Gireesh Shrimali

Cooperative interdomain traffic engineering using Nash bargaining and decomposition

We present a novel approach to interdomain traffic engineering based on the concepts of Nash bargaining and dual decomposition. Under this scheme, ISPs use an iterative procedure to jointly optimize a social cost function, referred to as the Nash product. We show that the global optimization problem can be separated into subproblems by introducing appropriate shadow prices on the interdomain flows. These subproblems can then be solved independently and in a decentralized manner by the individual ISPs. Our approach does not require the ISPs to share any sensitive internal information, such as network topology or link weights. More importantly, our approach is provably Pareto-efficient and fair. Therefore, we believe that our approach is highly amenable to adoption by ISPs when compared to past approaches. We also conduct simulation studies of our approach over several real ISP topologies. Our evaluation shows that the approach converges quickly, offers equitable performance improvements to ISPs, is significantly better than unilateral approaches (e.g., hot-potato routing) and offers the same performance as a centralized solution with full knowledge.

Cooperative Inter-Domain Traffic Engineering Using Nash Bargaining and Decomposition

We present a new inter-domain traffic engineering protocol based on the concepts of Nash bargaining and dual decomposition. Under this scheme, ISPs use an iterative procedure to jointly optimize a social cost function, referred to as the Nash product. We show that the global optimization problem can be separated into sub-problems by introducing appropriate shadow prices on the inter-domain flows. These sub-problems can then be solved independently and in a decentralized manner by the individual ISPs. Our approach does not require the ISPs to share any sensitive internal information (such as network topology or link weights). More importantly, our approach is provably Pareto-efficient and fair. Therefore, we believe that our approach is highly amenable to adoption by ISPs when compared to past naive approaches. We conduct simulation studies of our approach over several real ISP topologies. Our evaluation shows that the approach converges quickly, offers equitable performance improvements to ISPs, is significantly better than unilateral approaches (e.g. hot potato routing) and offers the same performance as a centralized solution with full knowledge.

Optimal Feed-in Tariff Schedules

We analyze the design of optimal feed-in tariff schedules under production-based learning. We examine least cost policies in a simple two-period model that focuses on bringing down the levelized cost of renewable technologies to a predefined target under two well-known dynamics: learning-by-doing (LBD) and economies of scale (EOS). We show that, when the levelized cost reduction target is stringent, subsidies are required in both periods, regardless of the dynamics. However, when the target is moderate, the optimal policy is to subsidize only in one of the two periods: under the LBD dynamics, it is optimal to subsidize as early as possible, whereas under the EOS dynamics, it is optimal to subsidize as late as possible. Under the LBD dynamics the prevailing factor is the impact of early investment on cumulative experience, whereas under the EOS dynamics the prevailing factor is capital depreciation. The key takeaway is that, based on the underlying dynamics, the policy maker needs to adopt fundamentally different kinds of policies to promote renewable technologies.

Have State Renewable Portfolio Standards Really Worked? Synthesizing Past Policy Assessments to Build an Integrated Econometric Analysis of RPS effectiveness in the U.S.

Renewable portfolio standards (RPS) are the most popular U.S. state-level policy tools for promoting deployment of electricity from renewable energy (RES-E). While several econometric studies have estimated the effect of RPS policies on RES-E deployment, results are sometimes contradictory. Thus, there is a need to not only reconcile these studies but also establish a definitive answer to the question whether RPS policies have been effective in increasing RES-E deployment. We use a panel – 50 states over 1990-2010 – of renewable deployment, policy design elements and market dynamics to capture the heterogeneity in state RPS design and state characteristics. We run time series cross sectional regressions with fixed effects, and verify the causal effect implied by our results with a nonparametric matching analysis. We trace discrepancies among past findings to differences in data classification and model specification. We show that the key feature of RPS policies – RPS stringency – has no effect, and in some cases a negative effect, on RES-E deployment. We also investigate the impact of individual RPS features. We show that renewable energy certificate (REC) unbundling as well as the presence of RPS in neighboring states is correlated with greater RES-E deployment locally, while other features – such as allowing greater flexibility to buy renewable energy from out-of-state – slows in-state RES-E deployment. We show that mandatory green power options have a positive effect on RES-E deployment while net metering schemes appear to have a negative impact. However, we were unable to detect any impact of an interaction effect between RPS and these policies. Finally, we investigate the impact of policies on the deployment of individual renewable technologies (biomass, geothermal, solar, and wind) and demonstrate differentiated policy impacts, such as biomass driving the result on RPS stringency.

Designing packet buffers with statistical guarantees

Packet buffers are an essential part of routers. In high-end routers, these buffers need to store a large amount of data at very high speeds. To satisfy these requirements, we need a memory with the the speed of SRAM and the density of DRAM. A typical solution is to use hybrid packet buffers built from a combination of SRAM and DRAM, where the SRAM holds the heads and tails of per-flow packet FIFOs and the DRAM is used for bulk storage. The main challenge then is to minimize the size of the SRAM while providing reasonable performance guarantees. We analyze a commonly used hybrid architecture from a statistical perspective, and investigate how small the SRAM can get if the packet buffer designer is willing to tolerate a certain drop probability. We introduce an analytical model to represent the SRAM buffer occupancy, and derive drop probabilities as a function of SRAM size under a wide range of statistical traffic patterns. By our analysis, we show that, for low drop probability, the required SRAM size is proportional to the number of flows.

Can bill-and-keep peering be mutually beneficial?

We analyze “Bill-and-Keep” peering between two providers, where no money exchanges hands. We assume that each provider incurs costs from its traffic traversing its as well as the peer’s links, and compute the traffic levels in Nash equilibrium. We show that Nash strategies are not blind, i.e., they are neither pure hot-potato nor pure cold-potato strategies. Rather, the Nash strategies involve strategically splitting traffic between a provider’s own links and its peer’s. We derive necessary and sufficient conditions for both the providers to be better (or worse) off in Nash equilibrium compared to the blind strategies. We also analyze society’s performance as a whole and derive necessary and sufficient conditions for the society to be better (or worse) off. In particular we establish that, under Bill-and-Keep peering, while it is not possible for two asymmetric providers to be both worse off, it is certainly possible for both to be better off.

The Impact of State Policy on Deployment and Cost of Solar PV: A Sector-Specific Empirical Analysis

Using a panel database for 27 programs in 16 U.S. states over 1998-2009, we assess the impact of 12 state-level policies on the cost and deployment of solar photovoltaic (PV) technologies for two sectors defined by system sizes: residential (

India's Energy Data: The Urgent Need to Move from Patchwork Systems to a National Energy Information Agency

We propose that India should create a national Energy Information Agency – an Indian EIA or “indEIA.” India urgently needs a dedicated, central agency to collect, collate, disseminate, and facilitate the analysis of all essential energy-related data. There are a number of government support programs, for everything from renewable energy to electricity access, in the billions of dollars. Their efficacy is rarely understood, in part because of data limitations. An entity like indEIA will be critical in helping India leverage the creative powers of the national and international research community to provide reliable, cost-effective, and clean energy to its citizens. In addition, granular data can enable the industry as well as policymakers to move towards more customized, dynamic, and nimble solutions, instead of relying on averaged, aggregated, and nongranular time-series or spatial data. Thus our proposal envisions indEIA as the primary vehicle for curating and maintaining India’s energy data, a vital national (and global) asset.

Introduction of Software Products and Services Through "Public" Beta Launches

Public “Beta” launches have become a preferred route of entry into the markets for new software products and web site based services. While beta testing of novel products is nothing new, typically such tests were done by experts within firm boundaries. What makes public beta testing so attractive to firms? By introducing semi-completed products in the market, the firm can target the early adopter population, who can then build the potential market through the word of mouth effect by the time the actual version of the product is launched. In addition, the information gathered through the usage of the public beta gives significant insights into customer preferences and consequently helps in building a better product. We build these marketing and product development implications in an analytical model to compare the different product introduction strategies like “skimming” or “penetration pricing” with beta launches. This analysis is done for products of branded and unbranded Web 2.0 companies like Google and Flickr etc. We also examine the impact of different monetization models like direct pricing and advertising on the beta launch strategy.