12.03.06
The case for making multicast a first-class service in the Internet
IP multicast is a must for the Internet. It provides many advantages over a plain unicast network. With demand growing toward more streaming media, IPTV, and video conferencing applications, multicast becomes more and more of a necessity. There are many advantages that multicast offers to these type of applications. The infrastructure for IP multicast already exists. It is supported by Ethernet and many multicast protocols are already implemented in routers.
There are many applications that would benefit from multicast technology including video conferencing, IPTV, real time audio, and IP telephone to name a few. These types of applications are growing in demand for one’s desire of being always connected. People want content on demand and they want it when they want it. Multicast provides an efficient way for people to receive the content they desire.
Multicast provides many advantages to the network that supports the success of the applications mentioned above. Besides the support of distributed applications, multicast also provides optimized network performance, resource reduction, scalability, and reduced network load. These benefits provide for a better working more efficient network. The cost is minimal since the current infrastructure already supports multicasting.
Multicast provides for an optimized network performance because of decreasing the amount of flow replications. Currently the network must execute several unicast statements, one for each flow, to achieve the same as one multicast statement. Multicast avoids these flow replications because of a better architecture to distribute the data. This will in turn save bandwidth that was wasted because of these unnecessary flow replications. As a result, new applications can be implemented without requiring a renovation to the network resources.
Scalability is another benefit of multicast. By effectively using the network resources by not wasting resources on unnecessary unicast statements, it reduces the traffic load of the sender. This also will allow applications and services to be accessed by many receivers. This will allow applications and services that use multicast to be easily dimensioned, so the sender can easily distribute these services to either a small or large group of receivers.
Multicast also frees up network resources. Optimally using the network resources with multicast will benefit the network as a whole by reducing traffic jams at the network equipment. This will allow more network resources to be available for additional applications and services and will allow the network to accomplish more without needing additional resources.
A free multicast backbone exists called MBONE. MBONE is a virtual network to connect multicast applications over the Internet. Its goal is to minimize the amount of data needed to provide multipoint audio and video conferencing. MBONE has been in existence since 1992 and was implemented because many routers did not support multicast. Now many routers do support multicast so MBONE is anticipated to become obsolete.
Ethernet and the Internet infrastructure already support multicast, so no major renovation to the network is needed. For example Cisco routers have supported multicast since 1995. They support multiple multicast routing protocols to accomplish things such as source discovery, multicast transport, multicast registration, multicast flooding, and source specific multicast. This names just a few of the multicast protocols supported by Cisco routers. The Microsoft server operating systems also support multicast. The point being that multicast is supported in both network hardware and software, but it is not supported at the ISP level as a service deployed to the entire Internet. One would be able to implement multicast in their respective Intranet, but multicast needs to be supported as a Internet service.
Multicast is standard on IPv6. It is not certain that IPv6 will ever be deployed, but multicast is to be a standard feature. The point is that IPv6, which is suppose to be the next version of IP is going to include multicast. They have identified that multicast is useful and many would benefit from it. IPv6 or whatever the next upgrade to IP is will most likely include support for multicast.
There has been much research in multicast and designing applications, services, and architectures that use multicasting. There have been architectures designed to support multicast with Mobile IP. Others have designed multicast support for commerce applications as well as distributed multimedia. This shows the demand for multicast is increasing and many would benefit from it. Again this shows that multicast should be implemented as a service to the everyday Internet.
Multicast benefits the network in many ways and should be added as a first-class service to the Internet. It has not been added because the ISPs make their profit on providing connections. Multicasting would essentially eliminate connection by the nature of the protocol. Instead ISPs should implement multicast as a service to those senders that would benefit. There are many benefits of multicast that would profit those who provide the distributed multimedia type applications. Many of those users would probably pay for the service because of the time and money it would save them. There is a need for multicast, but the ISPs would have to renovate their charging methods before they would implement it.
Multicast provides many advantages for the Internet and should be included as a first-class service. It allows for better resource allocation, optimized network performance, scalability, and support for distributed applications. Many applications would benefit from using multicast including video conferencing, streaming media, and IPTV. The provider of these applications would benefit the most because less resources would have to be allocated for the scheduling and processing of multiple unnecessary unicast operations.
Its not that multicast would be difficult to deploy either. Multicast has been implemented in the MBONE multicast backbone virtual network. It has also been implemented in many routers and server operating systems. Basically the current infrastructure already supports multicast and no renovation is necessary. The cost of implementing multicast as a service is minimal.
The major issue being multicast eliminates profit from the ISPs by eliminating routes. Instead of charging for routes, the ISPs could charge for the multicast service. Developing a charging system for multicast would cost the ISPs, but they would be able to make the money back by providing the service.
Multicast is an obvious must for the Internet. It provides many advantages for the users especially those using multimedia applications. More and more of these applications are being developed, which make multicast support even more of a must. Obviously there is a demand for the service. This can be seen from the research accomplished in this area along with the protocols developed and supported already.
traviskeshav said,
December 6, 2006 at 12:24 pm
I agree with a lot of your analysis in this essay. However, I’m not sure that we could reach a point where people would necessary accept paying more for something such as multicast that’s already part of the Internet architecture, rather than accepting the less-dynamic workarounds that exist now. It seems to me as if it’d be somewhat difficult to exactly figure out how to get ISPs to implement it in a manner that would be acceptable to both consumers and themselves. If the price is per multicast transmission, then it could be limiting as people would not want to use the service frequently for the risk of incurring a large bill. If the price is just to activate multicast for a certain user, then it could be abused by non-paying users sending their data to enabled users to multicast.
Perhaps I’ve become cynical over the course of this semester, but I don’t have much faith in ISPs investing their time and resources in finding a way to make it work.