At Ken’s suggestion, the TC decided to create an Actions and Open Issues “Parking Lot” for stalled or inactive items. Items moved to the Parking Lot at this meeting were:
Peter: Complete a zip file for SOA-RAF UML Models and load to KAVI
Ken: Peter Brown/Ken finalize adjudication of comments received as a result of review of OASIS SOA-RM by IEEE.
All: Gather Statements of Use
Ken: Will coordinate with William as to when Ken can put discussion of his “patterns” write-up on the TC meeting agenda..
Status of other Actions and Open Issues:
Martin: Update initial draft SOA-MSA side-by-side.—Martin posted a message to the TC list proposing a revised approach. Follow-up carried forward.
All: Review SOA-RM to identify potential areas for update.—Ken has started and will distribute initial review for discussion at next TC meeting.
B – Taskus Genuineus
Chair remarks—Ken recommended putting the patterns/ontology issue in the Parking Lot since he’s been unsuccessful getting status from William. (Rex said he hopes the action can be revived at some point as it might be useful.) Ken also recommended putting the IEEE item in the Parking Lot since he’s not heard anything recently.
Ken also mentioned an e-mail Bill Lewis (USA Civ) asking "what tools were we [SOA TC] thinking of when we defined the Description concept in SOA." Ken replied that no specific tools had inspired the TC’s thinking at the time (circa 2011) and observed that his experience is that very generic tools like links from a service catalog to Description text maintained in a Wiki are adequate.
Continuing discussion of various areas of contrast between SOA and MSA.
Rex said one of his takeaways from the recent NGINX Webinar was that there is a rudimentary but rapidly developing marketplace for APIs or microservices. Martin agreed that although the original and primary goal of MSA was support for rapid development of applications by independent teams, it does appear that a secondary effort is being made to gain the economic benefits of re-use via markets for microservices. Rex suggested the TC might perform a useful service by surveying the types of API or microservices repositories or registries that have appeared. He said a starting place might be the links to some of these repositories on Chris Richardson’s Web site.
Ken referred to a recent interview with a Google executive who said that Google has a massive repository of code available to all developers, and that this is effective since almost every conceivable function has already been coded by someone. Martin said he suspects that this approach only works if there are some disciplines added to how the repositories are managed: requirements for documentation of the function or microsevice code in the repository for example. . oogle. -- martin: doesn't work w/o added discipline. Rex said Google and AWS both have massive libraries of capabilities but manage them differently.
The TC discussed the fact that the context of services available on the Internet was very different now than when the first OASIS SOA work was done, and that some of the assumptions made back then about what would be economical and efficient as the distributed computing environment matured may have been wrong. For example, Rex suggested that SOA represented a “share everything” approach, whereas MSA seems to seek to “share nothing.” Martin suggested that actual experience implementing distributed systems seems to imply that "sharing takes too long" and speed is more important than re-use.
Review Committee Note procedures and templates
Rex suggested that the TC move to start trying to fill out the OASIS template for initiating a Committee Note on SOA-MSA comparison. He felt we have developed enough material to move to start documenting it. Ken agreed to find and review the OASIS template and instructions.
The meeting was adjourned at 12:30 PM ET. Next meeting in 2 weeks—May 24, 2017.
Actions carried forward –
Martin: Update initial draft template for side-by-side comparison of SOA vs. Microservices Architecture