Is anyone benchmarking/testing switches?

Shane86 · June 30, 2019

I came into a bit of an interesting situation today..

Earlier this week, I was building a new server on a live stream, so i needed an Ethernet jack near the computer i was streaming from. I dug into my "old stuff" pile of hardware that i had before i moved, and fetch an old TP Link TL-SG1005D that i had previously used between my home server and desktop when i wanted to split the drop at my desk.. patched it in.. online, everything seemed hunky-dory.

Today, i'm surfing around on that same streaming desktop and noticing lower performance from steam, funky page loads. run a few pings, and boom.. loosing packets.. Surely this is my ISP.. no. it's INSIDE THE NETWORK! (cue murdered teens)

Verified my server to server migration happening in the background is still happening over the direct wired cable between them, no real load on my Meraki 8-P switch.. then it hit me.. that TP-Link is still in the mix.

Yank it out, and internal pings go from 6-7ms to sub 1 to the router.

Holy Crap.

I'd been running this turd of a switch, figuring anything with the word Gigabit on it wouldn't begin to be tapped out for my home use for years.. and it's likely been messing with my universe for as long..

so this got me thinking.. I don't know of anywhere, or can find anything, of people benching budget networking gear, and how many people may be out there in the same situation.. oblivious to the impact on their network a cheap switch is actually having.

GzeroD · June 30, 2019

SmallNetBuilder used to do a lot of reviews in the space your looking at. they might not have anything for the newest parts but if your just looking to make sure whatever you pick up isnt terrible and weren't picking up the latest thing it should at least give you an idea of who is putting together quality kit.

Edit: Servethehome would also be a great resource to check out.

Windows7ge · June 30, 2019

Even cheap switches for the basic purpose of web browsing or transferring files will run just fine at a given speed. There's nothing to really benchmark in this context.

Now there are ways and reasons to benchmark switches for applications such as testing load capability, routing capability (for switches that are partial layer 3), among many other tests (for things like security), but for you I can just about guarantee the issue isn't how the switch is designed to function rather what is likely a hardware failure in some form be it a bad cable, failing interface, or other.