Today’s OC-ALC/IT infrastructure is closely linked if not very dependent upon a single product line by Microsoft Corporation from server/desktop operating systems to almost all applications. A single product line has become the defacto standard for the IT environment.

2.0 Problem Statement

Because the AF and OC-ALC are not using open standards for server and desktop software infrastructure the command has become totally dependent upon a single product line by one company. This has lead to spiraling costs through license upgrades when the product line changes with the operating system every 18 months to two years. Such a dependency has an inherit weakness because the health and security of the IT infrastructure is very dependent upon the health of the product line owner.

3.0 Anticipated Tasks/Methodology

Description

Objective of the CASI research project is to measure performance, security, dependability of the Red Hat Linux Advanced Server operating system against the Windows 2000 operating system. This objective thus requires an infrastructure, which requires three to four servers with Intel 64 bit CPU technology, and commercial off the shelf load enterprise testing software.

In addition, to provide direction to the RFP and subsequently to the proposal developer (the professor and/or students), please define the requirement (if any) for the use of the particular task:

Windows Advanced Server 2000 and Linux Red Hat 8.0 will be co-located on each server.
BEA J-Rocket will be the Java Virtual Machine (JVM) used on all BEA installations along with the usual BEA JVM so a comparison can be made.
Mercury Interactive (www.mercuryinteractive.com) software will be used enterprise testing by simulating large client loads. A separate proof of concept may be required by the AF and Mercury Interactive and can be arranged only by the AF and not the professor and/or students.
The application software will be BEA Platform 7.0 service pack 2 or BEA Platform 8.1 (Beta). BEA Platform 8.1 will be used if the product is near general release. This will be determined by the AF program manager.
No programming code is required, anticipated or permitted for this project.
Overall, results of the project may be published, but the various software suites other than Linux may not be identified without the expressed written consent of the AF or the participating vendors.

Anticipated Major Tasks

Below is a list of example activities included for your use (but are not limited to):

Kick-off and other Meetings/Interactions with OC-ALC program manager
Development and Review of Methodology and Data Requirements
Development of Program Plan
Generation of Solution
Development and Submission of Final Report and Briefing

4.0 Anticipated Schedule

Please provide the anticipated Project Period of Performance:

[x] Execution and Completion of Project between June 1 and August 31, 2003.

5.0 Anticipated Deliverables

The required deliverable is a report showing a case by case comparison of Windows 2000 versus Linux Redhat 8.0. The report will include but not be limited to scaled performance metrics showing CPU and RAM load, NIC load, and Garbage Collection incidents per operating system on standalone versus clustered servers. Other measurements will be determined during meetings with program manager.

6.0 Anticipated Benefits

Baseline measurement of performance of mid-tier application on Intel 64 bit platform
Baseline measurement of performance of mid-tier application with comparison between Windows Advance Server 2000 and Linux Advanced Server.
Baseline cost measurements of Linux and Windows
Baseline measurement of network, CPU and memory loads
Baseline measurement of clustered network, CPU and memory loads.
Indicators for reducing IT infrastructure costs.
Benefit analysis of Linux versus Windows.
Open standards versus product line benefits.
Other benefits to be determined