The Internet is composed of tens of thousands of Autonomous Systems (ASes) which interconnect with one another through customer-provider (transit) and settlement-free peering links. Their interconnection objectives are a function of their business type e.g. transit providers, content providers, enterprise customers etc. In order to achieve these objectives the ASes adopt a set of criteria which are used to assess potential and existing settlement-free peering relationships. These relationships are bilateral in nature, i.e., for two ASes x and y to establish a peering relationship, they must satisfy each other’s peering criteria. The set of criteria used by an AS for assessing potential peering relationships is referred to as its peering policy. We commonly observe three peering policies publicized by ASes in PeeringDB [1], an open portal where ASes voluntarily share information about their peering policies, – Restrictive, Selective and Open. The peering strategy adoption by ASes of different categories taken from a recent snapshot of PeeringDB [1] showed that Open is the dominant peering strategy among all AS categories, with more than 60% of ASes in each category using Open peering. The fact that 64% of NSPs (transit providers) use open peering is especially surprising. Transit providers prefer other ASes as their customers rather than peers, and so the fact that many transit providers use open peering is rather surprising. Why do transit providers tend to peer openly? In our previous work [2] we use agent-based computational modeling to study peering strategy adoption by transit providers. Our computational model incorporates most of the real world constraints e.g. geographic colocation, skewed distribution of traffic, economies of scale, multiple transit prices per AS etc. We employed computational modeling as incorporating all these constraints in analytical model quickly renders the model intractable. An extended version of [2] is under progress and is available at [3]. In these works we find that peering decisions are interdependent and myopic decisions and lack of coordination among ASes results in Open peering as an attractor among peering strategies for transit providers. Interestingly, we observe that this adoption of Open peering results in loss of economic fitness for a majority of transit providers. Further, large scale adoption of Open peering results in stable equilibria. In this paper we use game theoretic analysis to gain further insight into peering strategy adoption by transit providers in the Internet. We employ a much simplified variant of our computational model to keep the analytical approach tractable. Analytical and numerical results from our current work corroborate our previous simulation based results. Our results show that when transit providers have complete information about co-located providers they can optimize their economic fitness by adopting Selective peering strategy [1]. Further, any uncertainty in the system causes the providers to gravitate towards Open peering, a suboptimal equilibrium.