Cisco 6509 replacement?

IrishMLK

Weaksauce
Joined
Apr 4, 2009
Messages
101
Our 6509 has seen better days and we know it's past time to replace it... Question is - go with similar or just use multiple 48 port switches? Then we have old Cisco 2800s in our 9 IDFs that would need updated as well.

We have a mixed environment with about 300 active clients along with about 15 servers and an EMC AX4 over 4GB fibre brocades for backup.

Any suggestions with bang for buck in mind?
 
You haven't given us much info. This would be your core switch?

For IDFs, I'd go with 3750 stacks. For the servers, a 4948. or 4900M. What's up with the 2800s in the IDFs? What are they doing?

What's your budget?
 
The Cisco Catalyst 6500 is a very versatile beast; if it isn't performing up to your requirements, often times all you have to do is replace a line card, or the supervisor engines, to make it much faster. Can you post a 'show mod' from your 6509? It'll give us details on what sort of configuration the hardware is in.
 
Problem is, the original 6509 chassis is going out of support next year. Just depends on your requirements as to what you need. If you aren't pushing that 6500 you can go down to a redundant 4500. Or a new 6509-E. Also depends if 10Gb is in your future, and how much.
 
Yes, the 6509 is our core switch. The IDFs are fed by GbE fibre off the 6509 and they service the majority of our clients. All of our office clients - approx 100 - go to the 6509. The upcoming EOL is a determining factor.

Thanks for the help! Any non-Cisco suggestions?
 
Why do you have a 6509 switch feeding 2800 routers...which, I assume, feed access switches?

This design seems odd to me. I know the 6509 is fully capable of routing, but your design should have a true router at the core or the 2800's should actually be 3750's or similar layer three switch.

Without knowing what is running off of the IDF's, I would say replace the 2800's with 3750's or equivalent. What was the reason for the 6509? Just port density? What is your bandwidth requirements? You could stack multiple switches but it will probably be smarter to go with another single chassis unit. I doubt you need 9 blades, from your description so far.
 
There is a huge secondary market for 6500 series chassis and components. If you are happy with current feature set and don't have pressures in cooling and power, just keep the unit and buy some spares.

Network Hardware Resale and Prism Innovations offer support for EOL products at very competitive rates.
 
I would just up the 6500s, the last large company I worked with was using them, and they had no issues, I think they were going 10gbit on them.

Either way I feel it would be easier to stock a few extra parts and keep what u got. The 6500's are strong, but you should Always always, with a modular switch, have extra cards and power supplies on hand for an emergency.
 
Sorry the IDFs are 2900 switches. :eek:

Ok, that makes much more sense. In that case, why change anything? Just because you won't be able to get the newest IOS for it isn't a strong enough reason to upgrade it. As has been said, the secondary market is strong. You can get the parts for cheap. Hell, set up a second one and build in redundant paths.
 
Maybe I'm biased, but I disagree with keeping old/EOL hardware and buying gray parts for it. You can't legally download software updates which contain bug fixes and security patches, not to mention new features. You also have no one to go to for support (unless you're using a third party).
 
Maybe I'm biased, but I disagree with keeping old/EOL hardware and buying gray parts for it. You can't legally download software updates which contain bug fixes and security patches, not to mention new features. You also have no one to go to for support (unless you're using a third party).

Agreed. For access switches? Do whatever. I used to keep two on the shelf instead of a fast response SmartNet. For my core...it needs to be under 7x24x4 in most cases. Not grey market.

If you don't need the performance of a 6500 the newer 4500 switches are very capable and really reasonable.
 
It's hard to beat the flexibility of the 6500 series. Just upgrade line cards or sup engines if needed. Staying with a given vendor (Cisco) means you'll have more spare parts laying about in case of problems. In the 6500 league though it's usually important enough to have the $$$ 4 hour guaranteed response support contract on.
 
Well, when you have a 6500 as your core yet you post here for advice on replacing it, I suspect there isn't a support contract in place anyway. If there is a support contract in place then that would be the first place I would ask about upgrading. If I am correct and there isn't a support contract in place then there is nothing wrong with relying on the secondary parts market until the company wants to pony up.

Hell, I support the DoD and most of the equipment that I use is not supported by a Cisco contract. Every part I have had to order for replacement has come from the cheapest stores possible and most are not official Cisco distributors. The only thing that we use a contact for is obtaining IOS's and they are several versions behind because they are scrutinized before being authorized for use.

I can guarantee there are plenty of companies that are much further behind, mainly due to the cost of upgrading hardware. Most companies cannot afford to keep the newest hardware in place and do not see the benefit to constantly upgrading hardware, especially if the current hardware is meeting there needs and spare parts are easily, and cheaply, obtained.

Despite what us networking geeks would like, we don't control the purse strings.
 
I know of a company who replaced their 6509 with an HP E8212 and is happy with it. Lifetime warranty. Some local universities did too.

HP has a "Catalyst for Change" promo that gets an additional 20% off going on through the end of July too.
 
Thanks for all of the insight. I will get the config and talk to our suppliers about upgrading. After talking to my boss, if we can get another couple years on the 6509 by upgrading some things, we will concentrate on upgrading the IDFs now to offer more scalability when the core gets replaced.

I must say the HP stuff looks appealing. Especially if we want to make the transition to VOIP soon after the 6509 gets decommed. Price per port on POE with HP is easier to push past the bean counters... :D
 
The HP stuff is pretty nice. HP recently bought 3com and H3C. I have a customer with about 6000 PCs, they replaced a large chassis switch, similar to the 6509, with HP A5800 1 U switches. The 5800s have something called IRF, which creates a virtual switch. All your routing, switching, trunking etc... is configured as if it's one box, but if one of the 5800s dies, the others take over. it works pretty nicely, and comes with lifetime warranty and support. Additionally, you'll save a bunch of rackspace, power and cooling.

A number of years ago, pre-H3C purchase, we did an eval for a school district between HP, Extreme and Cisco. In the short term, the HP procurve solution was a little less expensive than cisco. But over time, given support costs, HP was far superior. You just can't beat their lifetime warranty.
 
You're comparing a stack of fixed (HP) switches to a 6500? Really?

An EOL cisco 6509 is going to be what, 10 years old? Even a relatively new SUP 720 for the 6500 is 400Mpps. You can get a 1U A5820 switch with 24 10GigE ports that'll do 360Mpps, and a lifetime next day warranty.

Again, that's a 1U box that has enough ports to run 20Gbps to each of his wiring closets. I suspect that's more than enough for 300 clients.
 
Just by my question you can tell I am new to network admin. The fabric, I have learned, on the 6509 is the bottleneck. The stacking ability is what made me interested in the HP stuff, not to mention the price-point for our relatively small company.

I don't foresee the need for 10GigE for anything other than the backbone to our EMC datastores in the future, but having the scalability would be nice as well.

Thanks again for everyone's input!! :cool:
 
Just by my question you can tell I am new to network admin. The fabric, I have learned, on the 6509 is the bottleneck. The stacking ability is what made me interested in the HP stuff, not to mention the price-point for our relatively small company.

I don't foresee the need for 10GigE for anything other than the backbone to our EMC datastores in the future, but having the scalability would be nice as well.

Thanks again for everyone's input!! :cool:

The backplane is the bottleneck?! What do you think the bottleneck in a stack is? You the you'll find a stack with a 720gbps-2tbps backplane?

Are you using a Sup1, Sup2 (without SFM), or Sup32 in your 6500? If so, then yea, the backplane isn't the fastest. Though, with HP's IRF, your "backplane" is the amount of ports you use for peer links.
 
Back
Top