February 2013 - Nsi-wg - lists.ogf.org

Re: [Nsi-wg] Issue 5 in ogf-nsi-project: Query on requester interface - is this needed?
by ogf-nsi-project＠googlecode.com 22 May '13

22 May '13

Comment #7 on issue 5 by imo...(a)gmail.com: Query on requester interface - is this needed? http://code.google.com/p/ogf-nsi-project/issues/detail?id=5 This discussion was not concluded, but I will capture the thoughts as were shared on the call: 1. Query as a separate WSDL was discussed, issue 22 has also been opened for this 2. The main use-case was monitoring or management agents that might not be NSAs. Unanswered questions are: a. Can query work if the entities querying are not NSA agents? b. Should query be a separate WSDL in any case? c. Which credentials - user credentials or peer relationships when walking the service tree? 3. There was support on the call for the need for query in both directions whether it be commercial networks or research networks scenario. This discussion was not deemed important for 1.0SC and resolution postponed till another date Inder

1 2

Issue 55 in ogf-nsi-project: correlationId is unneeded and complicates implementation significantly
by ogf-nsi-project＠googlecode.com 01 Apr '13

01 Apr '13

Status: Accepted Owner: ---- Labels: Type-Review Priority-Medium FoundInVersion-1.0 FixedInVersion-1.1 New issue 55 by thost...(a)gmail.com: correlationId is unneeded and complicates implementation significantly http://code.google.com/p/ogf-nsi-project/issues/detail?id=55 Description of Issue: The correlationId attribute in the WSDL messages is unneeded. Furthermore it requires correct implementations to persistently keep track of the correlationsIds of active requests. Discussion of Issue: The WSDL adds a correlationId to track the correspondence between a request and the subsequent update for the status. However this correspondence is unneeded as the request already contains information about the connection in question. However having the correlationId requires both the requester and provider to keep a persistent (as in on-disk) record of currently active requests, in order to be crash and restart safe, i.e., the service must be able to respond to requests that came before it crashed/restarted. The correlationId also means that state updates are not idempotent as the client can only receive a request (state update) once. However it cannot be guaranties that a message with a state update will only be send once due to the coordination of message dispatching and updating this information persistently, where there will always be a window which can lead to inconsistency, where the message must be retransmitted. Having to keep track of requests in this way, is called a state-full protocol (note that this does not mean a non state-full protocol doesn't have any state), and is usually considered a bad idea do implementation complexity and poor scalability in high-performance systems (arguably this is not what we have though), as it makes load balancing and redundancy significantly more tricky. Furthermore it must be remembered that not all requests will have a callback. E.g., a provision before start time (i.e., auto-provision) followed by subsequent terminate request, will result in a callback for terminate, but not for the provision request. Here the client must then clear all pending correlationIds, complicating the implementation (the client side issues are simple though). All this is further complicated by the fact that we haven't established if it is the provider or requester which is responsible for ensuring that correct connection state is updated. Resolution of Issue: Remove correlationId from the WSDL. This will make the state update messages (reserveConfirmed, etc.) idempotent, meaning they can be repeated. The requester will then update any state references about the remote connection to fit with the last received message (this also has problems, but none that doesn't exist in the current protocol).

1 3

Re: [Nsi-wg] Issue 41 in ogf-nsi-project: Do we replace the replyTo field with a topology/NSA configuration lookup?
by ogf-nsi-project＠googlecode.com 05 Mar '13

05 Mar '13

Comment #3 on issue 41 by takahiro...(a)gmail.com: Do we replace the replyTo field with a topology/NSA configuration lookup? http://code.google.com/p/ogf-nsi-project/issues/detail?id=41 I want to use the replyTo field. Currently, all NSA act as requester and provider. In this case, we are able to look up the csProviderEndpoint of requesters easily. However, in near future, some NSA act as only a requester, e.g. PerfSONAR and AutoEarth. It is difficult to describe and manage all requester endpoints as same as provider ones. Moreover, it causes a high barrier to publish new requesters. Therefore, a requester endpoint should be obtained using the replyTo field in the SOAP message payload.

1 7

Overly complex discovery WSDL
by Henrik Thostrup Jensen 04 Mar '13

04 Mar '13

Hi I've been looking at the discovery WSDL, and was a bit amazed with how complicated it turned up being. Now I have not been very active in NSI for the last couple of months, and have missed a couple of meeting, but I think we've ended up with something a lot more complex than what we discussed in Oxford and the mailing lists. For the request it looks like this: queryNsaRequest :: QueryNsaRequestType queryNsaRequest :: RequestType filter :: FilterType * nsaId :: NsaIdType * serviceType :: xsd:string * This is my own notation. Sorry if it is tricky to read. For the request it possibly to define summary of detailed request, but I haven't found any documentation on what should be returned for what. Secondly it is possible to have multiple filters, each specifying multiple nsaIds and serviceTypes. The parameters for each filter should be ORed, however there does not appear to be any documentation for how to deal with multiple filters. It also appears to be complete and utterly overkill as most NSAs will host 2 or 3 services. Now, for the response: queryNsaResponse :: QueryNsaResponseType discovery :: NsaDiscoveryType * # nsaId # softwareVersion # startTime # currentTime services :: ServicesType * service :: ServiceType #other :: any description :: xsd:string versions :: VersionsType * version :: VersionType # name # version # endpoint # wsdl # url # other :: any other :: any * capabilities :: CapabilitiesType * capability ft:TypeValuePairType # type :: xsd:string ! # targetNamespace :: xsd:anyURI value :: xsd:string * other :: any * Here we are dealing with a datastructure potentially 9 levels deep. Arguably, some of this is due to the ackward way XML/WSDL handles lists, but there are still 4 levels of nested lists. I won't go into any details here. I am not trying to do a witch-hunt on John here. In fact he may just have been listening to everyone, but what we ended up with here is pretty bad, and is in my opion very far from what we discussed in Oxford. I think we need to setup some review of WSDLs as well in order to ensure that these things do not spin out of control like this. Anyways, something to talk about tomorrow :-). Best regards, Henrik Henrik Thostrup Jensen <htj at nordu.net> Software Developer, NORDUnet

2 3

Release command
by John MacAuley 01 Mar '13

01 Mar '13

Peoples, I understand the usefulness of having a release command from a networking perspective, but I know a number of implementations are not supporting it at the moment (and some have no intention of supporting it). With my new "lets make the state machine as simple as possible attitude" I started to wonder what is the customer use case for this command? Chin had one where someone wanted to keep their reservation, but wanted to free up bandwidth for others to use. This is an interesting use case, but not the behaviour of the current command as documented. Removal of this command can significantly simplify the protocol and state machine. Would it be possible for people to provide their customer use cases for this command? I am wondering if it is truly needed, or perhaps, we have the incorrect behaviour defined. Thank you, John

5 7

minutes from today's NSI call
by Guy Roberts 27 Feb '13

27 Feb '13

.... Are available here: https://redmine.ogf.org/dmsf/nsi-wg?folder_id=6529 _____________________________________________________________________ Guy Roberts PhD Senior Transport Network Architect DANTE Cambridge, UK +44 1223 371316 DANTE is the project co-ordinator and operator of GÉANT, the high-speed pan-European research and education network that is transforming the way researchers collaborate. Learn more at: www.geant.net<http://www.geant.net/> Like us on: www.facebook.com/GEANTnetwork<http://www.facebook.com/GEANTnetwork> Follow us at: www.twitter.com/GEANTnews<http://www.twitter.com/GEANTnews> DANTE is the trading name of Delivery of Advanced Network Technology to Europe Limited registered in England & Wales. Registration Number 2806796. Registered Office - 9400 Garsington Road, Oxford Business Park, Oxford OX4 2HN. _____________________________________________________________________

1 0

Wednesday's NSI conf call
by Guy Roberts 26 Feb '13

26 Feb '13

The following is dial-in information for Wednesday's NSI call, Time: 7:00 PDT 10:00 EDT, 15:00 GMT, 16:00 CET, 24:00 JST 1. Dial Toll-Free Number: 866-740-1260 (U.S. & Canada) 2. International participants dial: Toll Number: 303-248-0285 Or International Toll-Free Number: http://www.readytalk.com/intl 3. Enter 7-digit access code 8937606, followed by "#" Agenda: - Discovery service - Preparation for OGF37, assignments of docs - Other tba _____________________________________________________________________ Guy Roberts PhD Senior Transport Network Architect DANTE Cambridge, UK +44 1223 371316 DANTE is the project co-ordinator and operator of GÉANT, the high-speed pan-European research and education network that is transforming the way researchers collaborate. Learn more at: www.geant.net<http://www.geant.net/> Like us on: www.facebook.com/GEANTnetwork<http://www.facebook.com/GEANTnetwork> Follow us at: www.twitter.com/GEANTnews<http://www.twitter.com/GEANTnews> DANTE is the trading name of Delivery of Advanced Network Technology to Europe Limited registered in England & Wales. Registration Number 2806796. Registered Office - 9400 Garsington Road, Oxford Business Park, Oxford OX4 2HN. _____________________________________________________________________

1 0

Re: [Nsi-wg] Topology comments
by Jeroen van der Ham 19 Feb '13

19 Feb '13

Hi, Thank you for you comments, I've taken most of them and have updated the document. A newer version is located at: https://redmine.ogf.org/dmsf_files/12981?download= On 13 Feb 2013, at 15:54, Jerry Sobieski <jerry(a)nordu.net> wrote: > objection: For _/NSI/_ the local ID need not be globally unique. For NSI the local id only needs to be unique within the scope of the network to which it belongs. > > The global uniqueness for local id object is an _/NML/_ requirement - not an NSI requirement. We should not impose this on NSI as it as it fundamentally changes the NSI framwork. > > We can state that the NML expects objects to have globally unique URNs, and that this is compatible with NSI topology. But that NSI does not impose the globally unique requirement onto the local ID portion. The NSI imposes a globally unique identifier for the Topology. Indeed the Port(Group) identifiers are not necessarily unique, but should be locally unique, and the combination of Topology and Port must be globally unique. > note: there is no way to verify that either the topology object or the localid object (ports) names are in fact globally unique. We simply use them. If we require ("must be") something in NSI, there must also be an agreed upon or specified means of determining if an identifier meets those requiremetns before it is used. Use without validation can create security holes that are easily exploited. If we require this, then we should state this as a security concern. Ensuring or verifying that something is globally unique is not possible. > comment: THis section should define what the source/dest identifiers are meant to do with respect to NSI: They indicate where a connection terminates (or originates), and thus must specifiy the NSI network domain in which the endpoint exists, and provide further resolution to identify the end point within that domain. Then map those functional requirements to the mechanisms for identifying each of those elemtns of the STP. No. This document specifies syntax and minimal context, nothing more. > > *2.3 THis seems convoluted.* Is this necessary? Why can't we have a simple list of TV pairs? What do you mean simple list of TV pairs? > > Further, how would one specify a list of Ports - any of which are valid termination points? Or a list of VLANS that extend across a disjoint set of ports that may be on different switches? Not possible in the current syntax. > > I am concerend that this syntax is I think too rooted in an ethenret VLAN service expectation. We need a more comprehensive and generalized means of identifying the set of terminals that can be used for an endpoint (or we will regress back to the exhaistive search problem.) The TVP pairs can do this if we eleboprate on how they build constraints that identify the candidate endpoints. No. It is rooted in the basic assumption that a group of ports is distinguished by a single labelset, whatever that labelset is. We have been only been able to define this deterministically with a single labelset (type-value pair). > *3.1.2 What is difference between "defined" relation and "allowed" relation?* an "allowed" relation is syntactically valid, but does not currently have any meaning associated to it. This means that a parser should at least accept it, and can then ignore it, but it can't throw an error. > 3.2.3 WHy do we have peersWith? While it makes sense for persistent NSAs managing networks or federations, plain old users "peerWith" other NSAs to make requests. So perhaps we should be clearer about what the peerswith relation really means or ought to mean. We want to be able to express control plane connections right? I welcome a better description here. > > *Sect 4. * Personal subjective thought: I understand the N3 syntax. But it seems rather convoluted just the same. The NSI concepts are very simple - why can't we find an expression of these simple concepts in a simple syntactic structure? While experts in this field may be able to sort thru this - it will be very difficult for network engineers to construct this who are not knowledgable of NDL or programmers or computer scientists. > > It seems to me there must be an easier more intuitive means of defining the NSI service environment. I think the Port/PortGroup/STP Group approach as defined is confusing as there is no underlying consistentcy to what a "Group" inherits from the base object. A group is not a grouping of another object - it seems to be a defined object in itself...which is confusing. An "intuitive definition" does not have a place in a standard. A Group is a special kind of object, which groups together other kinds of objects. You do not want to have a Group behaving like any other kind of object, except in very specific cases. This is why it is constructed like that. > Grouping is an important concept and I support it - but it needs to be generalized as a functional element. Grouping *is* a general concept. > Sect 4 page 8: The Location element is indeed useful, but I think we need to scope it better - i.e. it makes sense if the object associated with it is an object whose physical/geogrphic dimension is "small", e.g. a switch. It is less clear how it would apply to an entire Topology if that topology spanned a whole continent. If it is present for say NORDUnet.ets, what should we infer by that given that ports may be present in two different continents? The NML standard allows you to relate a Location object to any Network Object, so you can scope it any way you want. > Sect 5: If you assert an idenitifier must be globally unique - you must define how that is determined or validated. Otherwise, you put each NSA at risk of accepting a non-globally unique object name (for instance). If the NSI protocols require uniqueness, then it should be clearly stated why, and how that can be verified. Else, using a fake ID can cause protocl malfunction - a serious security problem. It is not determined or validated, and specifically stated as such in the URN, NML and NSI Topology standards security considerations. I am not aware of any feasible way of determining or validating globally uniqueness, nor does there seem much incentive in the NSI group to create something like that. Jeroen.

1 0

NSI conf call Wednesday
by Guy Roberts 19 Feb '13

19 Feb '13

The following is dial-in information for Wednesday's NSI call, Time: 7:00 PDT 10:00 EDT, 15:00 GMT, 16:00 CET, 24:00 JST 1. Dial Toll-Free Number: 866-740-1260 (U.S. & Canada) 2. International participants dial: Toll Number: 303-248-0285 Or International Toll-Free Number: http://www.readytalk.com/intl 3. Enter 7-digit access code 8937606, followed by "#" Agenda: ・ JvH & JS: Topology exchange document ・ Preparation for OGF37, assignments of docs ・ Other tba _____________________________________________________________________ Guy Roberts PhD Senior Transport Network Architect DANTE Cambridge, UK +44 1223 371316 DANTE is the project co-ordinator and operator of GÉANT, the high-speed pan-European research and education network that is transforming the way researchers collaborate. Learn more at: www.geant.net<http://www.geant.net/> Like us on: www.facebook.com/GEANTnetwork<http://www.facebook.com/GEANTnetwork> Follow us at: www.twitter.com/GEANTnews<http://www.twitter.com/GEANTnews> DANTE is the trading name of Delivery of Advanced Network Technology to Europe Limited registered in England & Wales. Registration Number 2806796. Registered Office – 9400 Garsington Road, Oxford Business Park, Oxford OX4 2HN. _____________________________________________________________________

1 0

MTL and MH
by Tomohiro Kudoh 12 Feb '13

12 Feb '13

Hi all, Here is my slide on MTL and MH. Tomohiro

2 2