Early Review of draft-ietf-ccamp-ospf-availability-extension-05
review-ietf-ccamp-ospf-availability-extension-05-rtgdir-early-hardwick-2016-07-27-00

Hello,



I have been selected as the Routing Directorate reviewer for this draft. The
Routing Directorate seeks to review all routing or routing-related drafts as
they pass through IETF last call and IESG review, and sometimes on special
request.
 The purpose of the review is to provide assistance to the Routing ADs. For
 more information about the Routing Directorate, please see ​

http://trac.tools.ietf.org/area/rtg/trac/wiki/RtgDir



Although these comments are primarily for the use of the Routing ADs, it would
be helpful if you could consider them along with any other IETF Last Call
comments that you receive, and strive to resolve them through discussion or by
updating
 the draft.



Document: draft-ietf-ccamp-ospf-availability-extension

Reviewer: Jon Hardwick

Review Date: 26 July 2016

IETF LC End Date: Not started yet

Intended Status: Proposed standard



== Summary ==

I have significant concerns about this document and recommend that the Routing
ADs discuss these issues further with the authors.



== Comments ==

This document addresses the situation where the maximum bandwidth of some links
in the network can vary over time, for example due to radio interference. 
Specifically, it addresses the path computation problem, where service placement
 must take these fluctuations into account.



OSPF is extended to flood the “availability” of such links, which is defined as
a sequence of <bandwidth, percentage> tuples.  Each tuple tells the percentage
of time at which the bandwidth of the link will be at least as much as
<bandwidth>,
 for example <100Mbps, 99.99%>.  This information is added to the ISCD TLV,
 defined in RFC 4203.  When services are being planned, the planner specifies
 the service’s availability requirement (e.g. 4-nines) and the path is computed
 accordingly.



The document is written in good English and easy enough to understand, and is
commendably brief, although probably too brief – see below.



== Major Issues ==



1) The specification of the protocol extensions appears incomplete. 
Specifically, section 1 says:



The extension

   reuses the reserved field in the ISCD and also introduces an

   optional Availability sub-TLV.



However, no further mention is made of the ISCD’s reserved field.  How is it
reused?



The following description of the Availability level is difficult to understand
and leaves me with several questions:

           This field is a 32-bit IEEE floating point number which

           describes the decimal value of availability guarantee of the

           switching capacity in the ISCD object which has the AI value

           equal to Index of this sub-TLV.



- What is the “switching capacity” of the ISCD object?  Do you mean “switching
capability” like PSC-1 etc.?

- What is the AI value?  This term is not defined.

- What do you mean by the index of this sub-TLV?  Do you mean the position of
this sub-TLV in the list of ISCD sub-TLVs?  What if somebody defines a new ISCD
sub-TLV in future – won’t that mess up the indexing?



2) The proposed extensions are not backwards compatible, so do not support
incremental deployment.  In section 3.2 it says:



   A node that doesn't support ISCD Availability sub-TLV SHOULD ignore

   ISCD Availability sub-TLV.



However, it’s not possible for this document to specify the behaviour of nodes
that do not support this document.  The issue is that the ISCD Availability TLV
is defined as a sub-TLV of the ISCD TLV, but the ISCD TLV defined in RFC 4203
 has no sub-TLVs.  Therefore, implementations of RFC 4203 that do not support
 this draft will reject the ISCD TLVs sent by implementations of this draft,
 because they will appear to have the wrong length.



== Minor Issues ==



There are a few cases I would like to see spelled out more clearly in the
document.

- How should a node interpret the absence of any ISCD Availability TLVs?

- How should a node interpret more than one ISCD Availability TLV for the same
availability level?  Is that even legal to send?

- Let’s say a node advertises a link with 10Gbps capacity, and includes an ISCD
Availability TLV saying that the link has a 99% availability level at 8Gbps . 
What (if anything) are we to conclude about its availability at 10Gbps?  At
5Gbps?



Section 5 (IANA Considerations).  What allocation policy should IANA use for
the new sub-registry?  See RFC 5226.



Best regards

Jon