Re: Network virtualization/isolation

[Daniel Lezcano <dlezcano@fr.ibm.com>]

Dmitry Mishin wrote:
> On Monday 04 December 2006 19:43, Herbert Poetzl wrote:
>> On Mon, Dec 04, 2006 at 06:19:00PM +0300, Dmitry Mishin wrote:
>>> On Sunday 03 December 2006 19:00, Eric W. Biederman wrote:
>>>> Ok.  Just a quick summary of where I see the discussion.
>>>>
>>>> We all agree that L2 isolation is needed at some point.
>>> As we all agreed on this, may be it is time to send patches
>>> one-by-one? For the beggining, I propose to resend Cedric's
>>> empty namespace patch as base for others - it is really empty,
>>> but necessary in order to move further.
>>>
>>> After this patch and the following net namespace unshare
>>> patch will be accepted,
>> well, I have neither seen any performance tests showing
>> that the following is true:
>>
>>  - no change on network performance without the
>>    space enabled
>>  - no change on network performance on the host
>>    with the network namespaces enabled
>>  - no measureable overhead inside the network
>>    namespace
>>  - good scaleability for a larger number of network
>>    namespaces
> These questions are for complete L2 implementation, not for these 2 empty 
> patches. If you need some data relating to Andrey's implementation, I'll get 
> it. Which test do you accept?

tbench ?

With the following scenarii:

  * intra host communication (one time with IP on eth and one time with 
127.0.0.1)
  * inter host communication

Each time:
	- a single network namespace
	- with 100 network namespace. 1 server communicating and 99 listening 
but doing nothing.