[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: FW: [soa-rm] Help in Sending this Note: Feedback on SOA-RM Spec (09)
-----Original Message----- From: Jeffrey A Estefan [mailto:Jeffrey.A.Estefan@jpl.nasa.gov] Sent: Monday, October 10, 2005 4:14 PM To: Matt MacKenzie Subject: Fw: [soa-rm] Help in Sending this Note: Feedback on SOA-RM Spec (09) Hello Matthew, I have not heard back from Duane regarding my comments below. Do you know if Duane on travel at this time or how we can follow-up on this issue and my feedback comments on the draft (rev 0.9) of the spec? It appears that my attempts to send Group e-mail is getting rejected. Thanks for your assistance, Regards... - Jeff Jeff A. Estefan Jet Propulsion Laboratory Office: (818) 393-5280 ----- Original Message ----- From: "Jeffrey A Estefan" <Jeffrey.A.Estefan@jpl.nasa.gov> To: <dnickull@adobe.com> Sent: Thursday, October 06, 2005 10:19 PM Subject: [soa-rm] Help in Sending this Note: Feedback on SOA-RM Spec (09) > Duane, > > My attempts to send an e-mail note to the soa-rm e-mail list keeps getting > thwarted by the mail server. I'm wondering if it's because I am currently > Applicant status. In the meantime, I wanted to send the following feedback > on the spec. > > Thanks... > > - Jeff > > ------------------ > > Editors, > > I will try and officially get these comments into the issues list but I know > the deadline for response was soon and I needed to provide some attachments. > Please see my feedback comments below. > > Regards... > > - J.A.E. > > Jeff A. Estefan > Jet Propulsion Laboratory > (818) 393-5280 > > ------- > > Editorial Suggestions: > 1. Make first instance of each concept definition in the Reference Model > subsections be of bold face type (e.g., 2.2.1 Service, 2.2.2 Service > description). > > 2. Consistent use of service-oriented architecture vs. service oriented > architecture (with/without hyphen). Is there an agreed upon convention? > > Observation: > There is no notion of "architectural style" in the definition of SOA. Nor > is there a definition of Architectural Style in the glossary. An SOA is an > example of an IT architectural style (just as there are other architectural > styles, e.g., event-driven architecture (EDA), client/server architecture, > information-centric architecture). One definition for Architectural Style > is as follows: > "A coordinated set of architectural constraints that restricts the > roles/features of architectural elements and the allowed relationships among > those elements within any architecture that conforms to that style." > > Finally, a few words about illustrations: > > 1. At a minimum, there should be a version of the illustration that Duane > has been using in his briefing slides and in the FAQ depicting SOA as it > relates to a larger SOA framework (e.g., reference architectures, SOA > implementations, etc.). [First attachment.] A future revision of the > SOA-RM document should show this diagram in UML 1.x or UML 2 format. > > 2. The object diagram that has been suggested depicting the key concepts of > the SOA-RM and their relationships covers primarily the functional elements > of SOA. They do not cover the non-functional elements, i.e., > quality-of-service (QoS) elements (e.g., security, transaction, system > management, etc.). In a future revision of the SOA-RM document, the QoS > concept elements and their relationship (as well as the any relationship to > the functional elements, typically through Policy) should be included. > > 3. One illustration that could be of great service to the community and > related to item 2 above would be a SOA stack diagram depicting the primary > functional and non-functional (QoS) elements. An example drawn from and IBM > Redbook I served on the review committee for early last year is attached. > This graphic would need to be updated to reflect the SOA-RM concepts. Stack > diagrams can indeed be contentious to develop, however, keeping simplicity > in mind , an example such as this has the benefit of a) serving as a > reference model on which implementation-specific specifications and > standards can be easily mapped (e.g., WSDL, SOAP, UDDI, SAML, XACML, WSRP, > etc.), and b) is extensible. It is probably too late to incorporate this > into the near term release of the SOA-RM specification document, but I'm > hoping to hear more input on this subject and whether serious consideration > should be given by the TC on adopting such a model for future revisions. >
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]