 It is very important to understand that when two different networks converge, the kind of applications that they offer also undergo a transition. When the telecommunication networks are going to merge with the internet, then they are going to peer with each other. As a consequence, a peer-to-peer service architecture is going to result in two. In this module, we shall understand what peer-to-peer services are from the standpoint of self-organizing networks. Then we look at an interesting concept of service offering and service soliciting, specifically known as seeding and leaching. The peer-to-peer networks actually are created in the form of a logical overlay on a physical network. So to begin with, there is no centralized entity, so peer-to-peer networks are formed without a centralized server. This results into the requirement to self-organize, and the kind of applications now which would be offered on this peer-to-peer network would be distributed services. In the absence of a single point of failure, these applications are going to be highly scalable and robust because as the number of users grows, the users themselves are going to be the service providers as well as the service users. The peers are logical entities because it is nothing but an overlay network. So the peer with joins as a client can also serve as a server. When it comes to offering the services, each node that enters into the network offers its processing power, that is the CPU cycles, the data rate that it can offer on the internet, and the hard disk it can share with other users. So this is what a certain node can seed into the network. When such kind of resources are pooled together, then different tasks can be shared between these peers, and any service which is required by a certain peer can be taken out of the network, and this process is known as leaching. Let's look at some of the types of peer-to-peer services. Number one, if this activity of forming peers is going to be an organized activity, it would result into structured peer-to-peer networks. These networks would have well-defined algorithms, known topologies, and well-defined and well-formulated policies. Since it is a very organized activity, it can be used for offering very large-scale services. So it means in today's next generation networks, we can think about offering peer-to-peer services by utilizing and organizing the infrastructure as we please. A more tangible and observable kind of peer-to-peer network formation is the unstructured network. Here, there is no a priori specific structure. However, just to bring in order and to make the peer-to-peer network operational, the centralized indexing or bootstrapping of the clients is required, and that is mostly done by a centralized entity or multiple of these entities. In this unstructured peer-to-peer formation, there could be some variants to it. It could be a pure peer-to-peer in which there is no supra node or no master entity. It could be centralized peer-to-peer where we have centralized indexing and the information sharing between peers is done under the eye of the centralized entity. Then we could have a hybrid of these called the hybrid peer-to-peer. In such case, we'd have nodes which are relatively smarter than and more authoritative than other nodes. These nodes are known as super nodes. So a mix and match or a blend of all these can also be realized. Let's do figurative comparison between the peer-to-peer and client server architectures and how their services vary. As you can see in the diagram, in the case of a client server architecture, the servers are designated machines and their roles do not change very often. So the clients send dedicated requests to known servers and get the responses. So there is no prior lookup or searching activity required. But in the case of pure peer-to-peer, we see that each peer needs to undergo a discovery process to know the most relevant peer it needs to contact. And after having discovered it needs to send a dedicated request to that peer and open a connection. And then that particular connection can be used for sending queries and responses from either side. What we conclude from here is that between a strict client server model and between a pure peer-to-peer model, a hybrid of or a combination of these could result into any desired kind of peer-to-peer network formation.