Seong-Lyong Gong

Opportunistic Scheduling and Adaptive Modulation in Wireless Networks with Network Coding

So far, many researches on network coding are performed with higher layer protocols such as MAC, routing, and flow control protocols without consideration of physical layer issues such as channel conditions of links. However, in wireless networks, the consideration of properties at the physical layer is important to improve system performance. Hence, in this paper, we study an opportunistic scheduling and adaptive modulation problem for wireless networks with network coding, which is a joint problem for MAC and physical layers. A similar problem was studied in considering an idealized system in which the data rate of each link is modeled with the Shannon capacity. They showed that to maximize the throughput of a transmission, the optimal subset of native packets that are encoded within a coded packet should be selected based on the channel condition at the destination for each native packet. Moreover, they also showed that it may not be the optimal selection to encode all possible native packets within a coded packet. In this paper, we consider a more realistic model than that of with practical modulation schemes such as M-PSK and MQAM. We show that the optimal policy that maximizes the throughput of a transmission is to encode all available native packets within a coded packet regardless of the channel condition at the destination for each native packet, which is a different conclusion from that of. However, we show that adaptive modulation, in which its constellation size in a coded packet is adjusted based on the channel condition of each destination node, provides a higher throughput than the scheme with fixed modulation, in which its constellation size is always fixed regardless of the channel condition at each destination node.

Frugal ECC: efficient and versatile memory error protection through fine-grained compression

Because main memory is vulnerable to errors and failures, large-scale systems and critical servers utilize error checking and correcting (ECC) mechanisms to meet their reliability requirements. We propose a novel mechanism, Frugal ECC (FECC), that combines ECC with fine-grained compression to provide versatile protection that can be both stronger and lower overhead than current schemes, without sacrificing performance. FECC compresses main memory at cache-block granularity, using any left over space to store ECC information. Compressed data and its ECC information are then frequently read with a single access even without redundant memory chips; insufficiently compressed blocks require additional storage and accesses. As examples, we present chipkill-correct ECCs on a non-ECC DIMM with x4 chips and the first true chipkill-correct ECC for x8 devices using an ECC DIMM. FECC relies on a new Coverage-oriented-Compression that we developed specifically for the modest compression needs of ECC and for floating-point data.

An average velocity-based routing protocol with low end-to-end delay for wireless sensor networks

To reduce energy consumption, in most MAC protocols for wireless sensor networks, listen-sleep cycles are adopted. However, even though it is a good solution for energy efficiency, it may introduce a large end-to-end delay due to sleep delay, since a node with a packet to transmit should wait until the next-hop node of the packet awakes. To resolve this issue, in this paper, we propose the average velocity-based routing (AVR) protocol for wireless sensor networks that aims at reducing the end-to-end delay. The AVR protocol is a kind of a geographic routing protocol that considers both location of a node and waiting time of a packet at the MAC layer. When a node can use information of n-hop away neighbor nodes, it calculates the n-hop average velocity for each of its one-hop neighbor nodes and forwards a packet to the neighbor node that has the highest n-hop average velocity. Simulation results show that as the knowledge range, n, increases, the average end-to-end delay decreases.

Cross-layer and end-to-end optimization for the integrated wireless and wireline network

In this paper, we study a cross-layer and end-to-end optimization problem for the integrated wireless and wireline network that consists of one wireline core network and multiple wireless access networks. We consider joint end-to-end flow control/distribution at the transport and network layers and opportunistic scheduling at the data link and physical layers. We formulate a single stochastic optimization problem and solve it by using a dual approach and a stochastic sub-gradient algorithm. The developed algorithm can be implemented in a distributed way, vertically among communication layers and horizontally among all entities in the network, clearly showing what should be done at each layer and each entity and what parameters should be exchanged between layers and between entities. Numerical results show that our cross-layer and end-to-end optimization approach provides more efficient resource allocation than the conventional layered and separated optimization approach.

DRAM Scaling Error Evaluation Model Using Various Retention Time

As process technology scales down, DRAMs become less reliable since current DRAM protection schemes cannot address scaling-induced errors effectively. In prior related work, faulty DRAM cells or inherent faults have been considered as permanent and easily detectable during manufacturing process. However, due to scaling, more errors occur intermittently during system operations and exacerbate DRAM reliability. To consider the increasing intermittent errors due to scaling, we propose a new error model using DRAM’s Variable Retention Time (VRT). Our model effectively describes latent inherent faults, which lead to intermittent errors, in addition to the permanent inherent faults. We also introduce an efficient methodology to accelerate simulations even with a myriad of inherent faults. We present parameter sweep results to provide better understanding or insight regarding DRAM scaling errors.

Joint optimization of link weight and link capacity expansion in communication networks

To satisfy a target utilization of each link in communication networks, re-routing of sessions and expansion of the link capacity could be considered. In this paper, we study a problem that finds joint optimal link weights for routing and capacity expansion so as to alleviate over-utilization of links in communication networks with minimum costs. We compare the performance of our joint algorithm with that of the algorithm that considers only link weight optimization [1], which shows that the joint optimization provides better performance than the link weight optimization only.

Link weight optimization for routing considering link utilization in communication networks

In this paper, we study the problem that finds optimal link weights for routing in communication networks with a target utilization for each link. Since the problem is known as NP-hard, we take a heuristic approach based on the simulated annealing method. However, to make the algorithm to be more robust and converge fast, we modify the basic simulated annealing method. Numerical results show that our algorithm provides a better performance than the basic simulated annealing method.

Joint optimization for integrated wireless and wireline networks

In this paper, we study a cross-layer optimization problem for a network that consists of one wireline core network and multiple wireless access networks. We consider four layers among five layers in the network layering architecture: application characteristics through the utility function and quality of service requirements in the application layer; end-to-end flow control in the transport layer; opportunistic scheduling in the data link layer; adaptive modulation and coding in the physical layer. We formulate a stochastic optimization problem considering above four layers and both wireline and wireless parts of the network jointly that results in a utility-based joint end-to-end flow control and opportunistic scheduling problem for the integrated wireless and wireline network. We solve the problem by using a dual approach and a stochastic sub-gradient algorithm. The developed algorithm can be implemented in a distributed way: vertically among four layers and horizontally among all entities in the network, clearly showing what should be done in each layer and each entity and what parameters should be exchanged between layers vertically and between network entities horizontally.