Scalable Multicast Data Streaming over Wireless Networks
In recent years, numerous large-scale
Wireless LANs (WLAN) have been deployed all over the world. However, the shortage of non-interfering
channels makes it a challenge for WLANs to efficiently support real-time multicast services. We have studied
the problem of association control in the case of multicast flows in order to optimize different objective functions
such as maximizing the number of accepted flows, balancing the multicast load among the access-points (APs)
and reducing the total load on all APs. For these problems we have designed approximation algorithms
and evaluated them using simulations. We have also studied the problem of efficient packet-by-packet scheduling
of real-time multicast flows. For mitigating interference, we allow APs to transmit simultaneously only
if they are mutually non-interfering and our objective is minimizing the fraction of time used by the APs for
servicing the multicast flows. We introduce two multicast strategies, the association strategy in which each user
is restricted to receive flows only from its associated AP and the non-association strategy in which a user may
also decode transmissions from other APs in its vicinity. Under both the strategies, the scheduling problem
of minimizing the multicast service time is NP-hard and we have developed simple approximation algorithms
with provable performance bounds. Our simulations clearly demonstrate that the proposed algorithms yield
efficient multicast scheduling.
In case of mesh networks the optimization must also consider the traffic in the backbone wireless network.
We have proposed a new metric that considers the overall network load which includes the traffic in the
backbone as well as the traffic on the last wireless hop to the user. By associating based on this metric rather
than RSSI (Received Signal Strength Indicator), we have shown that significant reduction in interference can
be obtained.
Most Relevant Papers