Link aggregation is useful for connecting switches with multiple Ethernet ports and other devices with multiple ports. There is, however, more to it than simply connecting two devices to each other, and this article will help you understand all of it.
Simply put, link aggregation is a method used to bundle individual Ethernet links together in a way that lets them act like a single logical link. When used between two devices, link aggregation also balances the traffic between them. The process is not too difficult to achieve once you understand how it works.
Before we move on, there are many terms that are used interchangeably with link aggregation, but they all refer to the same process and there is no need to get confused. Some of these are Ethernet bonding or teaming; link bonding, bundling or teaming; port aggregation, channeling or trunking; and Network Interface (NCI) bonding or teaming.
Understanding the terminology used within link aggregation will help you understand how the process actually works. Any link aggregated system has three major components:
- The link aggregation group or LAG is the group of ports that are being combined. LAGs are also called bonds or teams, but we will use the term LAG hereon.
- The Link Aggregation Control Protocol is the protocol that monitors connected devices and allows them to add or remove individual links from the LAG.
- The set of rules that decide which data packets will be sent through which links, and which ones won’t be sent at all, is called the scheduling algorithm.
Network attached storage or NAS devices allow certain combinations (switch to switch, switch to server, switch to a NAS, switch to a router, etc.) which are made possible through link aggregation. The NAS learns to treat the LAG as a single link. The individual data packets on the links involved are kept intact when being sent from one device to the other using the LACP. A LAG can also be made part of a VLAN.
The main benefit of link aggregation, as you have probably realized by now, is a better utilization of physical resources by balancing the data traffic being sent across different physical links, as well as by taking advantage of the multiple ports that might be available to you.
Link aggregation is hence also cost-effective. It increases bandwidth without you having to spend money on more cables or equipment. Increase is a bit of a misnomer here though; what link aggregation actual does is that it allows each physical link present in the LAG to function at a higher bandwidth than it would by itself.
LAGs are more reliable as well because if one of the physical links goes down, the traffic is immediately distributed to the other links and balanced out between the, at the same time.
Types of LAGs
We’re now going into slightly more detail about each component in link aggregation, starting with LAGs; or more specifically, the different types of LAGs. LAGs can be dynamic or static.
Dynamic LAGs use LACP, where static LAGs do not allow this option. Static LAGs are sometimes called manual LAGs because you have to negotiate connection settings yourself rather than have the LACP do it. However, some devices that use manual LAGs do offer additional help with configuring settings.
LACP is an IEEE standard that lets devices send each other Link Aggregation Control Protocol Data Units or LACPDUs so they can establish a link aggregated connection. What LACP primarily does is that it prevents the misconfigured LAG settings, which is the most common problem that arises for most users in this process.
This is why dynamic LAGs tend to be more user-friendly; their use of LACP helps prevent errors. LACP also minimizes packet loss (of data packets), because if one link in the LAG stops sending LACPDUs (like if a cable is accidentally unplugged), it removes that link from the LAG.
Note that both devices need to have LACP enabled for you to be able to set up a dynamic LAG connection.
Scheduling algorithms are complicated in essence, but on the surface, they are just rules that LAGs need to follow. For instance, all links in a LAG must be the same type of Ethernet and the links themselves must be identical. Scheduling algorithms work by using hash functions to look at the MAC, source, and destination IP addresses in the fields in Layer 2 or 3 headers.
First things first, you need to make sure that both the devices you are trying to connect support link aggregation. Next, configure the LAG on these devices. You have to do it individually for each device. When configuring these LAGs, make sure to use the same settings for port speed, MTU size or jumbo frames, etc. on both devices.
Now check that all LAG ports are a member of the same VLAN. If not, then you need to first add the LAG to your VLAN (you don’t need to add the ports to your VLAN individually; simply adding the LAGs does the trick).
Connect the ports on your devices while keeping in mind which ones you added to your LAG. Don’t connect the devices with Ethernet cables before you set up a LAG though. This creates a network loop, which can slow down or even stop traffic between devices on your network. The LAG does issue a warning at this point if you have done something wrong up until now. If no such alert appears, you may now connect your devices with an Ethernet or fiber cable.
Perform some final checks now and then you’re good to go. Firstly, the LED light for each port on the switch should be blinking green. And secondly, in the admin interface, it should say that the link is “up” (as opposed to, obviously, “down” for links that cannot establish an aggregated connection for whatever reason) for each device. Re-check your configuration settings if one or both of these is not the case for you.
Link aggregation will be helpful for your router most of all (as compared to other connections you might be using). Link aggregation also works particularly well with Windows 10. At any rate, now that you know how to, go and make use of all the ports on your devices that you weren’t utilizing before!