Mailing List Archive

On Thu, 2014-01-16 at 13:43 -0800, Carlos Alfaro wrote:
> Would be grateful if anyone can confirm whether using BR-MLX-10Gx8-X in an XMR with HSF, will give me line rate?

I could be wrong, but I don't think it'll work at all. That's an MLX
card; you can't use it with XMR management modules.

--
Richard

I believe the XMR will support MLX modules... not the other way around.
However, with the XMR and MLX replaced by the MLXe, I'm not sure if the
former platforms support the latest high-density interface modules.


On Thu, Jan 16, 2014 at 3:39 PM, Richard Laager <rlaager@wiktel.com> wrote:

> On Thu, 2014-01-16 at 13:43 -0800, Carlos Alfaro wrote:
> > Would be grateful if anyone can confirm whether using BR-MLX-10Gx8-X in
> an XMR with HSF, will give me line rate?
>
> I could be wrong, but I don't think it'll work at all. That's an MLX
> card; you can't use it with XMR management modules.
>
> --
> Richard
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>

The BR-MLX-10Gx8-X blade is an X blade.

-X will work in both the XMR and MLX or MLXe with XMR or MLX type management blades.
-X-ML will only work in the MLX or MLXe with MLX type management blades.

See also the module compatibility matrix.

http://www.brocade.com/downloads/documents/technical-notes/mlx-module-compatibility-matrix-tn.pdf

And yes it should give you line rate.
As far as I know with hSFM you will have a max bandwidth for the slot of 160Gbps.

Greetings,

Diederik


Sent from my iPad

> On 17 jan. 2014, at 01:06, Matt Kassawara <mkassawara@gmail.com> wrote:
>
> I believe the XMR will support MLX modules... not the other way around. However, with the XMR and MLX replaced by the MLXe, I'm not sure if the former platforms support the latest high-density interface modules.
>
>
>> On Thu, Jan 16, 2014 at 3:39 PM, Richard Laager <rlaager@wiktel.com> wrote:
>> On Thu, 2014-01-16 at 13:43 -0800, Carlos Alfaro wrote:
>> > Would be grateful if anyone can confirm whether using BR-MLX-10Gx8-X in an XMR with HSF, will give me line rate?
>>
>> I could be wrong, but I don't think it'll work at all. That's an MLX
>> card; you can't use it with XMR management modules.
>>
>> --
>> Richard
>>
>> _______________________________________________
>> foundry-nsp mailing list
>> foundry-nsp@puck.nether.net
>> http://puck.nether.net/mailman/listinfo/foundry-nsp
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp

Hello all,

To clear up the confusion, first off the NetIron platform chassis options:

MLX: can only handle -M modules.
XMR: can only handle -X modules, aka the full BGP routing table.
MLXe: is the successor to the MLX/XMR platform, and becomes MLX-like or
XMR-like depending on the management blades you install; management
blade type is distinguished by the -M or -X suffix in the SKU.

The main difference between MLXe and its predecessors is that the
backplane can handle high capacity cards such as:

BR-MLX-10GX24-DM
BR-MLX-100Gx2-X
BR-MLX-40Gx4-M

Now, you can still install a BR-MLX-10Gx8-X into a legacy XMR chassis
but HSF switching fabrics are required, the old SFM are not good enough.
That being said the ports on 10Gx8 cards are as far as I know all
wirespeed: I know for sure when installed in a MLXe chassis, but not
100% certain about full wirespeed when they are installed in a legacy
XMR chassis.

What I can tell you is that we have the exact setup Carlos describes in
production (BR-MLX-10Gx8-X blades in a legacy XMR chassis with HSF
fabrics), and it's running without problems: but we've never fully
maximized port bandwidth usage on all ports at the same time. Don't
think such a situation occurs too often in the wild, either, since
almost everyone will upgrade their capacity to spread load over more
ports rather than risking congestion.

Best regards,

Martijn Schmidt
i3D.net

On 01/17/2014 01:06 AM, Matt Kassawara wrote:
> I believe the XMR will support MLX modules... not the other way
> around. However, with the XMR and MLX replaced by the MLXe, I'm not
> sure if the former platforms support the latest high-density interface
> modules.
>
>
> On Thu, Jan 16, 2014 at 3:39 PM, Richard Laager <rlaager@wiktel.com
> <mailto:rlaager@wiktel.com>> wrote:
>
> On Thu, 2014-01-16 at 13:43 -0800, Carlos Alfaro wrote:
> > Would be grateful if anyone can confirm whether using
> BR-MLX-10Gx8-X in an XMR with HSF, will give me line rate?
>
> I could be wrong, but I don't think it'll work at all. That's an MLX
> card; you can't use it with XMR management modules.
>
> --
> Richard
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net <mailto:foundry-nsp@puck.nether.net>
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>
>
>
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp

Carlos, as the previous respondents mentioned, this will indeed work.
Additionally, with the HSF (even on the non -E chassis), this will be full
line rate for the 8x10. One other point, in the past you could run -X
modules in a MLX, but not -M modules in an XMR chassis. However, with the
5.6 code release, Brocade has opened this up and allowed more of a
universal model. However, if you do install a -M module in an XMR(or X)
chassis all modules will basically drop down to -M scalability (512k v4
routes in FIB).

Mike


On Thu, Jan 16, 2014 at 5:03 PM, i3D.net - Martijn Schmidt <
martijnschmidt@i3d.net> wrote:

> Hello all,
>
> To clear up the confusion, first off the NetIron platform chassis options:
>
> MLX: can only handle -M modules.
> XMR: can only handle -X modules, aka the full BGP routing table.
> MLXe: is the successor to the MLX/XMR platform, and becomes MLX-like or
> XMR-like depending on the management blades you install; management blade
> type is distinguished by the -M or -X suffix in the SKU.
>
> The main difference between MLXe and its predecessors is that the
> backplane can handle high capacity cards such as:
>
> BR-MLX-10GX24-DM
> BR-MLX-100Gx2-X
> BR-MLX-40Gx4-M
>
> Now, you can still install a BR-MLX-10Gx8-X into a legacy XMR chassis but
> HSF switching fabrics are required, the old SFM are not good enough. That
> being said the ports on 10Gx8 cards are as far as I know all wirespeed: I
> know for sure when installed in a MLXe chassis, but not 100% certain about
> full wirespeed when they are installed in a legacy XMR chassis.
>
> What I can tell you is that we have the exact setup Carlos describes in
> production (BR-MLX-10Gx8-X blades in a legacy XMR chassis with HSF
> fabrics), and it's running without problems: but we've never fully
> maximized port bandwidth usage on all ports at the same time. Don't think
> such a situation occurs too often in the wild, either, since almost
> everyone will upgrade their capacity to spread load over more ports rather
> than risking congestion.
>
> Best regards,
>
> Martijn Schmidt
> i3D.net
>
>
> On 01/17/2014 01:06 AM, Matt Kassawara wrote:
>
> I believe the XMR will support MLX modules... not the other way around.
> However, with the XMR and MLX replaced by the MLXe, I'm not sure if the
> former platforms support the latest high-density interface modules.
>
>
> On Thu, Jan 16, 2014 at 3:39 PM, Richard Laager <rlaager@wiktel.com>wrote:
>
>> On Thu, 2014-01-16 at 13:43 -0800, Carlos Alfaro wrote:
>> > Would be grateful if anyone can confirm whether using BR-MLX-10Gx8-X in
>> an XMR with HSF, will give me line rate?
>>
>> I could be wrong, but I don't think it'll work at all. That's an MLX
>> card; you can't use it with XMR management modules.
>>
>> --
>> Richard
>>
>> _______________________________________________
>> foundry-nsp mailing list
>> foundry-nsp@puck.nether.net
>> http://puck.nether.net/mailman/listinfo/foundry-nsp
>>
>
>
>
> _______________________________________________
> foundry-nsp mailing listfoundry-nsp@puck.nether.nethttp://puck.nether.net/mailman/listinfo/foundry-nsp
>
>
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>