Dante Certification Level 2 – Chapter 08: Unicast, Multicast & Broadcast

 

Understanding unicast, multicast, and broadcast is essential for designing efficient Dante networks, and this chapter breaks down each communication method with practical examples drawn from real AV systems. Unicast sends a packet from one host to a single destination and delivers the lowest latency, making it the default behavior for Dante audio and video flows. Broadcast sends data to every device in the local network, which is useful for processes like DHCP and device discovery but inefficient for media streams. Multicast bridges the gap, allowing a transmitter to send a single stream that the network duplicates only to subscribed receivers, which dramatically reduces bandwidth usage when distributing the same audio or video to multiple destinations. You will also learn how IGMP snooping and the IGMP querier control multicast traffic so it reaches only the ports that need it, and why Dante supports both IGMP snooping version 2 and version 3 for interoperability. By the end, you will know when to choose unicast versus multicast in Dante, how to recognize each flow type in Dante Controller, and how to design networks that scale efficiently with multicast video and audio.


Key Learning Objectives

 

By the end of this chapter, learners will be able to:

  1. Define unicast, multicast, and broadcast transmission and identify when each is used.
  2. Explain the role of IGMP snooping and the IGMP querier in managing multicast traffic on a Dante network.
  3. Recognize unicast and multicast flow indicators in Dante Controller.
  4. Compare the bandwidth and latency tradeoffs of unicast and multicast for Dante audio and video distribution.
  5. Apply best practices for configuring switches when multicast traffic is present in a Dante system.

Want access to all of the Dante Training Courses?

Enroll Now

View full transcript

Level 2 – Chapter 08 – Unicast, Multicast, and Broadcast

As we’ve seen, networking involves sending data from one device to another, but the method used to deliver that data can vary. Unicast, multicast and broadcast are the three primary types of communication on an IPV four network. Each serves a specific purpose and is used based on network requirements. Let’s dive into each one first.

Let’s start with unicast transmission. Unicast is the process of sending a packet from one host to another. This method is reliable and commonly used for internet traffic, like web browsing, emails and video streaming where each user receives a separate data stream. However, it can be inefficient if many users request the same data separately.

Think of uncast like making a direct phone call where only the intended recipient receives the message. Unicast is great when I want to send unique information to one destination, but what happens when I want to send the same information to multiple destinations? If I’m sending the same data to multiple destinations using unicast, I have to duplicate those transmissions over and over again. In this example, you can see that I’m putting a strain on the transmitter.

Eventually, it would run out of bandwidth. Not only that, if I have multiple switches, it will put additional data on the trunks. The data would multiply what is being sent across this link. Unicast is great for one-to-one communication where you want to achieve the lowest latency possible, even when splitting signals.

There may be times when I’m going to choose unicast, but if I’m trying to do a large distribution to many receivers, there are better ways to accomplish this. Let’s compare unicast with broadcast transmission. Broadcast is the process of sending a packet from one host to all hosts in the network. It’s the exact opposite of unicast transmission, and it allows communication without specifying a particular recipient, making it essential for processes like device discovery and network management.

If we use the analogy of a birthday party invitation, broadcast messages would be like creating a single group on my phone with all my contacts and sending them all a message at once, or going out on my street with a loudspeaker and inviting all the neighbors on my block. People I may not even know would receive the invitation. When a broadcast message is sent to the first switch, that switch will send it out to all ports. Of course, one of those links goes to another network switch, and that switch will send it out to all of its ports as well.

You’ll notice that broadcast messages don’t cross the router, and that’s a good thing. That means when your computer at home sends out A-D-H-C-P request, it stops at the edge of your home and it doesn’t go out across the internet to every network in the world. Broadcast transmissions are generally divided into two categories, limited broadcast and directed broadcast. A limited broadcast is a message that is sent to all devices within the local network, but does not extend beyond it.

It uses the IP address 2 5 5 2 5 5 2 5 5 2 5 5 as the destination address, and it is confined to the local network segment. When a computer sends a broadcast, all hosts in its local area network will receive it including the default gateway, but the router will not forward this broadcast. A directed broadcast on the other hand, is targeted at all devices within a specific network, including potentially remote networks. It uses the IP address with all host bits set to one.

For example, if the network address is 1 9 2 1 6 8 1 0 slash 24, the directed broadcast address would be 1 9 2 1 6 8 1 2 5 5. Directed broadcasts reach all hosts on a specified network even if they are not in the same local segment. In this case, routers can be configured to forward direct broadcast messages, but this is often disabled by default for security reasons when we’re sending large messages. The problem with broadcast is that it goes to every device on the network.

That means a lot of devices are getting data that they simply don’t need, and we could quickly overwhelm the interfaces of those devices. Generally speaking, broadcast is used as a method of locating other devices on the network, not as a means of routine communication. A-D-H-C-P request is a great example of where broadcast is useful. When a device first joins a network, it doesn’t know where the DHCP server is, so it sends that initial request as a broadcast message.

Of course, when it’s a Dante device, it’s transmitting audio or video data to other devices on the network. In this case, it will know where the other Dante devices are, so there are better alternatives for us to use. Multicast transmissions aim to split the difference between unicast and broadcast. It’s the process of sending a packet from one host to a select group of hosts potentially on different networks.

This means that the packet will not be sent to all hosts, but instead to a group of hosts that have specifically requested that data. Broadcast transmission could potentially saturate the network while multicast transmissions reduce traffic to only those hosts who request it in networks. If we are using only one switch, the switch creates a distribution list of hosts that want to receive multicast packets. If we are using multiple switches, one of them makes a note of that and coordinates with the other switches to build the distribution list.

Only one switch can be the master of the distribution list. When data for that multicast stream is sent, the switches can look at their distribution lists. The first switch realizes it only needs to send the stream to the second switch, and that second switch knows exactly which hosts have requested the stream. With multicast, the transmitting device only has to send the data once, and the only devices that receive it are the ones that specifically requested it.

It’s worth noting that the multicast stream IDs are actually IP addresses. This doesn’t mean that your device suddenly changed to a new IP range. These are just numbers that host subscribe to. Similar to how you can tune into a radio station or a television station, it’s worth noting that Dante does not require you to manage multicast IP addresses.

The only reason we go over this is if you’re on an enterprise network and the IT team asks you which multicast addresses are you using, you’ll know what they’re talking about. Multicast group IP addresses fall in the range of 2 2 4 0 0 0 through 2 3 9 2 55, 2 55, 2 55. To efficiently manage multicast traffic networks use something called IG MP snooping and an IG MP querier IG MP stands for Internet Group Management Protocol. Those will be covered in more depth in upcoming chapters, but briefly, IGMP snooping helps direct multicast traffic only to the ports that need it, reducing unnecessary bandwidth usage.

The IGMP query ensures that only devices that have explicitly joined a multicast group receive the multicast traffic. Most switches have IGMP snooping disabled by default, so if your switch does not support IGMP snooping, or if it is not enabled, your multicast traffic will behave like broadcast. Sending multicast traffic out all ports in your broadcast domain. This is what multicast distribution looks like without IGMP snooping, and this is what multicast distribution looks like with IGMP snooping.

The more multicast traffic you have on your network, the more likely you are to need IGMP snooping. When configuring IGMP snooping. You’ll want to enable it on all switches. However, there is also a feature called the IG MP query that you need to know about.

The IGMP query will be the device that keeps the master distribution list of where all the multicast subscriptions go. So while you want all switches to have IG MP snooping enabled, only one of them can be the IG MP querier. Finally, it’s helpful to know that there are multiple versions of IG MP snooping. Dante devices support IG MP snooping version two or version three.

If you’re on a network that has version three, you’re in great shape. If you’re sharing a network with another system that only supports IG MP snooping version two, no problem, Dante can accommodate that as well. Okay, so Dante can send audio and video flows as either unicast or multicast. In Dante controller, we use two icons to represent these two different types of flows.

The dot with the arrow indicates unicast, meaning you’re transmitting directly from one device to another. Multicast has an icon with a dot and a forked arrow indicating that the network is duplicating that flow for you, distributing it to two or more hosts who have requested it. This will become more important when you progress to level three, but it’s worth noting that uncast traffic can in fact cross a router to other networks. Multicast is also routable across subnets, but this is outside the scope of the Dante certification program and is not necessary for layer three Dante applications.

Another benefit of unicast is that it can achieve lower latency than multicast In Dante networks, unicast packets have the destination IP address in the packet, and as a result, it can traverse the switch more quickly. Multicast audio and video arrives with a multicast group IP address. The switch then needs to figure out who wants that packet, duplicating the packet for each path it must take. This extra work can start to overload your switches CPU.

So we have designed our multicast packets to place less of a load on the switch while of course allowing them more time to traverse the network. If you’re ever concerned about how much load your switch is taking on, managed switches will often offer a chart showing the CPU load. To summarize, we learned that unicast is a one-to-one communication method. It offers the best latency performance and is the easiest for your network to move even in layer three routed in environments.

When working with broadcast, keep these best practices in mind. First, try to avoid excessive broadcast. To reduce network congestion, use directed broadcasts only when necessary, and consider alternative methods like multicast. For more efficient network communication, limited broadcasts are local, non forwarded, and ideal for immediate network needs.

Directed broadcast while capable of reaching other networks requires careful handling due to security concerns and is not widely used because most applications use multicast. Instead, multicast efficiently distributes network traffic by delivering data to multiple recipients. Instead of duplicating data at the source, the switch handles traffic replication, ensuring that only necessary network segments receive the transmission. This optimizes bandwidth utilization and reduces network congestion.

If you’re using large amounts of multicast, IGMP snooping is required to ensure multicast only goes to the devices that have requested it. When you set it up, you must engage IGMP snooping on all switches, but remember, only one switch can be the IGMP query. Thanks for watching this chapter. Keep going as we cover some of these topics in more depth later in level two.

Perfect.