Is redundancy in a single core switch practical for a small campus?

[DANIEL WANG]

We have a small campus with a main building and a few prefabbed buildings, a microwave building that connects our various WAN sites, and redundant Internet connections. In the past we used two 4006 w/sup3, L2 port channel in between, and HSRP as our collapsed core, and more than a dozen of 3550 & 3524XL & 2950. It worked great for a few years. Due to network growth and 3-4 network outages in recent years that happened roughly once every 6-7 months caused by misbehaving sup3, accidentally introduced switching loops/misconfigured STP, etc. We decided to shut down one of the core switches and use it as a manual standby in case of core failures.

Now we are planning to move to new campus with similar layout as the old one, supporting about 250 users and 40 servers (VMs and physical). We wonder if we should go back to the original collapsed core model with two cores, or, just building a single core with enough redundancy built-in. Being a SMB, we are extremely cost-conscious. That was why we never did any upgrade for 10 years. Now that there is a chance, we'd like to do it as cost-effective as possible. Performance is not a big concern as we never really have problem with our current 100mbps to desktops and 1gbps for uplinks and ether-teamed Gbps servers setup. We do, however, plan to do 1Gb to desktops and  probably10Gb between access and distribution/core. Here are my questions:

Can we save money by using one 6500 chasis with redundant sup engines and possibly redundant line cards on important switch and server connections while still achieving virtually the same redudancy at the core as the dual core model? How practically reliable is one chasis? Information I got from various vendor engineers is that no one has ever seen a failed chasis.

If not, can we do redundant chasis (or backplanes) in one physical box? Is the cost then equivalent to the dual core setup?

Please advise at your earliest convenince. Any feedback will be greatly appreciated. Thank you so much!

[Joseph W. Doherty · Accepted solution]

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

A failed chassis is very rare.  I've never seen it happen, but one of my peers has seen it once.  In any case, there are some advantages to having separate cores, but one "box" redundancy might be sufficient for your needs.

Since you mention cost, besides the 6500, in-chassis redundancy can be had in the 4507R or 4510R if you don't need 6500 features.  Cisco also has a new 4500X switch that is to support VSS.  However perhaps the least expensive option might be a stacked 3750X.

[Leo Laohoo · Accepted solution]

I'd go with Joseph on this.  If cost is a major influence then going down with a "short" stacked 3750X makes sense.  Take note I said "short stack" because we've done financial comparison that if you stack up to three 3750X together it's cheaper than the 4510R+E with a Sup7E.  If you go more than four and up to six 3750X then it's cheaper to down the 4510R+E with a Sup7E.  I will not recommend anyone stacking a 3750/3750G/3750E/3750X with six or more in a stack.

[Joseph W. Doherty · Accepted solution]

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

Leo's makes an important point.  Larger chassis can be less expensive than similar number of stacked 3750s.  Actual break even point depends on kind of chassis and whether chassis has redundant supervisors.

Another point I didn't touch on, since OP noted performance isn't a big concern, the larger 3750 stacks don't perform as well as chassis with a true fabric.

What I had in mind was just a stack of two 3750Xs for redundancy, with dual MEC Etherchannel links to down stream devices; such as L2 switches (which also might be stacked, more 3750Xs [L2 license] or 2960Ss).  Your central servers could be connected to a dedicated switch (stack), perhaps with dual 10 gig or connected directly to the core (additional stack members?).