Incentives - Peer-to-Peer Computing: Applications, Architecture, Protocols, and Challenges - page 126

to guard against such potential malicious attacks to the incentive mechanisms.
In particular, whitewashing is widely considered as a very low cost technique
for a selfish or malicious user to work around the incentive scheme.
Obviously, there is plenty of room for future research about incentive mech-
anisms in P2P sharing environments. Most notably, revenue maximizing [Ye-
ung and Kwok, 2006b] in a hybrid P2P system (e.g., the so-called converged
wireless architecture where an infrastructure-based cellular network is tightly
coupled with P2P WLANs) is of a high practical interest because there are
more and more cellular subscribers trying to share their resources without the
intervention of the cellular service provider. In a data sharing environment,
server peer selection [Leung and Kwok, 2005a] is another important direction
because we believe that users care more about the quality of service achieved
than about the revenue or cost they incur in participation. In economics terms,
people, especially wired or wireless game players, are quite inelastic about the
costs. Nevertheless, energy conservation [Leung and Kwok, 2005c, Leung and
Kwok, 2005b] is still of a prime concern in any wireless P2P sharing network
because energy depletion cannot be compensated in any way by increased
revenue generated in a payment-based sharing system. Thus, perhaps in a
game theoretic setting, we should incorporate energy expenditure in the utility
function. Finally, topology control [Leung and Kwok, 2005d] in a wired (over-
lay) network or wireless (ad hoc) network is also important in the sense that
sharing is usually interest based, meaning that users naturally form clusters
with similar interests, and as such, related users would be more cooperative
in following the incentive protocols. Consequently, building an interest-based
sharing topology could be helpful in enhancing the effectiveness of sharing.
5.5 Case Study: PPLive
As in many contemporary practical P2P applications, apparently PPLive
does not incorporate any systematic incentive mechanism to promote peer
contributions that can possibly lead to an optimized overall performance.
Specifically, based on performance studies reported recently [Piatek et al.,
2010, Horvath et al., 2008, Vu et al., 2010], it is observed that even a tit-for-
tat-like mechanism as that used in BitTorrent clients is not implemented in
PPLive. The only premise that such a lack-of-incentive approach can be relied
upon is the proprietary nature of the PPLive client programs. Indeed, it is not
an exaggeration to say that PPLive is a centralized software system from an
implementation point of view because users currently have no control over the
client programs' behaviors. Nevertheless, such a situation is bound to change
in the near future. More importantly, it is highly probable that system per-
formance can be much enhanced if proper incentive mechanism is in place, as
evident by the fact that peers switch channels very often and participation in
multiple overlays is far from coordinated.