Agenda Bashing Sharon Chisholm asked for time to discuss syslog alarming document (draft-gerhards-chisholm-syslog-alarm-01.txt) KIKUCHI Yutaka asked for time to discuss/address tunnel work document during open mic after presentations? Dan Romanascu as AD - this presentaqtion has been made before (Chicago - IETF 69), and is more suited to IPPM No prob to re-presnt if thre is time, but doesn't think it belongs here in OPSAWG. Yutaka - IPPM chair is not sure it belongs there either, as it is an acive measurement. Scott Bradner as WG co-chair - if there is time we will discuss Dan Romanascu as AD - can we discuss initial charter draft (draft-ietf-opsawg-snmp-engineid-discovery-00.txt) Ted Seely as co-chair - as we had previously discussed, there is agreement to take to last call on mailing list - no objections in audience. --------------------------------- Agenda Presentations Dave Harrington - SMI to XSD Translations Take aways - make draft-li-natale-smi-datatypes-in-xsd-00 a OPSAWG doc (slide 10) Andy Bierman - question - is the output of the WG a doc that contains the algorithms for translations, or the module of types - becaue they are incompatible. Harrington - trying to limit it to only the translation of RFC 2578 & 9 Andy Bierman - that won't include addresses Sharon Chisholm - scope question. methods for defining netconf data types - seems like SMI translations want SNMP consistency, doesn't understand why you want to align? Dave Harringotn - not sure that what Yang and he (Dave) wants are the same thing. Sharon Chislom - we are disagreeing on XML type, not sure what the value is. Dave Harrington - if using an XML data structure to access MIB data, we want standardizeed XSD to acquire that. it is not to support NETCONF, or web based servcies, it is just if you are going to to from XML, here is a way to do it. If you are going to access MIB data, need a way to make sure you understand the baggage that comes with Sharon Chisholm - does comparing these things make sense? Dave Harringotn - BOF said yes. mail list said we don't want multiple versions of XSD flying ariound. wants to ask chairs to mod charter to admit/address goal is to get past issue of multiple versions of XSD - must i.d. diff version to do so Dan Romanascu - questioning the blocking effect? waht prevents progress / conclusions that it does more harm than good, vis a vis, Dave Harrington - lots of time spent debating - not sure its techical or personal Bob Natale - problem is we haven't taken use cases seriously. Need to push to ADs and use case for translating MIBS to XML via XSD. Disticnt point Netconf, rendering the XML for a given context. Leadership deceision? can we address these diff use cases and is there one decision that addresses all? be more forceful about use cases, and define them and when the baggage is appropriate? Andy Bierman - agrees with Bob. Use case for straight translation for using mgr tool for accessing snmp data w/o using snmp agent/tool? Security implications? Netconf issues are unique, and this issue isn't solving a general problem that is useful to everyone. Bob Natale - Asks is there a use case that is applicableto everyone? Dave Harrington - says if netconf wants to use a diff counter than SMI uses, then name it differently. Andy Bierman - says they do, Sharon - agrees with Bob and Andy and use cases, not 2 diff translations depending on how you look at it, doesn't think they're 2 diff translations for the same things because the use cases are very diff. Scott Bradner (as chair) - asks does the Yang team have anything to say in response, Respinders Name Unitelligable - agree that common data types would be nice - need to sit at table and see wjhat can be arrived at. Dave Partan - acknowledges and states to Dave Harrington & Bob Natale, want to work with you all and have same view Dave Harrington back on preso - diff use cases may drive diff XSDs, like to see otherwise, but... protocol indepenmdent would be best. Dave Harrington (?) - SHOULD HARMONIZING BE DONE IN OPSAWG? Scott Bradner (as chair) - asks do you have anything specific to be done? Dave harrington - start with doc, and move forward Bob Natale - about proposed approach - putting forth reqs that we expect to satisfy w/o considering why? Scott Bradner - IETF has been really good at max'ing defining req's, lets move forward. Bob Natale - Agree - this isn't a hard task Scott Bradner (as co-chair) - asks group - is this worth doing? Dave Harrington makes suggestion - lets compare some translations that have been done. Dave Harrinton's point is we are going for human readability Scott Bradner asks do we want to discuss that. (technology vs. human readability) Bob Natale- answerrs - in some cases comments or more info in definition of OID can impede and is meaningless to real reader. Scott Bradner - whats the balance? Bob Batale - important content vs. bloat comments. (summarized) Andy Bierman - does the doc deal with XML encoding issues? characters, etc...? Dave Harrington - not that he is aware of. Andy Bierman - at this level can it be ignored? Dave Harrington - not aware Phil Shaffer - making rules on how XML will be encoded? LIB SMI gives 2 ways to gen data , which path are you talking? SNMP as flat tables, or heirarchy? Dave harrington - doesn't know Bob Natale - plan in XMSDI to take another look at the smidump output options, and use what it does and incorporate that. Phil Shaffer - suggests heirarchical view Bob N - addressing file being displayed - and cosmetics of raw pattern vs, objects (left side is data types draft (romanscanu), right side is LIB SMI) Andy - not xsd friendly - xsd needs to be told how to process pattern (app info?) Jabber question - relayed by Dan R. for Dave Perkins - how do you figure out how to get heiarcical data from SNMP flatness Phil Shaffer - not my magic, but Juergen does it off the index code. Scott Bradners observation is at the high ordder bit they are very similar Dave Harrington - agrees, need tool generators to provide some guiadance readability readability by readers of the doc readability by readers/developers readability by tool users Andy Bierman has a question regarding current file being displayed - can you go over the bits data type (its diff)? Dave Harrington - netconf issues on how it is used with a union and an index. yang is more human friendly and would be, and friendliness to XSD LisT. Andy Bierman - lets come to an agreement and stick with it. Scott Bradner (as co-chair) - show of hands - consistent way of doing this is a worthwhile endeavor - yes (many) outweighs no's (zero) Open question - do it in this WG? Discuss with ADs to add to charter? Dan Romanascu as AD - draft applicability statement, etc... ? --------------------------------- OPEN MIC Sharon on SYSLOG - no slides - open discuss status of indiv. submission descritption of mapping defined SD output params resonable interest in CHI update & posted to mailing list wondering next steps? Scott Bradner as co-chair - what status Sharon Chisholm - standards track Dave Harrington - as chair of syslog WG - thinks its useful and not doing it there as syslog is in secutiry area - he supports the work in OPSAWG Ron Bonica as AD - please address informational versus proposed satndard Sharon Chisholm - defines some interoperability issues between vendors / alarm types , fields, tags , severity mapping diffs... Gene Govlinksy(SP?) - move forward with it Dan Romanascu as contrib. - believes its a useful doc as well Sharon Chisholm - doc aligns with ITU & Alarm MIBs. Scott Bradner as co-chair - can you do a sanity check with ITU since Alarm'ing types are near and dear to them Scott Bradner as co-chair - asks for show of hands for standards track - hands up - many yes's, zero no's Dan Romanascu as AD - draft naming led to confusion - will be taken as a WG work. ------------------------ KIKUCHI Yutaka discussing tunnel measurement performance draft (draft-kikuchi-tunnel-measure-req-02.txt) discussing two IDs focus on OAM point of view split from one ID into 2 prague ietf - IPPM not the right place (no OPSAWG then) (Basically looking for a home) draft has been rev'd should it be discussed here? Kikuchi thinks the 1st one belongs here but not metrics piece metrics alternatives not IPPM not within PMOLs charter (charter?) Scott Bradner as co-chair asks - can you address comments on mailing list in summary here? Yutaka thinks he already did Scott Bradner - terminology not clear enough - too ambiguous? Scott Bradner as co-chair - is working on a doc that defines perf. of measurement of tunnels to interest in the room. yes's 5 , no's 1, many nuetral. not comfortable expanding charter to accept the wrk. last call dan (AD) - ops & mgmt'abilty draft - dave h can you speak too? dave h - big issue from last mtg need input from others, Dave Kessens will deliver his review to Dave Harrington tomorrow presntation tomorrow in WG Chairs Lunch these are guidlelines for Ops & Mgmt for other WGs and protocol developers. Scott Bradner as co-chair asks - timeframes? Dave Harringotn - still looking for feedback from operators still looking for more reviewers in operator land Scott Bradner as co-chair - if you are comfortable after review, we can ask for last call Dave Harrington - we'll try to see if we can complete it within 30 days. ------------------------ mic closed session adjourned