Switching and Routing

Re: Redundant ethernet network project

Occasional Contributor

Redundant ethernet network project


I'm looking for reduntant network solution for a small firm. 

We are talking about 10 switch with 48 pt.

Based on my researchs (please correct me if I'm wrong) I understand that the best solution is star topology instead a ring topology.

So I have some question about HPE device.

1. Which devices are suitable for making a redundant star topology?
2. Which devices are suitable for peripheral nodes?
3, Star center switches and periferal switches shoulb connecter with spf module?
4. The star center switch needs to be in stack?

Thank you.  

Honored Contributor

Re: Redundant ethernet network project

Well...there is a single valid answer to each one of your questions because answers are strictly dependant to your desired topology (physical and logical).

  1. With regards to "Virtual Chassis" capability, actually you can decide to work with two main HPE/Aruba Switch series:
    1. HPE FlexNetwork (Comware 7 operating system based Switch series), those provide IRF technology to create a virtualized chassis made of multiple physical Switches interconnected as a Ring or Chain (Ring is better)...this can be deployed at Core level and also at Distribution/Access levels.
    2. Aruba (ArubaOS-Switch operating system based Switch series), those Switches provide:
      1. Backplane Stacking feature (with specific Stacking Modules and Stacking Cables) on models like Aruba 2920, Aruba 2930M and Aruba 3810M.
      2. Frontplane Stacking feature (called VSF, similar to IRF above) available on Aruba 5400R zl2 with v3 zl2 Modules only and on Aruba 2930F.
  2. With regards to Star topology versus Ring toplogy: you can also consider a Star of Chains/Rings...indeed if you have, at Core level, an IRF of two HPE FlexNetwork or a VSF of two Aruba's (or a Backplane Stack of two Aruba 3810M, as example) and then each access/distribution group is made of others Stacks...you will end up with a Start of Stacks and each Stack can be a Ring or a Chain of Switches (always better a Ring). Generally the Star's center should be an IRF or a VSF then your access/distribution should follow accordingly (other VSF's or other IRF's), within VSF and IRF you can deploy Ring or Chain topologies...consider that a Two-Members VSF/IRF means always you formed a Chain...Ring Topology enters in the scenario when you have more than two members to interconnect together.
  3. Connection between Star's center and periphery...it all depends on what througput, resilinecy and cable(s) lenght(s) you are dealing with (or you want to deal with). Generally you can do via Copper or Fiber Optics (SFP for 1Gbps and SFP+ for 10Gbps), use LACP as often as possible (redundant links).
  4. See above.
  5. I suggest you to familiarize with topologies described in this guide.

I'm not an HPE Employee
Kudos and Accepted Solution banner