SBIR-STTR Award

Analysis, Prediction and Optimization of Quality of Service in Computer Networks
Award last edited on: 4/4/02

Sponsored Program
SBIR
Awarding Agency
DOE
Total Award Amount
$824,144
Award Phase
2
Solicitation Topic Code
-----

Principal Investigator
Bjorn Frogner

Company Information

NetPredict Inc

750 Menlo Avenue Suite 200
Menlo Park, CA 94025
   (650) 853-8301
   sales@netpredict.com
   www.netpredict.com
Location: Single
Congr. District: 18
County: San Mateo

Phase I

Contract Number: DE-FG03-98ER82626
Start Date: 00/00/00    Completed: 00/00/00
Phase I year
1998
Phase I Amount
$74,940
New tools are needed to proactively manage the so-called Quality of Service provided to the end-users of computer networks, such as DOE's researchers. In particular, it is important to determine the impact of network and load modifications on Quality of Service before new applications are deployed. This project will develop an integrated suite of analytical software tools for network Quality of Service management. These tools will provide network managers with the capability to analyze, predict and help optimize the Quality of Service of an integrated system of applications, hosts, clients and networks, by combining the use of measured data and network device models. In Phase I, macro-level network models and trending techniques will be developed and network device performance models will be developed and validated. A framework and algorithms for network simulation and optimization will be developed. Finally, the architecture and user interface requirements for the associated software will be defined.

Commercial Applications and Other Benefits as described by the awardee:
By advancing the state-of-the-art of network management into true predictive and decision-support capabilities, large government and commercial markets are believed to exist for the resulting software tools.

Phase II

Contract Number: DE-FG03-98ER82626
Start Date: 00/00/00    Completed: 00/00/00
Phase II year
1999
Phase II Amount
$749,204
Due to the rapidly growing use of networks (LANs, WANs, Internet, etc.) in all segments of commercial, private and government activities, performance management has become an acute problem. In response to this need, a Quality of Service Analyzer that can determine where delays come from, identify ways of reducing these delays, and predict the effects of additional load, is being developed. The QoS Analyzer is founded on a model-based approach that involves automatic calibration to measured data. This approach, which originated in the process control and aerospace industries, has now been adapted to the network management problem. It provides high accuracy for monitoring, as well as predictive calculations. In Phase I, a theoretical framework was developed and a prototype implemented to solve customer problems and demonstrate the benefits of the technology. In Phase II, the underlying technical foundation will be extended. The focus will be on expanding the validity of the developed framework and searching for solutions that can provide specified quality of service. Commercial Applications and other Benefits as described by he awardee: By analyzing and identifying delays in computer networks, this product should improve productivity in both government and commercial markets.