anticipate this strategy will help propel HPC into mainstream usage, as it brings the power and familiarity of Windows to the HPC market. If you are looking to take advantage of HPC in your products and applications, I recommend starting with Microsoft Windows Compute Cluster 2003.

More Resources

• Article: HPC, the Message Passing Interface, and Microsoft: What You Need to Know to Control, Manage, and Communicate Over a High-Performance Computing Cluster with the Microsoft Platform



• Microsoft Windows HPC Community



• Web Seminar Series: Compute Cluster Server 2003: Technical for ISV
  This 8-part web seminar series will give ISVs the knowledge necessary to work with and begin to program for Microsoft Compute Cluster Server 2003. CCS provides a high performance computing (HPC) architecture where applications can be executed across multiple servers linked together in a cluster.
• Introduction to Compute Cluster Server 2003
• Install and Configure the Head Node
• Manually Deploying Compute Nodes
• The Job Scheduler
• Administrator Operations
• Automated Installations Using RIS
• 64-bit Software Development
• MPI and Parallel Implementations


• ISV Web Seminar Series: Longhorn Server—High Performance Computing with Windows Compute Cluster Server

* This article was commissioned by and prepared for Microsoft Corporation. This document is for informational purposes only. MICROSOFT MAKES NO WARRANTIES, EXPRESS OR IMPLIED, IN THIS SUMMARY.

Ty Anderson is a partner at Cogent Company, a Microsoft-based consultancy in Dallas that helps organizations use Microsoft technologies to grow their business. Ty is a regular contributor to Devx.com, wrote Office Programming 2003: Real World Applications (Apress), and loves to wear his yellow A's hat.