We will be installing two new 6513?s with Sup 720?s in our datacenter each residing on a separate floors in the same building. Connecting the switches is an 8 GB Ether Channel.
Would it be better to use HSRP or GLBP between switches?
Standard HSRP offers redundancy and with different group config you can have load balancing too.
But with the standard GLBP config you can achieve both redundancy and loadbalancing.
Although GLBP ans HSRP perform almost similar things, there are a few small differences between them. For example in GLBP a virtual MAC address is assigned to each device in the group whereas in HSRP only one virtual MAC address is shared by all the devices.
Although according to me there is no real benefit of using GLBP over HSRP. But I think HSRP is known to most of us, so its easier to implement!
Check out this link -->
Hope this helps,
Please rate all posts!
For me, I recommend using GLBP, you will achieve availability, and load- sharing. i think you will need to configure two groups if you need HSRP load-balancing "sharing". And, i worked with GLBP, it is very easy according to the configuration.
This is from my side!
HSRP and Virtual Router Redundancy Protocol (VRRP) target single points of termination in a Layer 2 access design. HSRP and VRRP provide a backup "standby" default gateway for users connected in the wiring closet.
Gateway Load Balancing Protocol (GLBP) is a Cisco extension to HSRP. GLBP allows users to use the HSRP standby router while functioning as a standby. This allows better utilization of network investments by load balancing traffic across the active and standby gateway routers.
As far as I know the only benefit of GLBP over HSRP is that we only need to have one virtual IP. So all the comouters have same IP configured as default gateway and still their data can be load balanced across multiple routers whereas to achieve the same with HSRP, we need multiple standby groups each with its own virtual ip address.
I would go with GLBP for its redundancy and load balancing.
I would not use it for providing gateways for hosts if any of your hosts are linux. We had to revert back to HSRP for hosts due to Linux having some issues (kernel issues) with it.
It really dosent matter, 6513&720's you looking at about 40gig throughput on the backplane so either way Load balanced(GLBP) or un-load balanced (HSRP) redundancy protocols will work. If you ever start to hit 40 gigs of throughtput you could distribute the load using GLBP, aside from storage and streaming content networks I dont know of many other networks that could easily hit that.
I know I used it back in 04 and it was still really buggy, blackholing packets. Had to revert to HSRP.
Consideration should be given to aligning VLAN paths to certain switches when employing service modules. Typically, active-active service modules such as the ACE are only active on a per-context basis. It might be better to use HSRP to ensure VLAN path and spanning-tree alignment to minimize taking extra switch hops.