Network newbie trying to learn something new...

[Gezus]

As the subject line says, I’m totally new to networking past setting up a consumer grade TP Link router… I’ve spent the last couple of days and nights researching articles, forums, videos, etc. trying to wrap my head around all things networking in an attempt to make setting up my Mikrotik devices easier. I’m realizing there is a lot of pre-configured items in consumer routers I wasn’t aware of…
With that being said, I have a few questions I’m hoping will help me understand how to configure my system (Hex S router and CRS326-24G-2S+RM switch).

So we’ll say my Gateway and DHCP server have IP Address 192.168.0.1 and will hand out addresses from 192.168.0.10 to 192.168.0.254 (Subnet is 255.255.255.0). Now Comcast sends my dynamic IP address to my modem (I use a personal Netgear modem with no routing capabilities) which in turn is sent to my Hex S router through Port 1. Is the IP Address of Port 1 the Gateway Address?

In order for any device connected to my router to get internet access, I need to bridge each port to port 1. In the standard “4-port run of the mill” consumer router, is this done automatically?

Now, let’s say I want to add my CRS Switch to the system for additional hosts. Because this switch is a physical piece of hardware different than the router, is it a different “network” that needs to be bridged to the router? If that is the case, does that mean I need to setup a completely different DHCP server on my router to handle the switch?

With respect to the switch (and actually the router as well), is every port considered a separate “network” that requires I bridge them together before anyone can talk to another one? Does each port have it’s own MAC address that can be assigned an IP address? If so, are consumer grade routers the same way?


I’ve watched several different videos on how to setup my router and switch, most of them were old (2016/2017) and there were several different ways certain things were done. Add to the fact that RouterOS allows you to select options which theoretically could achieve nothing and this makes following what is actually happening difficult for me to piece together and follow. If you’ve made it this far and are willing to answer my questions, I cannot thank you enough! I’ve literally spent the last 3 days and nights (I’m going on maybe 12 hours of sleep) trying to learn as much as I can and understand how to set this system up.

[pukkita]

So we’ll say my Gateway and DHCP server have IP Address 192.168.0.1 and will hand out addresses from 192.168.0.10 to 192.168.0.254 (Subnet is 255.255.255.0). Now Comcast sends my dynamic IP address to my modem (I use a personal Netgear modem with no routing capabilities) which in turn is sent to my Hex S router through Port 1. Is the IP Address of Port 1 the Gateway Address?

No. when you say “which in turn is sent” What do you mean exactly? Is there a dhcp-client on the HeX on ether1?

Remove ether1 port from bridge1 making sure is isn’t on any bridge, then add via ip > dhcp-client a new dhcp-client on ether1, making sure Add Default Route is ticked. This will auto setup the HeX to use the Netgear to reach Internet. No other dhcp-clients should exist.

In order for any device connected to my router to get internet access, I need to bridge each port to port 1. In the standard “4-port run of the mill” consumer router, is this done automatically?

That’s because you haven’t done the dhcp-client part. By doing that you turned it into a 5 port switch…

Consumer routers are pre-programmed. Routerboards are programmable. You can have the LAN bridge with 4 ports, or 3 or 2 or as you see fit.

Now, let’s say I want to add my CRS Switch to the system for additional hosts. Because this switch is a physical piece of hardware different than the router, is it a different “network” that needs to be bridged to the router? If that is the case, does that mean I need to setup a completely different DHCP server on my router to handle the switch?

No need for additional DHCP Server if as you point, the CRS is bridged to the HeX, putting both on same network segment.

With default CRS config (all ports bridged), make sure it uplinks to the HeX on a port (e.g. ether3) that actually belongs to HeX LAN bridge, bridge1. This way CRS and HeX will be on same broadcast domain (same network segment) and DHCP will be managed by the HeX.

Say you have a 8 port CRS; if you add all 8 ports to a bridge on it, then connect its ether8 to HeX ether4, where HeX LAN bridge has ether2-ether4, now CRS ether1-8 and HeX ether2-4 will act as a single switch with 12 ports.

With respect to the switch (and actually the router as well), is every port considered a separate “network” that requires I bridge them together before anyone can talk to another one? Does each port have it’s own MAC address that can be assigned an IP address? If so, are consumer grade routers the same way?

Yes, Yes, and it depends.

As stated earlier, Routerboards are 100% programable. Consumer grade routers are pre-canned devices.

In any Routerboard, if you add all the ports on a bridge, you create a switch, and any IP address should be assigned to that bridge interface, becoming the switch management IP.

If you leave all ports free, not assigned to any bridge, you get a 100% router device. You can assign IPs to any port; each port has its own mac address.

I’ve watched several different videos on how to setup my router and switch, most of them were old (2016/2017) and there were several different ways certain things were done. Add to the fact that RouterOS allows you to select options which theoretically could achieve nothing and this makes following what is actually happening difficult for me to piece together and follow. If you’ve made it this far and are willing to answer my questions, I cannot thank you enough! I’ve literally spent the last 3 days and nights (I’m going on maybe 12 hours of sleep) trying to learn as much as I can and understand how to set this system up.

As said, Routerboards are 100% programable, so requires the admin to actually know how networks work in real life to properly program them; nevertheless, there’s also QuickSet, which offers a preconfigured set of typical configs, similar to consumer router canned configs.

[Gezus]

Pukkita, thank you very much for your reply and I apologize for not getting back to you earlier. After reading, I needed to try and digest everything and I realized I’m getting hung up on bridges and how they function exactly, but just to make sure I’m understanding what you wrote I’d like to ask about your responses. If this is too much for this forum, I can ask the reddit community, just let me know!


So forgetting about the CRS for a moment… You said to take ether1 off of bridge1 and set it up as a DHCP-Client. What I’m gathering from doing this, is ether1 will be a DHCP-Client for my ISP which is the DHCP-Server, is this right? That would make sense. One video I watched said to bridge all 5 ports together and setup a DHCP-Server for all of them. I ended up only being able to access the internet from one computer.
From there, data coming into ether1 would be sent to the Gateway/Router, which in my case would be the HexS CPU? From there, the data will be directed where to go by the router. I’m also going to setup a DHCP-Server for my internal network. All of these things will be handled by the HexS CPU, so all three (Gateway/router/DHCP-Server) will have the same IP Address? So far, so good…?

Now for the bridge part… If I don’t bridge the remaining 4 ports (ether2-5), they will all act as separate networks, each requiring their own DHCP-Server (or static IP address) correct? What if I assign each port to the same subnet, forcing them to be in the same network?

If I do bridge the remaining 4 ports together, they become part of the same network? What if I assign each port to a different subnet, will that screw things up? Since a bridge can be a physical or virtual device, does the bridge itself have an IP Address/MAC Address? If so, does it need to be assigned one or either of these manually?

Hypothetically, if comp2 on ether2 wanted to talk to comp5 on ether5 and there was NO bridge, they would by default, be on separate networks and require the data to be sent to the router which would use their IP Addresses to direct traffic; however, if the ports were bridged together and acting like they’re on a switch, the data would be sent to the bridge, which would use MAC Addresses to direct traffic? In the case of the HexS they’re both using the CPU to accomplish the same goal, is there a benefit to doing either? I realize they’re different layers in the OSI model, but I’m not familiar with how it may affect speed or anything else in a negative or positive way.

[Gezus]

I ended up purchasing some classes on Udemy. Guy outlines how bridges work as well as how everything else is connected. Thanks again for the help!

[pukkita]

So forgetting about the CRS for a moment… You said to take ether1 off of bridge1 and set it up as a DHCP-Client. What I’m gathering from doing this, is ether1 will be a DHCP-Client for my ISP which is the DHCP-Server, is this right? That would make sense.

Yes.

From there, data coming into ether1 would be sent to the Gateway/Router, which in my case would be the HexS CPU? From there, the data will be directed where to go by the router. I’m also going to setup a DHCP-Server for my internal network. All of these things will be handled by the HexS CPU, so all three (Gateway/router/DHCP-Server) will have the same IP Address? So far, so good…?

Yes, that’s correct.

Now for the bridge part… If I don’t bridge the remaining 4 ports (ether2-5), they will all act as separate networks, each requiring their own DHCP-Server (or static IP address) correct?

Yes.

What if I assign each port to the same subnet, forcing them to be in the same network?

IPv4 doesn’t work this way. You wouldn’t be forcing anything, but creating an invalid configuration. Same network means same network segment (e.g. all in same bridge).

If I do bridge the remaining 4 ports together, they become part of the same network?

ether1-4 in bridge, and bridge device having four different, same network range IPs (e.g. 192.168.1.4/24 , 192.168.1.5/24, 192.168.1.6/24, 192.168.1.7/24 on the other hand, it is a valid configuration.

What if I assign each port to a different subnet, will that screw things up? Since a bridge can be a physical or virtual device, does the bridge itself have an IP Address/MAC Address? If so, does it need to be assigned one or either of these manually?

if they’re already part of a bridge? YES. Once you create a bridge, any L3 addressing shall be done on the bridge interface itself.

Bridges will always have a mac address. assigning an IP to the bridge is up to you depending on the need; in your scenario you need it. But there are others where isn’t required.

Hypothetically, if comp2 on ether2 wanted to talk to comp5 on ether5 and there was NO bridge, they would by default, be on separate networks and require the data to be sent to the router which would use their IP Addresses to direct traffic; however, if the ports were bridged together and acting like they’re on a switch, the data would be sent to the bridge, which would use MAC Addresses to direct traffic? In the case of the HexS they’re both using the CPU to accomplish the same goal, is there a benefit to doing either? I realize they’re different layers in the OSI model, but I’m not familiar with how it may affect speed or anything else in a negative or positive way.

HeX uses hardware offloading (if you enable it on the bridge and the ports) to forward between ports on the bridge, which means zero CPU load. Routing (L3) uses CPU. Bridge w/o hw offload uses CPU, but generally less than in L3.

Devices on a bridge can communicate directly, i.e. without frames passing through the CPU.

I think a look at https://help.mikrotik.com/docs/display/ROS7/Computer+Networks will clear things up, specially the ARP and Tying It All Together section.

[Gezus]

Pukkita,

Thank you very much for replying to my questions and for supplying the link for more information. It all helped tremendously. I’m glad to know I can post very basic and general networking questions on this forum and have them be answered without a “if you don’t know then you shouldn’t be playing with these things…” comment.