Browsing by Subject "congestion"
Now showing 1 - 2 of 2
Results Per Page
Sort Options
Item Detecting Tangled Logic Structures in VLSI Netlists(2010-10-12) Jindal, TanujThis thesis proposes a new problem of identifying large and tangled logic structures in a synthesized netlist. Large groups of cells that are highly interconnected to each other can often create potential routing hotspots that require special placement constraints. They can also indicate problematic clumps of logic that either require resynthesis to reduce wiring demand or specialized datapath placement. At a glance, this formulation appears similar to conventional circuit clustering, but there are two important distinctions. First, we are interested in finding large groups of cells that represent entire logic structures like adders and decoders, as opposed to clusters with only a handful of cells. Second, we seek to pull out only the structures of interest, instead of assigning every cell to a cluster to reduce problem complexity. This work proposes new metrics for detecting structures based on Rent?s rule that, unlike traditional cluster metrics, are able to fairly differentiate between large and small groups of cells. Next, we demonstrate how these metrics can be applied to identify structures in a netlist. Finally, our experiments demonstrate the ability to predict and alleviate routing hotspots on a real industry design using our metrics and method.Item Distributed Control Approaches to Network Optimization(2010-07-14) Sah, SankalpThe objective of this research is to develop distributed approaches to optimizing network traffic. Two problems are studied, which include exploiting social networks in routing packets (coupons) to desired network nodes (users in the social network), and developing a rate based transport protocol, which will guarantee that all the flows in a network (e.g. Internet) meet a delay constraint per packet. Firstly, we will study social networks as a means of obtaining information about a system. They are increasingly seen as a means of obtaining awareness of user preferences. Such awareness could be used to target goods and services at them. We consider a general user model, wherein users could buy different numbers of goods at a marked and at a discounted price. Our first objective is to learn which users would be interested in a particular good. Second, we would like to know how much to discount these users such that the entire demand is realized, but not so much that profits are decreased. We develop algorithms for multihop forwarding of such discount coupons over an online social network, in which users forward coupons to each other in return for a reward. Coupling this idea with the implicit learning associated with backpressure routing (originally developed for multihop wireless networks), we would like to demonstrate how to realize optimal revenue. We will then propose a simpler heuristic algorithm and try to show, using simulations, that its performance approaches that of backpressure routing. As the second problem, we look at the traditional formulation of the total value of information transfer, which is a multi-commodity flow problem. Here, each data source is seen as generating a commodity along a fixed route, and the objective is to maximize the total system throughput under some concept of fairness, subject to capacity constraints of the links used. This problem is well studied under the framework of network utility maximization and has led to several different distributed congestion control schemes. However, this idea of value does not capture the fact that flows might associate value, not just with throughput, but with link-quality metrics such as packet delay, jitter and so on. The traditional congestion control problem is redefined to include individual source preferences. It is assumed that degradation in link quality seen by a flow adds up on the links it traverses, and the total utility is maximized in such a way that the quality degradation seen by each source is bounded by a value that it declares. Decoupling source-dissatisfaction and link-degradation through an ?effective capacity? variable, a distributed and provably optimal resource allocation algorithm is designed, to maximize system utility subject to these quality constraints. The applicability of our controller in different situations is illustrated, and results are supported through numerical examples.