I ran some BBR tests under different scenarios and I am concerned that
under some conditions BBR can become very aggressive. It seems BBR flows
can decide that losses are un-correlated to congestion, and when in this
mode, they will maintain large cwnds regardless of the losses. In cases
when BBR i
On Tue, Sep 20, 2016 at 2:50 PM, Neal Cardwell wrote:
> On Tue, Sep 20, 2016 at 2:48 PM, Stephen Hemminger
> wrote:
>>
>> On Mon, 19 Sep 2016 23:39:23 -0400
>> Neal Cardwell wrote:
>>
>> > +/* INET_DIAG_BBRINFO */
>> > +
>> > +struct tcp_bbr_info {
>> > + /* u64 bw: max-filtered BW (app thro
On Tue, Sep 20, 2016 at 4:42 PM, Neal Cardwell wrote:
>
> On Tue, Sep 20, 2016 at 7:39 PM, Stephen Hemminger
> wrote:
> >
> >> NOTE: BBR *must* be used with the fq qdisc ("man tc-fq") with pacing
> >> enabled, since pacing is integral to the BBR design and
> >> implementation. BBR without pacing
On Tue, Sep 20, 2016 at 7:39 PM, Stephen Hemminger
wrote:
>
>> NOTE: BBR *must* be used with the fq qdisc ("man tc-fq") with pacing
>> enabled, since pacing is integral to the BBR design and
>> implementation. BBR without pacing would not function properly, and
>> may incur unnecessary high packet
> NOTE: BBR *must* be used with the fq qdisc ("man tc-fq") with pacing
> enabled, since pacing is integral to the BBR design and
> implementation. BBR without pacing would not function properly, and
> may incur unnecessary high packet loss rates.
Does it work with fq_codel?
On Tue, Sep 20, 2016 at 11:48 AM, Stephen Hemminger
wrote:
> On Mon, 19 Sep 2016 23:39:23 -0400
> Neal Cardwell wrote:
>
>> +/* INET_DIAG_BBRINFO */
>> +
>> +struct tcp_bbr_info {
>> + /* u64 bw: max-filtered BW (app throughput) estimate in Byte per sec:
>> */
>> + __u32 bbr_bw_lo;
On Tue, Sep 20, 2016 at 2:48 PM, Stephen Hemminger
wrote:
>
> On Mon, 19 Sep 2016 23:39:23 -0400
> Neal Cardwell wrote:
>
> > +/* INET_DIAG_BBRINFO */
> > +
> > +struct tcp_bbr_info {
> > + /* u64 bw: max-filtered BW (app throughput) estimate in Byte per sec:
> > */
> > + __u32 bbr_bw_
On Mon, 19 Sep 2016 23:39:23 -0400
Neal Cardwell wrote:
> +/* INET_DIAG_BBRINFO */
> +
> +struct tcp_bbr_info {
> + /* u64 bw: max-filtered BW (app throughput) estimate in Byte per sec: */
> + __u32 bbr_bw_lo; /* lower 32 bits of bw */
> + __u32 bbr_bw_hi;
This commit implements a new TCP congestion control algorithm: BBR
(Bottleneck Bandwidth and RTT). A detailed description of BBR will be
published in ACM Queue, Vol. 14 No. 5, September-October 2016, as
"BBR: Congestion-Based Congestion Control".
BBR has significantly increased throughput and redu