The design and analysis of control policies in multi-hop wireless networks has mainly focused on  
throughput and stability as measures of performance. While these metrics are important, the  
development of analytical techniques to study the delay performance is crucial to provide guarantees on  
the quality of service, network dimensioning (choice of buffer sizes, capacity of links), etc. However,  
analyzing the delay of throughput- optimal (queue-length based) scheduling policies in such systems is  
extremely difficult due to the complex correlations that arise between the arrival,   service, and the  
queue length processes. 
In this talk, I will describe a novel methodology to establish fundamental lower bound on the expected  
delay performance of the system. Our analysis is supported by extensive numerical studies to suggest  
that the lower bound captures the important elements (interference, multiplexing and fading) of a  
wireless network. For the tandem queue network, where the delay optimal policy is known, the  
expected delay of the optimal policy numerically coincides with the lower bound. I will also present  
sharper upper bounds for important scheduling policies such as the Maximum Weighted Matching  
(MWM) type of scheduling policies.