netdev - Re: TCP Vegas per-socket alpha, beta, gamma parameters

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <AD8C2BD7-0633-4927-8155-D4297D7B71C1@fb.com>
Date:   Mon, 19 Jun 2017 14:30:00 +0000
From:   Lawrence Brakmo <brakmo@...com>
To:     Lars Erik Storbukås <storbukas.dev@...il.com>,
        Netdev <netdev@...r.kernel.org>
Subject: Re: TCP Vegas per-socket alpha, beta, gamma parameters

Hi Lars,

You would need to modify the source code to support per connection alpha, beta and gamma parameters. Although you can modify the parameters at runtime, they would apply to all connections using Vegas, both existing and new ones. 

You may also want to consider using TCP NV (New Vegas) instead. It works better in current networks and servers.

- Lawrence Brakmo
  

On 6/19/17, 6:48 AM, "netdev-owner@...r.kernel.org on behalf of Lars Erik Storbukås" <netdev-owner@...r.kernel.org on behalf of storbukas.dev@...il.com> wrote:

    I'm looking into how I can modify alpha, beta and gamma parameters for
    TCP Vegas on a per-socket basis (if at all possible). I want to modify
    the sending rate per-socket dynamically  (not system wide) when a
    delay-based algorithm is used.
    
    Is it possible to modify the parameters of Vegas without modifying the
    source code of Vegas itself?
    
    Would setting the (system wide) alpha, beta and gamma parameters be
    updated immediately? Would an ongoing transmission get the new values
    of alpha, beta or gamma when they are modified, or is it read only
    when initialising a Vegas transmission?
    
    Background for question:
    I'm working on the implementation of a Deadline Aware, Less than Best
    Effort. A framework for adding both LBE behaviour and awareness of
    “soft” delivery deadlines to any congestion control (CC) algorithm,
    whether loss-based, delay- based or explicit signaling-based. This
    effectively allows it to turn an arbitrary CC protocol into a
    scavenger protocol that dynamically adapts its sending rate to network
    conditions and remaining time before the deadline, to balance
    timeliness and transmission aggressiveness.
    
    / Lars Erik Storbukås (storbukas.dev@...il.com)