What Do You Really Know About Multicast in Local Area Networks?

 

How are high-bandwidth applications like IPTV and video able to be sent over any LAN environment to many recipients safely and optimally without causing havoc with network bandwidth?

Well, good design practices in networks for a start - and another is to use multicast.

The vast majority of systems that transport any sort of AV traffic over the IP network today use multicast as their preferred method of transport. (I say the vast majority as some, believe it or not, use other less efficient methods, such as broadcast….)

And they use multicast with good reason. It is the most efficient mechanism on the network available for transporting information (especially large-bandwidth data) from one source to many receivers. Unlike non-real time traffic such data which is transported as unicast, the network has to treat high-bandwidth, real-time traffic like video that needs to be received by a number of endpoints simultaneously, differently. It uses multicast to do this.

In this post we cover some multicast fundamentals and find out why it is so important to AV applications over the IP network such as IPTV.

 

Some Multicast Basics.

So what is IP multicasting? In a sentence, it is a bandwidth-saving technology that reduces network traffic by simultaneously delivering a single stream of time-sensitive data such video to (up to) thousands of recipients.

This can be achieved because the IP multicast function sends the source traffic to numerous receivers by replicating the multicast packets in the network, thus avoiding sending the same stream across the network incrementally – sending many copies of the data. This not only uses minimal network bandwidth, but also helps lighten the load on both source and receiver devices.  Therefore, high-bandwidth traffic such as video (whether compressed, for example MPEG-4 or uncompressed) can be sent efficiently and optimally over the network.

In the ‘How Multicast Works’ example shown below, streams (the coloured arrows) are only sent to the receivers that request them. Not all the streams from the source devices (AVoIP encoders) are sent to all recipients. This is achieved in the main by using the IGMP (Internet Group Management Protocol.)

Receivers send IGMP messages (reports) and to the sources via the network in order to join multicast groups…and therefore receive the multicast streams.

The IGMP Group Idea

So IP multicast is based on the idea of groups. A multicast group is any number of receivers that are receiving the same source information. If they want to receive another stream, so ‘watch’ another channel in real terms, then they need to join another group that that stream is in. To do this they need to leave the current group they are in, and request to join another group. In this way, only streams that are requested are sent out across the network to the requesting hosts.

IGMP groups are not bound by area or location or any physical restrictions – these devices can be located anywhere on the network.

So the use of the IGMP protocol is highly recommended as it prevents hosts on a local network from receiving traffic from a multicast group they have not explicitly joined. This prevents ‘flooding’ of multicast traffic to all ports on a broadcast domain (or the VLAN equivalent.) It is therefore especially useful for bandwidth-intensive application such streaming video. 

 

A Word About Multicast IP Addressing

There is a special range of IP addresses that the Internet Assigned Numbers Authority (IANA) have allocated for multicast – the IPv4 Class D address space. This range is between 224.0.0.0 and 239.255.255.255. Any address in this range identifies any number of hosts that have joined a group – essentially any hosts that want to receive multicast traffic assigned for that particular group.

However the addresses 224.0.0. through to 224.0.0.255 are reserved for local-link addressing, and so should not be forwarded by routers or other network devices. Routing protocols also use this address space for updates – for example OSPF uses 224.0.0.5 and EIRGP uses address 224.0.0.10.

Although the rest of the range, so 224.0.1.0 to 239.255.255.255 is classed as globally scoped and can be used for general multicast traffic, there are a few exceptions. Certain address have been reserved for use only for certain functions – for example, 224.0.1.1 is reserved for Network Time Protocol (NTP)

So unlike the other classes of IP address that specify unicast IP addresses that are ‘device specific’ to network hosts, Class D space addressing is used to indicate the group address or the destination address of IP multicast traffic.

 

In Conclusion

And there you have it – the basics of IP Multicasting in a few paragraphs. Think of it simply of just a way of sending IP traffic (datagrams) to a group of interested parties (receivers) in a single transmission. Using IP multicast allows you to employ a point-to-multipoint connectivity strategy that allows a single sender’s transmission to be received by numerous receivers simultaneously.

But before you even start implementing any multicast capability on your network, be aware that not all network switches are created equal.

You not only need to ensure that the network switch you select supports multicast (most support IGMP do these days) but also - and perhaps more importantly - to ensure that it is configurable for multicast provision. Many so called ‘Web-Smart’ switches have certain configurable parameters, but multicast is usually not one of them.

A fully managed, multicast capable switch is the best way to go to implement IP multicast and optimise your network for high-bandwidth traffic.