Spectrum Leasing via Cooperative Opportunistic Routing Davide Chiarotto * , Osvaldo Simeone and Michele Zorzi *‡ * Department of Information Engineering, University of Padova, Italy CWCSPR, New Jersey Institute of Technology, New Jersey, USA California Institute of Telecommunications and Information Technology, UC San Diego, USA Email:{dchiarot,zorzi}@dei.unipd.it,osvaldo.simeone@njit.edu Abstract—A spectrum leasing mechanism is proposed for the coexistence between a primary and a secondary network that is based on cooperation and opportunistic routing. The primary network consists of a source and a destination communicating via a number of primary relay nodes. In each transmission slot, the next hop is selected in an on-line fashion based on the decoding outcomes in the previous transmissions according to the idea of opportunistic routing. The secondary nodes may serve as potential next hops for the primary network, but only in exchange for leasing of spectral resources so as to satisfy secondary quality- of-service constraints. Four policies based on spectrum leasing via opportunistic routing are proposed that provide different trade- offs between gains in throughput and overall energy expenditure for the primary network. Analysis is carried out for networks with a linear geometry and quasi-static Rayleigh fading statistics by using Markov chain tools. Index Terms—Cognitive radio networks, property-rights, spec- trum leasing, cooperative transmission, opportunistic routing I. I NTRODUCTION Regulation of the coexistence of primary and secondary networks is considered to be a key issue in the design of future wireless systems. Among the proposed paradigms, the approach proposed in [1], [2] circumvents the problems associated with the so called commons model (i.e., sensing and receiver-centric interference [3], [4]) and standard spectrum leasing (i.e., pricing [5]) by leveraging the idea of spectrum leasing via cooperation. In exchange for cooperation, primary users lease part of their spectral resource to secondary users, that in turn accept to cooperate if their desired Quality-of- Service (QoS) requirements are satisfied. The cooperation between primary and secondary users in a multihop scenario is done by means of opportunistic routing, which aims at increasing the throughput of multihop networks over fading channels by exploiting the channel diversity of- fered by the availability of multiple possible next hops. In particular, selection of the next hop is made in an opportunistic fashion based on the channel conditions of previous transmis- sions of a given packet, thanks to appropriate feedback from the decoders [6], [7]. This paper aims to improve the performance of the primary network in terms of throughput and energy expenditure by using secondary nodes as potential next hops in an oppor- tunistic fashion. In order to exploit the additional diversity provided by the secondary users via spectrum leasing, four This work was partially supported by the U. S. National Science Foundation under Grant # CCF-0914899, and by the European Commission under the MEDIEVAL project (grant agreement no FP7-258053). Fig. 1. A primary linear multihop network (grey circles) with k hops and a secondary network (white circles) aligned with respect to primary relay nodes. routing schemes are proposed to offer different gains in terms of primary throughput and energy consumption. II. SYSTEM MODEL We consider the system sketched in Fig. 1, in which a primary and secondary network coexist via spectrum leasing. The primary source P 0 wishes to communicate with the primary destination P k , at a normalized distance of one, possibly via multihop routing. There are two sets of additional nodes, which are placed along two parallel linear geometries with vertical distance Δ V . The first set is composed of k 1 primary nodes P 1 ,...,P k-1 whose only role is that of forwarding information from P 0 to P k and they are equally spaced with inter-node distance Δ H =1/k. The second set of nodes consists of secondary (unlicensed) nodes S 1 ,...,S k-1 , aligned with the primary, and thus with the same inter-node distance. We will also consider a partial secondary deployment in which only one every α secondary nodes in Fig. 1 is active so that the number of secondary nodes is k/α 1 (assumed to be an integer) with inter-node distance αΔ H . For simplicity, where not stated otherwise, we will assume α =1 in the following. Secondary nodes access the channel only if spectrum is leased by the primary network, as will be discussed below. All devices considered work in half-duplex mode (i.e., they cannot receive and transmit at the same time) and transmission is organized in slots, each corresponding to the transmission of a packet. The process starts when the source transmits a packet with transmission rate R bits/s/Hz in the first slot. In the following ones, retransmissions take place, if necessary, according to a Type-I HARQ process (i.e., retransmissions are not combined at the destination). Retransmissions in each slot may be performed by the source P 0 , or by the primary or the secondary relays, as long as the latter have correctly decoded in the previous slot. After the packet is correctly delivered to the destination, the primary source transmits a new packet and the process repeats.