[ovs-dev] [PATCH] ipfix: implement flow caching and aggregation in exporter

[Romain Lenglet]

On Oct 1, 2013, at 11:40 AM, Neil Mckee <neil.mckee at inmon.com> wrote:

> Just a comment:  one of the architectural differences between IPFIX and sFlow is the location of the flow-cache:   with sFlow the cache is implemented in external software on a central server,  which can then easily export IPFIX if required (while also providing real-time network-wide visibility into flows-in-progress,  which is relevant for SDN use cases).
> 
> http://blog.sflow.com/2013/08/restflow.html
> 
> Neil

In OVS, the IPFIX flow cache is optional in the exporter. It can be disabled. So one can get either sFlow-like caching behavior, or usual IPFIX caching in the exporter.
Thanks for your comment.
--
Romain Lenglet

> 
> 
> On Aug 21, 2013, at 1:50 PM, Ben Pfaff <blp at nicira.com> wrote:
> 
>> On Wed, Aug 21, 2013 at 01:47:45PM -0700, Romain Lenglet wrote:
>>> ----- Original Message -----
>>>> From: "Ben Pfaff" <blp at nicira.com>
>>>> To: "Romain Lenglet" <rlenglet at vmware.com>
>>>> Cc: dev at openvswitch.org
>>>> Sent: Wednesday, August 21, 2013 1:25:59 PM
>>>> Subject: Re: [PATCH] ipfix: implement flow caching and aggregation in exporter
>>>> 
>>>> On Tue, Aug 20, 2013 at 05:54:27PM -0700, Romain Lenglet wrote:
>>>>> Implement a per-exporter flow cache with active timeout expiration.
>>>>> Add columns "cache_active_timeout" and "cache_max_flows" into table
>>>>> "IPFIX" to configure each cache.
>>>>> 
>>>>> Add per-flow elements "octetDeltaSumOfSquares",
>>>>> "minimumIpTotalLength", and "maximumIpTotalLength" to replace
>>>>> "ethernetTotalLength".  Add per-flow element "flowEndReason" to
>>>>> indicate whether a flow has expired because of an active timeout, the
>>>>> cache size limit being reached, or the exporter being stopped.
>>>>> 
>>>>> Signed-off-by: Romain Lenglet <rlenglet at vmware.com>
>>>> 
>>>> I'm happy with the code.  I don't exactly understand what this
>>>> provides to users, so can you provide a brief item to add to NEWS that
>>>> explains it?  (Is the benefit directly user-visible?)
>>> 
>>> This basically makes IPFIX usable.  Without caching it generates one
>>> message for every sampled packet, which can be a lot of traffic.
>>> This can be useful as it makes it behave like sFlow, but it is not what
>>> users usually expect from an IPFIX exporter.
>>> 
>>> I added this item into NEWS:
>>>   - Added configurable flow caching support to IPFIX exporter.
>> 
>> Thanks, that should help our users (and does help me).
>> _______________________________________________
>> dev mailing list
>> dev at openvswitch.org
>> http://openvswitch.org/mailman/listinfo/dev