emergency-if — archive

[David Ellis]

Don and others; Thank you for the detailed e-mail. We need to put your use case into EA. As I read this however, there seems to be a specific application and technology "structured Restful or other URI" technology choice made. We should evaluate these issues and determine if they are real requirements for the EDXL-DE routing standard and if so work to provide a means to solve them. The requirements you outline still look like metadata describing content (Content object) and ways to "pull/poll" for data at rest. You are correct in your assessment of UCORE and UICDS in my opinion. However; the use of URL and "string" in EDXL-DE 1.0 example(s) and the other artifacts were due to the rapid nature of the EDXL-DE 1.0 standard. Also, Many of the requirements you specify are really IPAWS like and this was the primary reason we pushed the standard out so fast. The theory of the URN was not an accident and it had significant subcommittee discussion. Unfortunately we did not have a "Jeff" documenting subcommittee meetings or audio recordings. I think we need to again have detailed discussions about requirements and not rush to put out a OASIS EDXL-DE 2.0. The real problem is much of the technical talent which developed the theory of the EDXL-DE 1.0 have left the TC. In fact, there has been several attempts by these standard developers and other technical/consumer groups to create a different technical committee to move DE work or create a more universal "interoperability" EDXL-DE like standard. A truly secure policy based information exchange capability is the "Holy Grail" of solving numerous interoperability needs. The need to dynamically determine situational (incident centric in OASIS EM requirement space) participant roles and enforce information exchange policies between them in time evolving scenario is critical to any small/large information exchange. Unfortunately, the determination of when a small incident will become a larger and potentially more information restrictive (terrorist attack) cannot be determined in advance and there is a need to have a seamless capability to scale information exchanges. One of the reasons the DoD GIG is having so many problems with exchanging information is because it does not use the TCP/IP packet and metadata stack pattern/concept for exchanging data. They keep implementing "Service" which must be dynamically queried "pull/poll" for evaluating policy and routing metadata evaluations (SAML/XACML). These type of system inevitable fail when they are scaled. Even the internet DNS system which is used to create routing metadata (sender creation) is out of band, highly decentralized in authoritative updates and not used by intermediate security and routing infrastructure. The IP packet uses IP addressing and ports to implement routing and security decisions when in transit. For those of us who were a part of the Arpanet (AFWL) and DECNET development, these lessons learned should not be relearned in our evolving application metadata routing implementations. This is an important national/international need and OASIS EDXL-DE 1.0 was a good first effort in meeting the need. Dave