[RRG] draft-farinacci-lisp-05

[Iljitsch van Beijnum <iljitsch@muada.com>]

A few comments:

There is no version field in the LISP header. There should be. At  
least a few "set to zero on send, ignore on receive" bits that can be  
used for extensions without bumping the version number is also a good  
idea.

"In order to eliminate the need for a mapping lookup in the reverse  
direction, the ETR gleans RLOC information from the LISP header."

I find this undesirable because this way, the behavior of the system  
can be different depending on which end initiated the communication.  
Trusting information supplied directly by the other end is also  
problematic security-wise. I would much prefer it if both ends did an  
independent mapping lookup.

"LISP Locator Reach Bits: in the LISP header are set by an ITR to  
indicate to an ETR the reachability of the Locators in the source site."

I wonder how useful this is in practice. First of all, having 32 RLOCs  
is way too many. Second, depending on the way xTRs are deployed (see  
my next message), it could be quite hard for one xTR to know the  
status of any others for a particular EID. But more fundamentally,  
this assumes that reachability is a binary property: something is  
reachable or it isn't. In reality, it's more complex: a certain  
location may be reachable from some parts of the network and not from  
others. Even if it is reachable, it may be preferable to use a  
different RLOC. As such, I think it makes much more sense to determine  
which RLOC is going to be used by a given ITR for a given EID on a  
case-by-case basis rather than broadcast some reachability bits.

There are also potential issues with synchronizing the numbering of  
the RLOCs in the mapping system and in this field.

"Record TTL: The time in minutes the recipient of the Map-Reply will  
store the mapping. If the TTL is 0, the entry should be removed from  
the cache immediately. If the value is 0xffffffff, the recipient can  
decide locally how long to store the mapping."

Don't think this is a good idea. Experience with the DNS TTL has shown  
that people may be tempted to set both unreasonably low TTLs or ignore  
TTLs. Mandated minimum and maximum caching times would make the  
protocol more deterministic and remove an opportunity for people to  
shoot themselves in the foot.

" Priority: each RLOC is assigned a priority. Lower values are more  
preferable. When multiple RLOCs have the same priority, they are used  
in a load-split fashion. A value of 255 means the RLOC MUST NOT be  
used."

Isn't it more natural to use 0 as the special case value?

"Note that the destination RLOC address MAY be an anycast address if  
the tunnel egress point may be via more than one physical device."

This makes me rather uncomfortable, as anycasting could get in the way  
of reachability testing. We have room for a number of RLOCs, why not  
put in RLOCs for all physical ETRs rather than anycast them?

Is the "Loc-AFI" field the AFI for an RLOC? Why is this field only 8  
bits while other AFI fields are 16 bits?

"2. Locator unreachability is determined by an ITR by receiving ICMP  
Network or Host Unreachable messages."

These can be spoofed or filtered/not generated, so both their presence  
and their absence don't necessarily mean anything. I suggest the only  
action triggered by such a message should be a one-time reduction in  
the time until a new mapping request is done. (I.e., if this happens  
every 300 seconds, reduce by 240 seconds for the first ICMP message  
but not subsequent ones, which means an immediate request 80% of the  
time but a maximum of 1 request per 60 seconds.)

"3. ETR unreachability is determined when a host sends an ICMP Port  
Unreachable message."

Which host? The one holding the EID? Or the ETR? The former won't  
happen if the ETR is unreachable, and an ETR being connected to the  
network but not being able to function as an ETR seems like a corner  
case to me.

So basically the use of mapping requests/replies is the only reliable  
(implicit) reachability detection mechanism. However, it is largely  
unspecified how ITRs should use this mechanism to determine  
reachability.

"perform a route-returnability check"

Return routability check?

"the use of a 6-byte Nonce field in the LISP encapsulation header"

The nonce field is only 32 bits in the LISP header earlier in the  
document.

" In practice, this is not really a problem. Hosts typically do not  
originate IP packets larger than 1500 bytes. And second, an informal  
survey of ISPs has been taken where nearly all ISP link MTUs are  
either 4470 bytes or support Ethernet jumbo frames of 9180 bytes.  
Therefore, we don't anticipate any problems with prepending additional  
headers."

Even if we assume that all the links within a transit AS support  
sufficiently large MTUs, this doesn't address the situation where ISPs  
interconnect over a shared ethernet infrastructure (i.e., an internet  
exchange). Also, the placement of ITRs/ETRs is left fairly open, with  
the suggestion that ETRs can be placed in end-user networks. In that  
case, it's very likely that there is at least one hop in the path that  
is limited to a 1500-byte MTU. I think it would be helpful to  
explicitly signal the MTU and possibly the maximum size of packets  
that can be fragmented that an ETR supports back to ITRs.

In my opinion, it would be beneficial to remove pretty much all of the  
text that doesn't pertain to actual LISP operation from this draft,  
and move that which is still useful (some of it is a bit stale after  
five iterations) to a new "LISP architecture" document.

--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg