Category: Networking

  • Why I No Longer Use a TP-Link ES208G (or any front port facing switch) as My Home Lab Main Switch

    Estimated read time is 5 minutes.

    The quick answer

    It is cleaner from the front viewing angle. It requires fewer parts and overall rack space (1U vs 2U minimum).

    Intro

    Whether you’re just connecting a few things up your router or you’re a seasoned Ubiquiti Chad, there are options out there for us regarding network switches featuring rear ports that even a data center technician agreed with on a Facebook post I regret not snipping for you as evidence.

    This article is not tied directly to two models of switches, and can be used in general with the understanding that many switches exist that are very similar from many different brands.

    Why I Bought the TPLINK ES208g

    I realized I needed a switch. I had multiple PCs we’ll call “nodes” without Wi-Fi cards. An ethernet cable from my router to my switch then each node connected to the switch was a working pair

    The 1 gigabit speed ended up being plenty for me, and still is – as I’m not a Ubiquity bro yet. Anyways, upon stacking my gear with this TPlink switch, things became a little squirrelly.

    • All node ethernet ports were on the back of rack, forcing me to run longer ethernet cables to reach the front of switch.
    • End result was unsightly, switch situated backwards to avoid longer than necessary cables, link LEDs were no longer visible,
    • all in all, not a setup I was happy with

    I ended up turning the entire switch backwards to eliminate this because my cables were short. I had purchased this 5-pack of 1-foot copper cables on Amazon.

    Research led me to a keystone patch panel. This is a blank panel where you take very short leads of ethernet from the front of switch, and connect to the front of the patch panel. Then on the backside of the keystone panel (inside the server case) you take a separate lead and connect that to the ethernet port on your node(s).

    The Big Dilemma

    A front port switch + Keystone switch = 2U rack space minimum, plus approximately double the amount of additional parts and cables. See model below:

    Front-port network switch connected through a keystone patch panel.

    Don’t get me wrong, the above setup can look cool. The short jumpers become the exhaust manifolds of your energy hungry technical project where if it were the 1950’s, would actually be a v8 muscle car and actual exhaust manifolds. This was also when the actual clouds were where our important data was stored.

    Twice the space, twice the parts, twice the opportunity of failure; not necessarily required to function either. I didn’t want to buy new parts or 3d print them, figure out all those custom cable lengths and spend another $35 on cables and connectors. I ended up finding a simpler efficient solution. Bring in the forgotten all-star to us in the home lab world.

    The Rear Facing Port Switch

    Modeled below:

    Rear-port Cisco SG200-style switch with Mini PCs.

    It is much less complicated and achieving the same objective. I found a commonly available device: Make = Cisco and Model = SG200-08. It was designed for small businesses. Perfect as I want my setup to be as “enterprise grade” without spending a ton. The standard SG200-08 can be powered through PoE on port 1, but it does not supply PoE to connected equipment. The larger SG200-08P version can provide PoE through four ports. I bought it on eBay for about $30. eBay is probably the best place to find these in this modern day. As for newer models with rear facing ports, they’re available and your best bet is online as for some reason they’re much less common than a front port models in stores.

    My advice home lab builders is to get a rear port facing switch first for the following reasons:

    • 50% Less rack space. Especially if you don’t have a rack, this will look much neater hiding all the ports and connections in the rear. You still get LED connection indicators on the front so it remains interesting to look at.
    • Less chance of a wire being unplugged or snagged on something
    • Cleaner front facing appearance
    • Form factor is roughly equivalent to comparable front facing port switch of same port count
    • No need for several more cables, parts, or keystone panel

    Why Not to Choose A Rear Port Switch

    • Difficult access when working on network, i.e., plugs are in the back and not always accessible
      • Only an issue in a very large rack butted up against the wall. Most home labs are smaller and can easily be moved for rear access
    • Potentially unaware these switches exist or unable to find one in retail environment

    Here is my current home lab setup.

    More detail of my home server coming soon, but in short it is 4 compute nodes running Proxmox OS. What I find most exciting now is firing up all nodes on BOINC, hearing them warm up and fans on high as they work hard for science. Web hosting and design, home network ad blocking, and self hosted cloud storage have been a real treat to learn about too.

    Related reading: My original TP-Link ES208G home-lab article

  • TP-Link Omada Managed Switch Guide

    Estimated reading time: 4 minutes

    From a budget minded home lab beginner

    Introduction

    A TP-link ES208G from amazon just arrived in the mail and I want to understand it before I use it in my small home server. It is an 8 port managed network switch. I want to take you through the learning I’m doing in this project. From what I’m looking at, it is something a normal person at home like me can figure out – that is, unless I’ve been lying to myself about being normal. Side note, this is not a full enterprise grade switch giving you every setting under the sun. Its function is to provide more settings than your typical “Plug and play” “Easy switch.” For context, I wanted a real server feel. I want to experience some higher level server management tasks along the way and this particular switch was $28, only $3 dollars more than most new reputable easy switches online.

    Getting started

    I began with the question “how do I get into the switch to make adjustments?” I’m sitting at Starbucks right now with my server running at home connected to the switch in port 2, and the output of the router feeding port 1. I learned right away that remotely accessing your switch is a no-no unless you have configured your VPN (I’m using Tailscale) a certain way to where it allows you access to your home IP. More on that later. I expected it to be like routers I’ve worked on before where the default IP is consistent across routers as well. Easily enough, that’s all I needed though was to log in to my router, click on my devices connected, in my case there were 10 unassigned ones. The switch was identifiable as device “ES208G” with an IP address of 192.168.4.101

    Type that IP into your web browser and you will find your device setup page. Create username and password first.

    This is where an easy switch and managed switch differ from one another.

    1. First thing I want to do is give the switch a permanent IP address to keep the login process consistent.
      • In the System Info dropdown, select IP Settings. You want to disable DHCP settings to ensure a static IP.
      • Choose and enter an IP that is freed up. I chose 192.168.4.101 was free because that’s what it originally was and I don’t need any extra help confusing anything.
      • Subnet mask should typically be 255.255.255.0
      • Gateway is a little more tricky, but find your routers IP address and enter that into Default Gateway. Click Apply. Make sure window is wide enough, my apply button was hiding off to the right side.
    • Rename the switch. Let’s be honest, Tina has a much better ring than Camry. This will be in the initial first few setup questions, for me it was.
      • I named mine OmadaSwitch1
      • With potentially many network devices, this will make the switch easier to recognize.
      • Select Apply
    • Rename Ports.
      • With this switch not being a full enterprise grade piece, we actually cannot rename the ports. However, if you’re following along with a different model/brand, this is one thing to do a little clicking around for. If I were able to, I name ports something like below:
        • Port 1 – Router Uplink
        • Port 2 – HP ProDesk Server
        • Port 3 – Security Camera/Access Point/Desktop/etc.
          • For my current build, I have: Modem –> Router –> Switch –> Server with the “–>” symbol indicating an ethernet cable
          • in my case I could create a txt file for instance with this information and potentially give to my local AI model to retrieve one day in case I forget.
    • Leave speed and duplex on Auto
      • Unless we are trying to solve a specific issue here, my sources lead me to leave Auto Negotiation enabled, Speed set to Auto, as well as Duplex.
    • Verify the link
      • In the Switching>Port Settings screen, you should see your configurations such as Speed + 1000M or 1 Gbps.
      • You should also see full duplex the ports you have in use’s state equaling “Enabled.”
      • In the Monitoring>Traffic Summary menu, you should see a number of Tx bytes and Rx bytes indicating that data is indeed being Transmitted and Received, respectively.
    • Final steps
      • Ensure you have applied all your settings first, then close your browser instance. Re-open and verify your settings remain applied.

    This is a good place to leave off for now. While this guide covered the basic configuration needed to bring my managed switch online, there’s still a lot left to explore. Features such as cable diagnostics, traffic monitoring, MAC address tables, VLANs, Spanning Tree Protocol (STP), QoS, link aggregation, and security settings each deserve their own deep dive. Rather than enabling features just because they’re available, I’d rather understand what problem each one solves first. As my homelab grows, I’ll revisit these tools one by one and document what I learn along the way.

    Next Article: Why I no longer use the TP Link ES208G for my home lab main switch.