 In a capitalist based Global society It is the economics which matters. There's a well-known phrase money makes the mayor go We are going to look at the economics Side of the content delivery networks. So the design considerations are now going to be Economics informed or finance informed. We'll start off with an interesting example of Pasture with some herdsmen with their own sheep. It's known as the tragedy of commons It's the most well-known example in gaming theory Then we'd look at the incentives that would be expected by the CDNs from the content providers Then we'd look at the game theory as in network So what is tragedy of commons? Imagine a pasture that accommodates certain fixed number of sheep and this is a common good that serves the Community of certain herdsmen on an individual level each herdsmen or a shepherd is Interested in maximizing personal interest more sheep or wool more meat more money and one odd herdsmen or a shepherd a shepherd could imagine and Increase certain sheep or cattle and it would be all fine But there's a problem. The problem is all are human rational thinkers. So all herdsmen eventually come to the same conclusion So this results into a large increase in sheep When there are so many sheep The Cross that the sheep consume is depleted and all sheep die off starvation eventually this actually means that the problem of Not accommodating the global good versus individual gain results into this disaster coming to the networks paradigm If we look at the same issue of each individual's incentives or Selfish interests are at odds with the societal gain. This can be termed as an optimization problem or an incentive misalignment issue Let's look at an example. It's from a certain research paper that 70% of Nutella file-sharing system They were actually leeches or freeriders. They didn't contribute any data to the network and As little as 1% of the Nutella file-sharing community was serving 50% of all the requests Now this is against the very core spirit of Peer-to-peer file-sharing systems This means that we have to somehow come at common terms Through which we can encourage Individuals to behave more responsibly so that the individual interests are met while maximizing the overall Incentives or the profit of the community Specifically for the content delivery networks. Let's look at possible incentivization from the Individual CDN provider from the content provider or someone who is a user and wants to Participate as as as a helper in the CDN community Let's look at the participants incentives from from the CDN designer perspective a CDN designer or a CDN itself would be interested in Incorporating or Involving the end users as helpers for let's say content distribution Something which is very much similar to peer-to-peer architecture you could rethink the YouTube central servers now getting distributed as peer-to-peer content providers in the global community Similarly, the CDNs could reach out to each other as Though they are competitors for the same market, but they could join hands for the infrastructure sharing and coverage of a certain demographic and Another possible incentive or a desirable goal for a content provider is to Identify the infrastructure that is part of the CDN which is most loaded And which would become the hotspot eventually becoming a failure point. Now, these are the incentives which have to be addressed for CDN how such Optimization or trade-off be achieved one possible way to look at it is in terms of a Typical game like tic-tac-toe So what is game theory game theory is a theoretical approach towards studying the interactions between network elements as a game in which the Participating entities are not cooperative their rival and they take decisions which have certain cost for at the network level So this game theory provides us the foundation and one of the most well-known Approach approaches is the Nash solution or the Nash equilibrium. It says that a certain Solution or a certain balance between non cooperating entities could be carved out and a certain equilibrium stage could be reached where The individual user behavior could be predicted in how best or how far a user would go to achieve its own interests and infer the eventual outcome of the game The game is not a user versus another user but the overall consumption of the resources to Maximize certain objective functions that the entire network community is about so we can think about Social optimum or the social best that could be achieved by maximizing the Individual good or the individual role of all Users who are actually maximizing their own utility function The basic introduction of game theory is meant to address certain CDN related issues the best and most well researched areas in application of game theory for CDN include traffic routing traffic routing as As we know as the shortest path algorithms like Dijkstra's or Bellman Ford or for Autonomous systems in the form of border gateway protocol which works on Hot portrait or outing or any other optimization function This is traditional traffic routing problem. However, game theory looks at it from a different perspective There is a well-known paradox known as presses Paradox that actually has certain findings which are counter-intuitive It says that in increasing the network capacity By adding a new route Or a link to a path May adversely affect the performance of the entire network It means the congestion could actually increase Now the problem is further Exhibited if we look at how bad the situation could could get once the users are allowed to freely choose their traffic path This is with regards to traffic routing now for another issue in CDNs that is a caching replication is To find out where exactly data should be hosted temporarily or Permanent to satisfy certain objective function It means we are trying to look at increasing the area areas or the Locations where data could be stored at the same time We do not want certain existing players Which are storing data to exit or to defect from the network. So this is some kind of trade-off requiring problem Then we have the network formation network formation as we know is Is the formation of the? topology or the graph now This could be at any level it could be at the physical layer it could be at the data link layer It could be at the network layer or it could be at the application layer like an overlay So the the the concern is what are going to be? The new links for an incoming player that wants to establish its links to the existing Network elements So it means the network is now going to be extended Now with that extension coming from a new node would some of the existing edges be deleted When the new node joins in Similarly a new node If it in has to increase its own utility function Will it have to always increase the links or it might consider Decreasing its own links to maximize its own utility Now these are the important areas that require the application of game theory that is basically a probabilistic model to look at how the Economics of CDNs could be improved. We are referring to the same book thus far