SITIS Archives - Topic Details
Program:  SBIR
Topic Num:  N07-064 (Navy)
Title:  The use of Self-Managing Software Agents to Improve Application and Network Reliability and Data Accessibility
Research & Technical Areas:  Information Systems

Acquisition Program:  SEA073R
  Objective:  Develop and demonstrate the application of certifiably accurate, self-managing software agents to improve continuity of operations of the shipboard application environment. In addition, demonstrate how Semantic Web software agents can be employed to improve cross-venue data sharing capability for application to improve support of the operation, material maintenance and tactical war fighting missions of Naval submarines.
  Description:  Modern Submarines operate with the support of many IT applications running on a common shipboard tactical and logistics support network. The non-tactical shipboard network software environment has become increasingly complex to maintain with the delivery of the Non-Tactical Data Processing System (NTDPS), the Distance Support software suite, the Submarine Non-tactical Delivery Interface System (SNADIS), and the Navy Tactical Command Support System (NTCSS). This is furthered complicated by Fleet management desire to have near-real time readiness reporting and assessments sent off-hull on a regular basis to support optimize Fleet mission deployment. The level of expertise required to maintain these capabilities can easily exceed the skill set of shipboard network administrators. While considerable effort has been made by software developers to simplify the administration of individual applications, the aggregation and level of integration of these applications taken as whole is a very challenging environment to maintain a high Quality of Service (QoS). When problems occur, they can go undetected and/or unresolved, because there is nobody on board who is sufficiently proficient to understand and restore the lost capability. The potential for this condition is ever increasing as more and more network applications with cross-application dependencies get added to this environment every year. The use of self-managing/self-healing software agents/components could offset this lack of expertise and be used to correct and restore the non-tactical software environment to the desired optimum conditions. In addition, there is a Fleet need to evolve these non-tactical shipboard information systems to work more effectively and synergistically to support shipboard tasks. Vast amounts of data are provided to and generated by ships for use under the guise of shipboard maintenance, operations and training. The nature of shipboard operations are such that this data can often be consumed and acted upon by many activities; however, the data is seldom utilized for purposes other than those for which it was originally developed. This condition exists because there is no overarching search/discovery/association capability sitting in the shipboard information environment. As consequence of this inefficiency, data is often recreated as if new vice being repurposed and associated from an existing ship data store or application. Many man-hours and dollars are being expended and wasted by the Fleet as result of this condition. This SBIR would develop and demonstrate the use of self-managing software agents to address ongoing maintenance and support requirements of the evolving sophisticated shipboard network environment. In addition, this SBIR would demonstrate the use and viability of Semantic Web software technology/tools as a solution service/application to improve data access and interoperability. These self-managing software agents need to be able to adapt to handle changing user needs, system faults, changes in the operational environment, and resource variability. Such a software system must configure and reconfigure itself, augment its functionality, continually optimize itself, protect and recover itself, all the while keeping the complexity of the system hidden from the user. This technology is particularly important where the complexity of software continually challenges the organic skills of shipboard network administrator. The requirement to address the enclave environment of the shipboard networks is needed because current industry research and practice concerning self-adaptive/self-managing systems tends to focus on the big solutions providers working in a large fully connected enterprise environment. The support of all IT applications running on a common shipboard tactical and logistics support network is the responsibility of a small group of individuals that have this assignment either as collateral duty on older submarines or as dedicated function on newer submarines. These individuals are required to be capable of operating and maintaining all the hardware and software resources of the shipboard network. While formal training is provided by the Navy to help these individuals to become proficient and qualified to perform these duties, the enormous complexity of the evolving heterogeneous software environment on ships makes this job extremely challenging and creates opportunities for system failures. The operational nature of submarines does not always permit the use of shore-based resources to augment and assist the shipboard network administrator; consequently, there is a strong need to offset this condition with software agents that can detect, analyze, and initiate the use self-healing, self-optimizing, and self-configuring software. In addition, recent developments in the area of Semantic Web technology show substantial promise for augmenting the existing data sources with semantic markup so that the existing database can be processed as a true machine searchable knowledge base. Knowledge-base data mark-up technology is now emerging for production use after a substantial incubation period with significant DOD (DARPA) and University support. The World Wide Web Consortium (W3C) formally published the core specifications in 2004 permitting the development of initial sets of Semantic Web tools. However, this is a technology in which knowledge assertions are developed through a computational process. In doing so, there is, by design, little human involvement, and it is critical that the software agents be trusted both as to their accuracy to implement their design function and as to their reliability in carrying out their designated functions. This effort has four specific goals: 1.) To evolve an open software architecture within the existing operational environment that includes the methods, algorithms, techniques, and tools that can be used to support dynamic adaptive behavior in software; 2.) To demonstrate the ability to create useful scenarios in which data developed for one mission area can usefully be applied to another mission area with emphasis on bridging the logistics-support tactical-mission gap; 3.) To show that the software agents employed can be formally certified to accurately perform their design process; 4.) To demonstrate that the agent management capability employed is self-healing in its execution with emphasis on detecting and correcting abnormal system and software states.

  PHASE I: Perform general high-level assessment of the types of faults causing problems in current submarine operations based on interviews with system personnel (but not invasive examination of the actual operational system). Analyze the potential data sharing scenarios on an operational submarine and develop an operational concept description, select technology and design a demonstration platform. Specify and evaluate the needed software engineering tools to certify the software agents and self-healing capability that monitors their activity in execution so as to detect and correct session failures, and prepare an implementation plan for Phase II. Develop a candidate-shopping list of the kinds of tool characteristics that might be applicable. The candidate tools will include measurement and assessment tools for measuring a real system to facilitate the adjustment of operational agents for long-term operation. Working with Government representatives, select a comprehensive list of system monitoring and self-healing tools for a solution set targeted at the Submarine environment. Develop an acquisition/development plan to assemble a solution set composed of these identified tools.
  PHASE II: Execute the acquisition/development plan developed in Phase I and assemble a candidate tool set. Adapt specific measurement and monitoring tools for subsequent installation on a Government system. Demonstrate performance of this tool set on a contractor-created test system. Working with Government System personnel (who may be required to conduct actual measurements), conduct an assessment of system operations on an actual submarine platform employing the measurement tools developed earlier in Phase II. Develop a list of requirements for Phase III pilot capability. Update software and other tool assessments as needed to assure optimal software selection for Phase II demonstration. Acquire and/or develop agent software modules needed to execute scenario. Acquire and/or develop certification tools for agents. Acquire and/or develop self-healing, monitoring and restoration tools for activity monitoring capability. Load test system with representative data as mutually agreed with the Government. Rigorously test both the self-healing and certified data integrity using the tools in accordance with a multi-dimensional testing plan. Based on assessment, design set of run-time tools and agents for Phase III pilot system. Prepare implementation plan for Phase III pilot system test and demonstration.

  PHASE III: Conduct additional operational assessment and perform detailed requirements analysis for selected shipboard Pilot system and refine system design. Working with Government representatives select specific requirements to be satisfied by the Pilot capability. Design, acquire, implement, field, and support operation of a pilot production-level capability on selected submarine platform to be specified by the Government. Develop and/or acquire additional software agents as needed to implement the user requirements that are identified in the pilot testing. Conduct system test and reassess system performance after pilot system operational. Document measured benefits of installed system performance tools. Prepare full system description and other documentation as needed to replicate the operation on other Naval platforms. Update technology recommendations as needed to assure that recommended implementation is employing the current technology. PRIVATE SECTOR COMMERCIAL POTENTIAL/

  DUAL-USE APPLICATIONS: While the technology proposed for this solution set targeted at enclave intranets, there are many such legacy intranet systems in the government and the private sector and there will be for the foreseeable future. If the technology is demonstrated and proven in this effort for these legacy enclave environments, the commercial potential is virtually certain and will be in great demand. The approach is directly applicable to many commercial data environments and not unique to military environments.

  References:   1. “Toward Architecture-based Self-healing Systems”, Dashofy, E.M.; van der Hoek, A.; Taylor, R.N.; The ACM Digital Library, Proceedings of the first workshop on Self-healing systems, 2002 (There is a substantial amount of material available on the subject of self-healing software online in the ACM Digital Library) 2. “Self-Managing Software”, Hinchey, M.G.; Sterritt, R.; Computer (An IEEE Journal) Volume 39, Issue 2, Feb. 2006 Page(s): 107 – 109 3. “Services and autonomic computing: a practical approach for designing manageability” Kapoor, V.; Services Computing, 2005 IEEE International Conference on; Volume 2, 11-15 July 2005 Page(s): 41 - 48 vol.2 4. Resource Description Framework (RDF): Concepts and Abstract Syntax, W3C Recommendation, February 10, 2004, Graham Klyne, Jeremy Carroll, eds. 5. RDF/XML Syntax Specification (Revised), W3C Recommendation, February 10, 2004, Dave Beckett, ed. http://www.daml.org/ The DARPA Agent Markup Language (DAML) Program

Keywords:  Semantic Web; autonomics; self-healing software; RDF; XML; Knowledge base

Questions and Answers:

No questions posed on this topic at this time

Record: of