The future of the Catalyst 6500 Series

[Kerry Kriegel]

In my many years as a network engineer I have experienced almost all of the trials and tribulations from the explosion of the internet.  The increase of bandwidth, transport technology, transport media, number of routes, even the transition from IPv4 to IPv6 - over the last decades has been truly amazing.  It has also, at times, been truly painful.

I am speaking specifically of the EOS and EOL lifecycle of equipment that necessitates the occasional “fork-lift” replacement of backbone hardware.

Several years ago we went through that process and installed the Cisco flagship Catalyst 6500 series at our edge and core layers in multiple data centers.  With the promise of “investment protection for decades to come”, we thought we were on our way.  At the very least, I thought they would still be here until after I retired.  Over the years we have needed to upgrade the Sup modules to the present SUP-720-3BXL, fiber modules up to 10G, etc., etc.  These upgrades were on a per slot, as needed basis.  They were also exactly what we were expecting when we put the chassis in place.

Then, a few months back, when the v4 route table went over 512K, we made a painful discovery – TCAM.

The adjustment to hold more v4 routes, at the cost of holding fewer v6 routes, was put into place, and all we had to do was reboot each chassis.  Not too bad.  But wait, the limit is 1 million routes, total – period.

Google and I have been searching for the announcement from Cisco that the next generation supervisor cards will soon be released and all of the TCAM issues will be resolved, thus protecting my investment for at least one decade.  The most interesting discovery was references to a 2009 video by John McCool, Cisco Senior Vice President/General Manager for the Data Center, Switching and Services Group.

http://blogs.cisco.com/ar/cisco_svp_john_mccool_on_future_of_catalyst_6500

The video is no longer available.  :-/

It is not my intention to start a discussion about the different methods of getting around the limits that are now on the horizon for the 6500.

I would rather have a discussion about its future.  Does it have one, or is it being abandoned by Cisco?  Are there any “next generation” plans for the 6500?

Am I in for another fork-lift upgrade?  Like it or not?

[Jon Marshall]

Kerry

I don't work for Cisco so these are just my opinions.

Firstly the 6500 has lasted a remarkably long time already when compared to other switch models mainly due to it's installed user base. To expect any specific platform to still be there until you retire (obviously depending how old you are :-)), is somewhat optimistic to say the least.

The fact that Cisco have managed to adapt the 6500 from a device that originally supported a shared 32Gbps bus between all modules to one supporting up to 80Gbps per slot with the sup2T and maintained a certain amount of backwards compatibility is quite an achievement.

In terms of having to do a complete replacement of all your kit it depends on what you mean.

Undoubtedly the 6800 is seen as the successor to the 6500 and you will see customers gradually upgrading to that chassis. However that does not mean you need to throw away everything you have now. 

The sup720 provided pretty much full backwards compatibility in terms of modules while delivering more bandwidth, up to 40Gbps per slot in a 6509.  The sup2T provided less compatibility in terms of modules ie. not all were supported but some were and again increased the bandwidth capacity up to 80Gbps per slot in a 6500 chassis.

With the 6800 modular switch the same sup2T now provides up to 220Gbps per slot although obviously how much of that bandwidth can be used is still dependant on the module itself.

So it can be a gradual migration and if you already have sup2Ts then these can be reused as can some, but not all, of your modules in the new chassis.

I suspect this is what Cisco will do ie. I don't think the 6500 is going away any time soon because that will give customers time to migrate but the 6500 architecture ie.the chassis itself is now the limiting factor so it will eventually have to replaced.

Not sure I have totally answered your question but just my thoughts on it.

Jon

[Joseph W. Doherty]

Disclaimer

The Author of this posting offers the information contained within this posting without consideration and with the reader's understanding that there's no implied or expressed suitability or fitness for any purpose. Information provided is for informational purposes only and should not be construed as rendering professional advice of any kind. Usage of this posting's information is solely at reader's own risk.

Liability Disclaimer

In no event shall Author be liable for any damages whatsoever (including, without limitation, damages for loss of use, data or profit) arising out of the use or inability to use the posting's information even if Author has been advised of the possibility of such damage.

Posting

I'm pretty much in agreement with Jon.

I suspect the 6500 series is on the way out, to be succeeded by the 6800, just as the 6500 replaced the 6000 series; or the 6500E replaced the 6500.

I also want to mention, that at some point, improved architecture requires some serious hardware upgrades.

For example, the 6000 chassis doesn't support fabric, i.e. the SFM, SFM2 or sup720.

To use a SFM or SFM2 you needed a sup2.

To fully leverage 6500 fabric, you want all cards to have a fabric interface.

To use a sup2T, you need a 6500E (which also supports more power within the chassis for PoE line cards).

So, it's certainly nice to be able to perform some incremental hardware upgrades along the way, but major hardware improvements often do require fork-lift, or close to fork-lift, upgrades.

BTW, don't believe any vendor's promises.  The only guarantees you can really count on are death and taxes.  ;)

[Leo Laohoo]

I agree with Jon and Joseph.  

6500E is on it's final leg.  The only question left is WHEN is the EoS going to be announced.  Timing is everything.  If it's up to Cisco it's last year.  

The "most" bandwidth/slot you can get with the 6500E is 80 Gbps and Cisco can't squeeze anything more than that.  The most-logical upgrade path for the 6500E is the 6807-X (and NOT the 6880-X).  

If you can wait until next year, I recommend you do so.