From: David Miller <davem@davemloft.net>
To: eric.dumazet@gmail.com
Cc: christoph.paasch@uclouvain.be, netdev@vger.kernel.org,
ycheng@google.com, ja@ssi.bg
Subject: Re: [PATCH net-next v2 2/5] tcp: metrics: Add source-address to tcp-metrics
Date: Fri, 10 Jan 2014 17:37:11 -0500 (EST) [thread overview]
Message-ID: <20140110.173711.1541969326146414340.davem@davemloft.net> (raw)
In-Reply-To: <1389222826.31367.28.camel@edumazet-glaptop2.roam.corp.google.com>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Wed, 08 Jan 2014 15:13:46 -0800
> On Wed, 2014-01-08 at 23:43 +0100, Christoph Paasch wrote:
>> Hello Eric,
>>
>> On 08/01/14 - 09:55:51, Eric Dumazet wrote:
>> > On Wed, 2014-01-08 at 16:05 +0100, Christoph Paasch wrote:
>> > > We add the source-address to the tcp-metrics, so that different metrics
>> > > will be used per source/destination-pair. We use the destination-hash to
>> > > store the metric inside the hash-table. That way, deleting and dumping
>> > > via "ip tcp_metrics" is easy.
>> >
>> > Note that this has the following problem :
>> >
>> > Some applications use a set of source IP addresses to overcome the 64K
>> > port limitation.
>>
>> Ok, did not know about that.
>>
>> > tcp_metrics uses a hard-coded TCP_METRICS_RECLAIM_DEPTH value of 5,
>> > meaning that cache wont be able to store more than 5 source IP addresses
>> > (reaching one particular remote IP).
>>
>> Maybe we could do something like the below (yet untested). That way we allow
>> up to 32 entries with the same destination but different source and still
>> only 5 with different destinations.
>>
>> I guess 32 * 64K connections is enough. :)
>> We could also make TCP_METRICS_RECLAIM_DEPTH(_DST) a tunable.
>
> Well, not sure if this is a problem anyway, and if we want extra
> complexity for this rare use case, considering tcp metrics for
> high number of flows sharing a common path is unlikely to be useful
> (with exception of Fast Open, but again it must be rare)
I think we are overthinking this, even for the aforementioned case.
If people report that this is a real problem they are hitting, and
not just with constructed test cases, we can work on a solution.
next prev parent reply other threads:[~2014-01-10 22:37 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-01-08 15:05 [PATCH net-next v2 0/5] Make tcp-metrics source-address aware Christoph Paasch
2014-01-08 15:05 ` [PATCH net-next v2 1/5] tcp: metrics: rename tcpm_addr to tcpm_daddr Christoph Paasch
2014-01-08 15:05 ` [PATCH net-next v2 2/5] tcp: metrics: Add source-address to tcp-metrics Christoph Paasch
2014-01-08 17:55 ` Eric Dumazet
2014-01-08 22:43 ` Christoph Paasch
2014-01-08 23:13 ` Eric Dumazet
2014-01-10 22:37 ` David Miller [this message]
2014-01-10 23:10 ` Hannes Frederic Sowa
2014-01-08 15:05 ` [PATCH net-next v2 3/5] tcp: metrics: New netlink attribute for src IP and dumped in netlink reply Christoph Paasch
2014-01-08 15:05 ` [PATCH net-next v2 4/5] tcp: metrics: Delete all entries matching a certain destination Christoph Paasch
2014-01-08 15:05 ` [PATCH net-next v2 5/5] tcp: metrics: Allow selective get/del of tcp-metrics based on src IP Christoph Paasch
2014-01-10 22:38 ` [PATCH net-next v2 0/5] Make tcp-metrics source-address aware David Miller
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20140110.173711.1541969326146414340.davem@davemloft.net \
--to=davem@davemloft.net \
--cc=christoph.paasch@uclouvain.be \
--cc=eric.dumazet@gmail.com \
--cc=ja@ssi.bg \
--cc=netdev@vger.kernel.org \
--cc=ycheng@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox