Associate
- Joined
- 2 Sep 2013
- Posts
- 2,303
Hi all, more of an academic "I would like to know more" type of question here this time, as I already have the network up and running and doesn't appear to have any issues.
So the home network I've got up and running at my parents place is essentially a 10g network (cat6 connections) with the "core switch" being made up of two Mikrotik CRS304-4XG-IN switches that have been daisy chained up to serve the various rooms and devices. It all works fine (full 10g 1.11-1.12GBps file transfer speeds at base MTU settings, etc), but I read something about it being undesirable to daisy chaining the switches via ports to form a larger (core) switch as being undesirable due to being limited to those port speeds; That got me wondering about if the network I have setup would experience this same issue that was discussed or not? I'm of the mind "no", but just wanted more experienced minds on networks to give a heads up on it or not.
So the purple boxes are switches in the diagram, with the two on the upper middle right being the two Mikrotiks mentioned.
The cyan boxes are wireless access points (And the router).
The blue boxes are the 10g devices.
So the item I read about online, says that the daiy chain approach is not ideal since the link is limited to the ports data rate rather than the data plane at the back of a singular switch (used as a core switch). This I get, especially if all devices operate at full speed and constantly hammering the whole network. However, the setup of the network has all the 10g active devices (which only communicate with each other when I ask them to - for file transfers or backup or Remote AI work; are all user induced) are all connected to the second daisy chained switch and all other devices connected to the first switch. So should that mean that the issue with the link to daisy chain up switches to effectively create a larger core switch is not expected to be an issue here on this network given it's use style and layout?
In the future (when there's more in the family that might be using the network more at the same time), I'm thinking that's when I need to considering making a switch to a singular larger switch for core switch use? But for now - given it's really just me doing any heavy network use (the rest is VOD and Youtube by my parent) it should mean the port link issue is not a concern?
Thanks for any clarification/confirming and if there's anything you noticed off with what I've posted.
In case anyone is wondering why I went for two switches instead of one larger one, there was several considerations for this including:
1. Price: It costs less to grab two of these switchs vs a similar switch with the same amount of ports on.
2. Space/Size/Wall Mounting: My parents mobility and falls risk (trip hazzard) being a heavy concern, means you can't have stuff just there that they might try and avoid and then increasing their risk of a fall, so as close to the wall as possible, etc and why no larger switches, etc. or cabinet or rack to house everything.
3. Fanless: These are fanless vs the fanned ones with the amount of ports needed. Unfortunately I need to be able to have the equivalent of Superman hearing in case I am called, so fanned switches even on low settings is just going to reduce the chance of me hearing my name being called for assistance for anything, etc.
4. Power Delivery/UPS: To make sure I can keep the first switch alive for 3+ hours (older concern in case rolling blackouts are introduced by the grid), it can only be done with 1 switch being kept alive. A larger switch would go through more power and send that UPS up time of over 3 hours to below it. We shouldn't need it to last that long, but it was a design choise/consideration. (This has been bypassed by having a second UPS that powers the second switch of the "core switch" and the living room TV in case we need to monitor anything on the news).
5. Migrating Hardware: Had an older QNAP 10/2.5 switch that has been moved to a different room to faciliate WOL and 10g connection there. So 1 (10g Mikrotik) switch alone wasn't enough to connect everything up.
So the home network I've got up and running at my parents place is essentially a 10g network (cat6 connections) with the "core switch" being made up of two Mikrotik CRS304-4XG-IN switches that have been daisy chained up to serve the various rooms and devices. It all works fine (full 10g 1.11-1.12GBps file transfer speeds at base MTU settings, etc), but I read something about it being undesirable to daisy chaining the switches via ports to form a larger (core) switch as being undesirable due to being limited to those port speeds; That got me wondering about if the network I have setup would experience this same issue that was discussed or not? I'm of the mind "no", but just wanted more experienced minds on networks to give a heads up on it or not.
So the purple boxes are switches in the diagram, with the two on the upper middle right being the two Mikrotiks mentioned.
The cyan boxes are wireless access points (And the router).
The blue boxes are the 10g devices.
So the item I read about online, says that the daiy chain approach is not ideal since the link is limited to the ports data rate rather than the data plane at the back of a singular switch (used as a core switch). This I get, especially if all devices operate at full speed and constantly hammering the whole network. However, the setup of the network has all the 10g active devices (which only communicate with each other when I ask them to - for file transfers or backup or Remote AI work; are all user induced) are all connected to the second daisy chained switch and all other devices connected to the first switch. So should that mean that the issue with the link to daisy chain up switches to effectively create a larger core switch is not expected to be an issue here on this network given it's use style and layout?
In the future (when there's more in the family that might be using the network more at the same time), I'm thinking that's when I need to considering making a switch to a singular larger switch for core switch use? But for now - given it's really just me doing any heavy network use (the rest is VOD and Youtube by my parent) it should mean the port link issue is not a concern?
Thanks for any clarification/confirming and if there's anything you noticed off with what I've posted.
In case anyone is wondering why I went for two switches instead of one larger one, there was several considerations for this including:
1. Price: It costs less to grab two of these switchs vs a similar switch with the same amount of ports on.
2. Space/Size/Wall Mounting: My parents mobility and falls risk (trip hazzard) being a heavy concern, means you can't have stuff just there that they might try and avoid and then increasing their risk of a fall, so as close to the wall as possible, etc and why no larger switches, etc. or cabinet or rack to house everything.
3. Fanless: These are fanless vs the fanned ones with the amount of ports needed. Unfortunately I need to be able to have the equivalent of Superman hearing in case I am called, so fanned switches even on low settings is just going to reduce the chance of me hearing my name being called for assistance for anything, etc.
4. Power Delivery/UPS: To make sure I can keep the first switch alive for 3+ hours (older concern in case rolling blackouts are introduced by the grid), it can only be done with 1 switch being kept alive. A larger switch would go through more power and send that UPS up time of over 3 hours to below it. We shouldn't need it to last that long, but it was a design choise/consideration. (This has been bypassed by having a second UPS that powers the second switch of the "core switch" and the living room TV in case we need to monitor anything on the news).
5. Migrating Hardware: Had an older QNAP 10/2.5 switch that has been moved to a different room to faciliate WOL and 10g connection there. So 1 (10g Mikrotik) switch alone wasn't enough to connect everything up.