------------------------ Reviewers Comments
-----------------------------

Reviewers Familiarity:               9
Evaluation of work and contribution: 9
Significance to theory and practice: 9
Originality and novelty:             7
Relevance for NET-COOP:               10
Readability and organization:        9
Overall recommendation:              9

**(0 = no mark set by reviewer)**

Summary:
                   The paper makes an original and important
contribution to congestion control and fits nicely in the
program of NET-COOP.

Comments to the author:
          The paper introduces the Multi-Level Comparison
(MLC) queue  management scheme. MLC is a stateless active
queue management scheme that makes TCP/IP network to
approach max-min fairness. The paper is a very important
contribution and it is well written. Therefore, the reviewer
definitely recommends to accept it.



---------------------------------------


Reviewers Familiarity:               3
Evaluation of work and contribution: 7
Significance to theory and practice: 5
Originality and novelty:             6
Relevance for NET-COOP:               7
Readability and organization:        9
Overall recommendation:              6

**(0 = no mark set by reviewer)**

Summary:
                   This paper is clear and well written. It
perfectly fits main topics of this Conference. The authors
present a class of stateless AQM schemes called MLC(l), with
l as an index. They show that for an arbitrary network
topology of TCP flows, the resource allocation obtained
using MLC(l) converges to max-min fair as l increases.

Comments to the author:
         


---------------------------------------


Reviewers Familiarity:               7
Evaluation of work and contribution: 7
Significance to theory and practice: 6
Originality and novelty:             6
Relevance for NET-COOP:               8
Readability and organization:        8
Overall recommendation:              8

**(0 = no mark set by reviewer)**

Summary:
                   Theauthors propose a multi level
comparison AQM algorithm for routers, and provide
performance analysis for this. The algorithm is essentially
randomly samples the queue at each incoming packet and drops
the packet depending on the number of samples that belong to
the same flow as te arriving packet. The authors show that
this scheme provides a power-law drop (power law wrt the
throughput of flows).   The performance analysis seems solid
in Sec 2.2 and there is sufficient algorithmic novelty in
this paper.

Comments to the author:
         


---------------------------------------


============== END OF REVIEWS ===============