On Jul 6, 2010, at 5:18 AM, Lixia Zhang wrote:
On Jul 5, 2010, at 1:52 PM, Paul Francis wrote:
Hi Gang,
Zartash Uzmi and myself have posted an internet draft of FIB
aggregation
that is basically a follow on to Lixia et.al. draft
http://tools.ietf.org/html/draft-zhang-fibaggregation-02.
Note that the "zhang" in draft-zhang-fibaggregation-02 is really
Beichuan Zhang.
It's mostly Beichuan and Lan Wang's work.
You can find it at http://www.ietf.org/id/draft-uzmi-
smalta-00.txt. We
expect to give a short presentation of it at the next GROW meeting.
Any comments appreciated.
will read and comment.
Since ORTC is the optimal algorithm that does not introduce extra
routable space, I'm interested in the difference between this work
and ORTC. Just did a quick search in the draft and found the following:
-----------
SMALTA's one-shot aggregation is derived from ORTC
[Paper.Draves-ORTC]. Like its precursor, one-shot runs three passes
over a tree data structure: ... [LW: omitted some text here]
Unlike ORTC, SMALTA one-shot is not provably optimal, though it is
very close to optimal. This is because SMALTA places some minor
constraints on the ORTC algorithm in order to allow the aggregated
table and the original non-aggregated table to be implemented as a
single data structure.
---------
Beichuan can correct me, but I think ORTC algorithm allows the
aggregated table and the original table to be implemented as a single
data structure. In fact, we have implemented ORTC using a patricia
trie. We have a paper to appear in Globecom that talks about this.
Lan
_______________________________________________
GROW mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/grow
