Cascading switches

A customer asked us for a price for 21 switches.
They need sfp in and out of every switch but only 5 ports in each switch. It’s just measurement data from the building that will pass through this network, nothing else.

However, I have never built a network with 21 switches connected in series. I would never do this for a network otherwise but the fiber cable is already there, going from house to house.

The connection for this to the internet is only 2mbit/s so there are really small amounts of data.

Are there any pitfalls for me here?

Best regards

Johan

No real pitfalls to my knowledge.
Sure, such topology is a big single point of failure but apart from that it will work just fine, especially for such low flow of data.
There is no option to form some sort of “ring” ? Are these 21 buildings/switches really “in a line”

You probably have to disable (R)STP since it will hit the diameter limit.

Rapid Spanning Tree protocol (RSTP) → The maximum allowed network diameter for the RSTP protocol is 40 switches.

But anyway, with only 1 fiber line I would disable any STP anyway since it is not relevant and of no value

Thanks for your input here :slight_smile:. Feel alot more confident now that it will work, however not optimal :stuck_out_tongue:.

@jvanhambelgium - Just curious, why do you want to turn off STP considering there will likely be multiple devices connected to each switch? BTW, I suspect there might be some kind of BMS/HVAC management system hooked up to each building.

STP has nothing to do with number of devices connected to each switch, it has to do with loop detection and prevention. While one can never be sure there won’t be any loops (that’s possible as soon as switch/bridge has more than one port), it’s the matter of probability to create a loop and how much damage can happen until the loop is detected (manually) and broken.

If access to switch ports won’t be possible to anyone, then probability of creating a loop is slightly lower … and xSTP does come with certain “benefits” which may outweigh the basic functionality in certain cases.

Thanks @mkx, I’m quite aware of the functionality. In this case ‘devices’ additionally includes L2 communication links that some BMS systems automatically generates for extra redundancy. It might also mean possible redundancy between the switches, as most fibers (presumably multimode in this case) are installed with at least one pair which I’d recommend to use in any case.

With that said, I was merely wondering about the reasoning behind @jvanhambelgium’s motivation to disable STP in this particular case. I can’t see any obvious benefits; in fact, quite the opposite, if you want to add redundancy as described above. Personally, I would never disable STP just for the sake of it.

There are no benefits in disabling STP for sure and I was only looking at the uplink “line” not the different endpoints.
Without some form of STP somebody could cause some havoc when connecting an endpoint the wrong way, having other switches introduced that you are not aware of and forming a loop etc,etc.

So yeah, just leave it on :smiley:

Yeah, that makes sense.