• Objective
• Status
• Results
• Patents pending
• Team members
• Sponsor

Objective

Current communication mechanisms such as protocol suites, client-server and Remote Procedure Calls, suffer from too much internal computer communications overhead to provide the communications performance required for future distributed systems applications (e.g., Network of Workstations - NOW) and to take advantage of the new, fast network technologies such as ATM and fiber optic based gigabit networks.

• Speed up the computer internal communications mechanisms for network communications
• Maintain 100% interoperability with all existing applications, operating systems, communication protocols, network interface cards and networks.

Improving communications throughput by a factor of 2.5 to 10 (an order of magnitude) without requiring changes to existing programs, hardware and networks is what makes the INCA project of interest.

Visit the INCA project web site.

Status

We have implemented INCA part 1 on two UNIX operating systems, Solaris 2.5.1 and LINUX 2.0 and have simulated INCA part 2.  INCA part 1 consists of software only which can be used by all existing applications and computer systems to improve their network data throughput by nearly an order of magnitude.  INCA part 2 involves cache memory operation modifications and has been simulated yielding an additional 13.5% to 46.6% network data throughput improvement.   Work continues to further improve performance while maintaining interoperability.

Results

To date, INCA has demonstrated a minimum of 260% and a maximum of an order of magnitude (1000%) improvement in network data throughput while maintaining 100% interoperability - requiring no changes to existing application programs, computer architectures, operating systems, protocols, network interface cards and networks.

INCA is one of the few methods that speeds up small message throughput on the same order as large message throughput.  INCA is the only known mechanism that speeds up small and large message throughput by up to an order of magnitude while requiring virtually no changes to existing applications, systems and networks.

Patents pending

Due to the potential commercial application of INCA, our sponsor has applied for two patents, one for INCA part 1 and one for INCA part 2.  As a result, implementation specifics are not for the public domain.  Our sponsor can be contacted for INCA product information.

This project is complete 

Team members

• Prasanth Gopalakrishnan
• K. Schug  (MiTech Inc.)
• Srivani Jade (CacheFlow)
• Suneetha Gullapalli (Lucent Technologies)
• Prof. Anura P. Jayasumana
• Dr. Klaus Schug

Sponsor

Phase 1 was supported in part by MiTech Incorporated, 9430 Key West Avenue, Rockville, Maryland 20850