Mailing List Archive: nsp: foundry

MLXe-4 vs ASR1006 vs MX80

Index | Next | Previous | View Threaded

drew.weaver at thenap

Apr 26, 2011, 5:23 AM

Post #1 of 16 (1796 views)
Permalink

MLXe-4 vs ASR1006 vs MX80

Has anyone seen any bake offs, or side by side comparisons between the MLXe-4, ASR1006 or the MX80?

I am trying to pick a router for a small regional DC and these seem attractive but I don't know all of the caveats and subtleties.

Any thoughts?

nick at foobar

Apr 26, 2011, 5:37 AM

Post #2 of 16 (1773 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

On 26/04/2011 13:23, Drew Weaver wrote:
> Has anyone seen any bake offs, or side by side comparisons between the
> MLXe-4, ASR1006 or the MX80?
>
> I am trying to pick a router for a small regional DC and these seem
> attractive but I don't know all of the caveats and subtleties.

MLXe-4: 400G backplane, 1m ipv4 prefixes, far faster than either the
ASR1006 or the MX80, but fewer features, particularly in terms of hqos

MX80: 60G backplane, 2m ipv4 prefixes, rich feature set

ASR1006: 5G output with ESP10, 4m ipv4 prefixes. Rich feature set.

Really, you need to decide what you need from a router and then choose the
box which provides what you need. These three boxes are completely
different systems with different strengths and weaknesses. If you need raw
muscle, the mlxe4 leaves the other boxes in the dust. If you want
cartloads of features, you'll probably need either the mx80 or the ASR1k.

Nick

_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

drew.weaver at thenap

Apr 26, 2011, 5:42 AM

Post #3 of 16 (1773 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

On 26/04/2011 13:23, Drew Weaver wrote:
> Has anyone seen any bake offs, or side by side comparisons between the
> MLXe-4, ASR1006 or the MX80?
>
> I am trying to pick a router for a small regional DC and these seem
> attractive but I don't know all of the caveats and subtleties.

MLXe-4: 400G backplane, 1m ipv4 prefixes, far faster than either the
ASR1006 or the MX80, but fewer features, particularly in terms of hqos

MX80: 60G backplane, 2m ipv4 prefixes, rich feature set

ASR1006: 5G output with ESP10, 4m ipv4 prefixes. Rich feature set.

Really, you need to decide what you need from a router and then choose the
box which provides what you need. These three boxes are completely
different systems with different strengths and weaknesses. If you need raw
muscle, the mlxe4 leaves the other boxes in the dust. If you want
cartloads of features, you'll probably need either the mx80 or the ASR1k.

Nick,

Currently on our edge/border routers we run Sampled Netflow, iACLs, uRPF, BGP IPv4/IPv6 (full feeds, several peers), OSPFv2/v3 and I think those are all of the features we use.

We just don't want to end up with a platform that isn't upgradable and has a bunch of crippling hardware limitations.

thanks,
-Drew

_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

jmadrid2 at gmail

Apr 26, 2011, 6:53 AM

Post #4 of 16 (1764 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

I currently use two MLXe-16 and can attest to the fact that you are
not going to get anywhere near 1m routes. More like 500K max. FYI.

On Tue, Apr 26, 2011 at 8:42 AM, Drew Weaver <drew.weaver [at] thenap> wrote:
> On 26/04/2011 13:23, Drew Weaver wrote:
>> Has anyone seen any bake offs, or side by side comparisons between the
>> MLXe-4, ASR1006 or the MX80?
>>
>> I am trying to pick a router for a small regional DC and these seem
>> attractive but I don't know all of the caveats and subtleties.
>
> MLXe-4: 400G backplane, 1m ipv4 prefixes, far faster than either the
> ASR1006 or the MX80, but fewer features, particularly in terms of hqos
>
> MX80: 60G backplane, 2m ipv4 prefixes, rich feature set
>
> ASR1006: 5G output with ESP10, 4m ipv4 prefixes. Rich feature set.
>
> Really, you need to decide what you need from a router and then choose the
> box which provides what you need. These three boxes are completely
> different systems with different strengths and weaknesses. If you need raw
> muscle, the mlxe4 leaves the other boxes in the dust. If you want
> cartloads of features, you'll probably need either the mx80 or the ASR1k.
>
>
> Nick,
>
> Currently on our edge/border routers we run Sampled Netflow, iACLs, uRPF, BGP IPv4/IPv6 (full feeds, several peers), OSPFv2/v3 and I think those are all of the features we use.
>
> We just don't want to end up with a platform that isn't upgradable and has a bunch of crippling hardware limitations.
>
> thanks,
> -Drew
>
>
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp [at] puck
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>

--
It has to start somewhere, it has to start sometime. What better
place than here? What better time than now?

_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

david.hooton at platformnetworks

Apr 26, 2011, 6:58 AM

Post #5 of 16 (1750 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

On 26/04/11 11:53 PM, "Jose Madrid" <jmadrid2 [at] gmail> wrote:

>I currently use two MLXe-16 and can attest to the fact that you are
>not going to get anywhere near 1m routes. More like 500K max. FYI.

What are peoples experiences with XMR's?

Dave

_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

nick at foobar

Apr 26, 2011, 6:59 AM

Post #6 of 16 (1773 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

On 26/04/2011 14:53, Jose Madrid wrote:
> I currently use two MLXe-16 and can attest to the fact that you are
> not going to get anywhere near 1m routes. More like 500K max. FYI.

the line card FIBs support 1m ipv4-sized slots. If you use ipv6 / mpls /
l3vpn / IGPs / L2 / etc, all entries in these will take up one or more
slots. So yes, 1m ipv4 prefixes will generally not be enough to run a DFZ
of 1m prefixes.

This is the usual disclaimer for statements of the form: "my FIB supports X
prefixes".

Nick

_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

cmarlatt at rxsec

Apr 26, 2011, 8:34 AM

Post #7 of 16 (1759 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

On 04/26/2011 08:37 AM, Nick Hilliard wrote:
> MLXe-4: 400G backplane, 1m ipv4 prefixes, far faster than either the
> ASR1006 or the MX80, but fewer features, particularly in terms of hqos

IIRC the MLX only does up to ~512k IPv4 routes. The XMR is the chassis
that'll support higher than this.

Regards,

Chris Marlatt
_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

nick at foobar

Apr 26, 2011, 9:00 AM

Post #8 of 16 (1759 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

On 26/04/2011 16:34, Chris Marlatt wrote:
> IIRC the MLX only does up to ~512k IPv4 routes. The XMR is the chassis
> that'll support higher than this.

sorry, you're correct here - the XMR does 1m prefixes. More correctly,
it's the line cards which support this, as this chassis uses distributed
forwarding.

In their data sheet for the MLX series, Brocade note:

> Industry-leading scalability up to:*
> • 10 million BGP routes
> • 1 million IPv4 routes in hardware (FIB)

the * refers to "Scalability limits depend on configured system parameters,
NetiIron XMR module types, licenses, system profile selected, and routing
database complexity."

This is stretching reality a little more than usual, or perhaps the
documentation writer just got the two mixed up.

Nick

_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

bdflemin at gmail

Apr 26, 2011, 9:49 AM

Post #9 of 16 (1769 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

On Apr 26, 2011, at 8:58 AM, David Hooton wrote:

> On 26/04/11 11:53 PM, "Jose Madrid" <jmadrid2 [at] gmail> wrote:
>
>
>> I currently use two MLXe-16 and can attest to the fact that you are
>> not going to get anywhere near 1m routes. More like 500K max. FYI.
>
> What are peoples experiences with XMR's?

We operate a network with 15 XMR 4000's and have been very happy. We
pull transit feeds from 4 upstreams in geographically diverse
locations. We run OSPF internally and support IPv4 and IPv6 unicast
and multicast. We do NOT do much with L3 VPNS. Brocade has been a good
partner and helped us with special pricing on some larger projects.

Three "gripes":
1) We have all 1st generation linecards and can't get VLAN stats. Easy
way to fix it (buy new card) but we can't afford the new linecards.
2) I want to point a VLL to a non-routerID IP!
3) Regex for things like as-path filtering is poor compared to Cisco
and Juniper.
General) Documentation is FAR behind Cisco and Juniper, IMHO.

The price per port + energy consumption make the XMR a slam dunk for
our all-Ethernet network.
_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

mmoriniaux at prosodie

Apr 26, 2011, 9:50 AM

Post #10 of 16 (1764 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

Hello All,
Last pres we got the MLXe does 1m in FIB, our current MLX only do 512k
(IIRC you can put regular MLX cards in an MLXe but then you only get the
512k FIB)

If you are lokking for something in the MX80/ASR1k class why not look at
the CER?
Dirt cheap, same feature set as MLX but only 1U. The CER coming out this
summer support 1.5m routes in FIB.

Best Regards,
Michel Moriniaux

> -----Original Message-----
> From: foundry-nsp-bounces [at] puck
> [mailto:foundry-nsp-bounces [at] puck] On Behalf Of
> Chris Marlatt
> Sent: Tuesday, April 26, 2011 5:34 PM
> To: Nick Hilliard
> Cc: foundry-nsp [at] puck
> Subject: Re: [f-nsp] MLXe-4 vs ASR1006 vs MX80
>
> On 04/26/2011 08:37 AM, Nick Hilliard wrote:
> > MLXe-4: 400G backplane, 1m ipv4 prefixes, far faster than either the
> > ASR1006 or the MX80, but fewer features, particularly in
> terms of hqos
>
> IIRC the MLX only does up to ~512k IPv4 routes. The XMR is
> the chassis that'll support higher than this.
>
> Regards,
>
> Chris Marlatt
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp [at] puck
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>

_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

jmadrid2 at gmail

Apr 26, 2011, 10:01 AM

Post #11 of 16 (1770 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

Brad,

I have similar gripes with the Brocade documentation. The boxes themselves are workhorses and Cisco should pray to handle BGP as well as Brocade does. That being said, their documentation leaves A LOT to be desired. Thus far, we have run into some issues with our MLX's and to be honest, I have prayed that they would somehow magically become MX80's overnight, but as of now they haven't :)

Sent via BlackBerry by AT&T

-----Original Message-----
From: Brad Fleming <bdflemin [at] gmail>
Sender: foundry-nsp-bounces [at] puck
Date: Tue, 26 Apr 2011 11:49:01
To: <foundry-nsp [at] puck>
Subject: Re: [f-nsp] MLXe-4 vs ASR1006 vs MX80

On Apr 26, 2011, at 8:58 AM, David Hooton wrote:

> On 26/04/11 11:53 PM, "Jose Madrid" <jmadrid2 [at] gmail> wrote:
>
>
>> I currently use two MLXe-16 and can attest to the fact that you are
>> not going to get anywhere near 1m routes. More like 500K max. FYI.
>
> What are peoples experiences with XMR's?

We operate a network with 15 XMR 4000's and have been very happy. We
pull transit feeds from 4 upstreams in geographically diverse
locations. We run OSPF internally and support IPv4 and IPv6 unicast
and multicast. We do NOT do much with L3 VPNS. Brocade has been a good
partner and helped us with special pricing on some larger projects.

Three "gripes":
1) We have all 1st generation linecards and can't get VLAN stats. Easy
way to fix it (buy new card) but we can't afford the new linecards.
2) I want to point a VLL to a non-routerID IP!
3) Regex for things like as-path filtering is poor compared to Cisco
and Juniper.
General) Documentation is FAR behind Cisco and Juniper, IMHO.

The price per port + energy consumption make the XMR a slam dunk for
our all-Ethernet network.
_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

frnkblk at iname

Apr 26, 2011, 11:52 AM

Post #12 of 16 (1750 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

CER only has two 10GB ports. =(

Frank

-----Original Message-----
From: foundry-nsp-bounces [at] puck
[mailto:foundry-nsp-bounces [at] puck] On Behalf Of Moriniaux Michel
Sent: Tuesday, April 26, 2011 11:50 AM
To: Chris Marlatt; Nick Hilliard
Cc: foundry-nsp [at] puck
Subject: Re: [f-nsp] MLXe-4 vs ASR1006 vs MX80

Hello All,
Last pres we got the MLXe does 1m in FIB, our current MLX only do 512k
(IIRC you can put regular MLX cards in an MLXe but then you only get the
512k FIB)

If you are lokking for something in the MX80/ASR1k class why not look at
the CER?
Dirt cheap, same feature set as MLX but only 1U. The CER coming out this
summer support 1.5m routes in FIB.

Best Regards,
Michel Moriniaux

> -----Original Message-----
> From: foundry-nsp-bounces [at] puck
> [mailto:foundry-nsp-bounces [at] puck] On Behalf Of
> Chris Marlatt
> Sent: Tuesday, April 26, 2011 5:34 PM
> To: Nick Hilliard
> Cc: foundry-nsp [at] puck
> Subject: Re: [f-nsp] MLXe-4 vs ASR1006 vs MX80
>
> On 04/26/2011 08:37 AM, Nick Hilliard wrote:
> > MLXe-4: 400G backplane, 1m ipv4 prefixes, far faster than either the
> > ASR1006 or the MX80, but fewer features, particularly in
> terms of hqos
>
> IIRC the MLX only does up to ~512k IPv4 routes. The XMR is
> the chassis that'll support higher than this.
>
> Regards,
>
> Chris Marlatt
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp [at] puck
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>

_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

cmarlatt at rxsec

Apr 26, 2011, 12:45 PM

Post #13 of 16 (1754 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

On 04/26/2011 12:50 PM, Moriniaux Michel wrote:
> Hello All,
> Last pres we got the MLXe does 1m in FIB, our current MLX only do 512k
> (IIRC you can put regular MLX cards in an MLXe but then you only get the
> 512k FIB)

I have seen similar. However on my MLXe's the MP only wants to let me
configure a maximum of 524288 for the ip-route in system-max - even
though my LPs show twice the available memory as the "gen-1" cards.

I suppose it could be a cam profile issue but I wont have a need to
support that many routes for these devices so I haven't determined where
the discrepancy is yet.

Regards,

Chris Marlatt
_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

georgeb at gmail

Apr 26, 2011, 6:32 PM

Post #14 of 16 (1747 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

On Tue, Apr 26, 2011 at 12:45 PM, Chris Marlatt <cmarlatt [at] rxsec> wrote:

> On 04/26/2011 12:50 PM, Moriniaux Michel wrote:
>
>> Hello All,
>> Last pres we got the MLXe does 1m in FIB, our current MLX only do 512k
>> (IIRC you can put regular MLX cards in an MLXe but then you only get the
>> 512k FIB)
>>
>
> I have seen similar. However on my MLXe's the MP only wants to let me
> configure a maximum of 524288 for the ip-route in system-max - even though
> my LPs show twice the available memory as the "gen-1" cards.
>
> I suppose it could be a cam profile issue but I wont have a need to support
> that many routes for these devices so I haven't determined where the
> discrepancy is yet.
>
>
> Regards,
>
> Chris Marlatt
>

MLXe uses XMR cards, that is why it gets 1 million routes. The MLXe is
basically a converged MLX/XMR

james.braunegg at micron21

Apr 28, 2011, 6:25 AM

Post #15 of 16 (1786 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

Dear All

Micron21 completed a three month in depth internal investigation between Brocade MLXe XMR the Cisco ASR1004 and the Juniper MX80 around 8 months ago, evaluating which product to invest in. If you wish to know specific details of our study please feel free to contact me directly.

In summary please find the below comments

To give you some background Micron21 is an Australian based Datacentre, we provide datacentre services at an extremely high density, our network is mission critical and supports some of the largest Australian companies, in saying that we are very proud in our decision choosing the Brocade MLXe XMR platform as our routing platform.

Brocade was the right choice for us and since our migration to the Brocade MLXe XMR platform we have also developed an amazing relationship between our company and many senior people within almost every department at Brocade.

In fact owning a Brocade solution makes us feel extremely proud that we choose Brocade, you can't help but smile even more when you see the Brocade One "rap" video knowing your part of an evolving revolution.

To be honest and unbiased Brocade hands down for our application within the datacentre environment was a much better product when you compare the raw processing power to almost anything on the market. Whilst Brocade might not have the "Marketing perception" or glossy paper brochures of the ASR or MX80 it truly is a much better product when you look the technical features you will actually use for a routing platform within a high density environment.

Kindest Regards

James Braunegg
W: 1300 769 972 | M: 0488 997 207 | D: (03) 9751 7616
E: james.braunegg [at] micron21<mailto:james.braunegg [at] micron21> | ABN: 12 109 977 666

[Description: Description: Description: M21.jpg]

This message is intended for the addressee named above. It may contain privileged or confidential information. If you are not the intended recipient of this message you must not use, copy, distribute or disclose it to anyone other than the addressee. If you have received this message in error please return the message to the sender by replying to it and then delete the message from your computer.

From: foundry-nsp-bounces [at] puck [mailto:foundry-nsp-bounces [at] puck] On Behalf Of Drew Weaver
Sent: Tuesday, April 26, 2011 10:23 PM
To: foundry-nsp [at] puck
Subject: [f-nsp] MLXe-4 vs ASR1006 vs MX80

Has anyone seen any bake offs, or side by side comparisons between the MLXe-4, ASR1006 or the MX80?

I am trying to pick a router for a small regional DC and these seem attractive but I don't know all of the caveats and subtleties.

Any thoughts?

Attachments:

image001.jpg (2.62 KB)

nick at foobar

Apr 29, 2011, 5:09 AM

Post #16 of 16 (1710 views)
Permalink

Re: MLXe-4 vs ASR1006 vs MX80 [In reply to]

On 28/04/2011 14:25, James Braunegg wrote:
> In fact owning a Brocade solution makes us feel extremely proud that we
> choose Brocade, you can’t help but smile even more when you see the Brocade
> One “rap” video knowing your part of an evolving revolution.

http://www.youtube.com/watch?v=Kdc-dcaM50s

What has been seen cannot be unseen.

Nick
_______________________________________________
foundry-nsp mailing list
foundry-nsp [at] puck
http://puck.nether.net/mailman/listinfo/foundry-nsp

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact Gossamer Threads