Considering buying a new switch. Typically not all to big an issue. Check Datasheet and fine. For smaller use cases this is probably feasible but I experienced never the less reaching some limits. So what to do?
Right now I started to do a benchmark and wanted to test out in which configuration a network switch actually maxes out. Besides max- MAC entries, max routing entries or multicast groups particular the number of differen VLANs results in memory consumption, since each one needs dedicated buffering.
On top most VLANs need IP configurations set up and depending on your applications probably multiple of them. Unfortunately in my case exactly the later.
Now you may consider configuring max configurations some fun, but after several loops to find the setup maximum, this does not exactly look like productive work. To create multiple configurations which then might be
Continue reading →
Approaching a certain quality level of switching and routing, high availability evolves to be an obligation. In these terms, according to the different OSI service layers, there are many high availability protocols, securing the according network services. The Spanning Tree family as STP, RSTP, MSTP, PVST, protocols for link aggregation as LACP and layer three routing redundancy services like VRRP.
These protocols have the advantage, being vendor independent standards and presume to be interoperable. But either design gets complex, interoperability keeps its caveats or ressources are simply disabled and take over in failiure. Thats not exactly performance driving.
So vendors created stacks – which failed otherwise, or they started to create systems of higher complexity which proprietary created load sharing high availability clusters in the Continue reading →