 The content delivery networks are assumed to be mostly situated close to the core content delivery network architecture, primarily residing in the data center. But for new users who access these content delivery networks through mobile, they are more interested in getting services which are more akin to their and more mindful of their constraints in mobile environments. So the information dissemination or the object sharing distribution in mobile CDNs is an important concern. Here we are talking about mobile users and the content delivery network for mobile users is termed as mobile CDNs. So we'll take a formal definition We'll look at the architecture and some more details. So mobile CDNs are basically a surrogate service as you understand that the origin server once burdened needs some kind of support of surrogate service. Now these surrogate servers deliver the content in mobile network infrastructure, hence the name. So typical CDN architectures have to take into effect, take into account some mobility specific requirements. That is mobile networks have scarcity of resources, understandable, compare a mobile phone with a laptop or a desktop. Then the users are on the move. So it means some kind of ambulatory behavior of users is a limiting factor in the smooth provisioning of services in mobile environments. Then the user is interested in getting more location specific services and the network is not always supporting. So there are frequent flaky links and frequent disconnections. With this in view the architecture that could support mobility that is a mobile CDN has to take into account some servers, the surrogate servers, which will replicate the origin server content. Now these surrogate servers have to be located close to the telecommunication or the wireless infrastructure, including the mobile switching centers and the base station controllers. Now there has to be some kind of infrastructure including both wired and wireless to deliver content to optimal surrogate. The surrogates have been deployed, but now the mobile phone or the mobile user is in the proximity of certain surrogate. So the closest has to be accessed and some kind of feedback mechanism to measure and ensure quality of service should also be there. Then we also need to have a cashier manager for content replication across these surrogates and once these cashier managers have and cashed the content, the content location manager has to now facilitate and schedule data provisioning to the users which are in close proximity. This could be pulled, the data could be pulled or the data could be pushed either way. Then last but not the least an accounting mechanism for collecting logs and information so that the strategy for replication, caching, surrogate deployment and surrogate access mechanisms could be improved. Now let's look at a typical mobile content delivery network. On top right we have the origin server. This origin server is being supported through surrogates. These surrogates are distributed across Europe, Asia, US and these surrogates are providing localized coverage to different wired-com wireless infrastructures. Let's take an example of surrogate server in the United Kingdom. Here there's a cellular network say 4G LTE offering internet connectivity to the mobile users. So intuitively and by evidence it can be seen that the mobile users must get the content replicated in the surrogate server in the United Kingdom. A comparison can be made between what are typical CDNs in terms of their features and how do they differ from mobile CDNs. Mostly there are a lot of similarities, but once we look at a user location, it is fundamentally different. In mobile CDN it varies. The surrogate server placement has to be close to base stations versus the ISP in case of CDN. Then the replica maintenance cost becomes very high in mobile CDN. The services have to be location-based services and the user-oriented services because the user is using a limited form factor constraint device. The content outsourcing policy has to be in agreement with the origin server. So it has to be pushed-based because the surrogates are mostly on the receiving end. The mobile CDN architecture under centralized infrastructure is very simple to understand. We have various network providers, the cellular providers. Since it is a centralized infrastructure, there is a central CDN provider and it is connected to various cellular and Wi-Fi providers and is very easy to maintain a lot of consistency is there. But in case of ad hoc networks, the infrastructure becomes slightly more tricky. Some kind of consistency, replication, mirroring and presence of content issues arise because the connectivity is intermittent. So you see here, we have ad hoc network like mobile ad hoc network, which is the collection of mobile phones in ad hoc mode or the hotspot mode which are connecting to the content delivery network through each other. Then there is a vehicular ad hoc network vannet in which the vehicles with their built-in wireless connectivity talk to each other to reach out to another CDN provider. Now the proximity is a very important concern because the batteries of these wireless devices get drained or depleted rather very quickly and these devices in case of ad hoc mode have to do a lot of forwarding or relaying which otherwise was not a requirement in central wireless infrastructure. So you see a comparison can be made between the ad hoc deployment and the centralized deployment in terms of the complexity that would appear. This is from Raj Kumar bhaiya content delivery networks.