Wireless CommunicationsIt is well known that interference between devices that communicate over the same spectrum is the primary bottleneck of data speeds in modern wireless communication systems. Much of my research in wireless communications has involved the study of interference alignment - a promising technique to manages interference in wireless communication systems. In my doctoral work, I developed and studied interference alignment for various communication scenarios. In my postdoctoral stint, I am exploring the use of lattice codes for wireless systems, in particular, their usefulness for interference alignment. To learn more about interference alignment, click here. Some sample publications are listed below.
Viveck R Cadambe, Syed A Jafar, Interference Alignment on the Degrees of Freedom of the K user Interference Channel, IEEE Transactions on Information Theory, Aug 2008.
Krishna Gomadam, Viveck R Cadambe, Syed A Jafar, A Distributed Numerical Approach to Interference Alignment and Applications to Wireless Interference Networks, IEEE Transactions on Information Theory, Jun, 2011
Vasilis Ntranos, Viveck R. Cadambe, Bobak Nazer, Giuseppe Caire Integer-Forcing Interference Alignment, IEEE International Symposium on Information Theory (ISIT), Jun, 2013
Distributed Storage SystemsIn modern data storage systems such as data centers, the data is stored in a distributed storage system with multiple servers. In such systems, it is important that the data is available to the user, even though the system components could be unreliable, specifically that the servers could fail. Classical coding theory provides several powerful techniques to design such failure tolerance with a small overhead in terms of storage footprint. In modern data storage systems, there are often additional requirements in several applications that lead to interesting research problems. I have done some work in this domain that I summarize below.
Repair Efficient Erasure Codes
It is of interest in storage systems that when a server has failed, or is unavailable for some other reason (e.g., when it is away for an upgrade), the data be reconstructed as efficiently as possible from the other available servers. In my research, I have studied code design for reconstruction of data by incurring a small cost of repair, both in terms of access and bandwidth. Some related publications are listed.
Viveck R. Cadambe, Syed A. Jafar, Hamed Maleki, Kannan Ramchandran, Changho Suh, Asymptotic Interference Alignment for Optimal Repair of MDS Codes in Distributed Data Storage, IEEE Transactions on Information Theory, May 2013
Viveck R. Cadambe, Arya Mazumdar An Upper Bound on the Size of Locally Recoverable Codes, IEEE International Symposium on Network Coding (Netcod), Jun, 2013.
Coding for Distributed Computing Theory
Recently, I have studied a problem that lies in the intersection of erasure coding and distributed computing theory. When a distributed storage system is used as a shared memory for applications of distributed computing, it is important that the data should appear to the users as if stored on a single centralized system even though the data is stored in distributed servers, i.e., the storage system should be atomic. I have recently designed efficient atomic shared memory algorithms that are based on erasure coding, and incur relatively smaller communication and storage costs than their replication based counterparts. The following technical report provides details.