Ramya Raghavendra

No "power" struggles: coordinated multi-level power management for the data center

Power delivery, electricity consumption, and heat management are becoming key challenges in data center environments. Several past solutions have individually evaluated different techniques to address separate aspects of this problem, in hardware and software, and at local and global levels. Unfortunately, there has been no corresponding work on coordinating all these solutions. In the absence of such coordination, these solutions are likely to interfere with one another, in unpredictable (and potentially dangerous) ways. This paper seeks to address this problem. We make two key contributions. First, we propose and validate a power management solution that coordinates different individual approaches. Using simulations based on 180 server traces from nine different real-world enterprises, we demonstrate the correctness, stability, and efficiency advantages of our solution. Second, using our unified architecture as the base, we perform a detailed quantitative sensitivity analysis and draw conclusions about the impact of different architectures, implementations, workloads, and system design choices.

A case for adapting channel width in wireless networks

We study a fundamental yet under-explored facet in wireless communication -- the width of the spectrum over which transmitters spread their signals, or the channel width. Through detailed measurements in controlled and live environments, and using only commodity 802.11 hardware, we first quantify the impact of channel width on throughput, range, and power consumption. Taken together, our findings make a strong case for wireless systems that adapt channel width. Such adaptation brings unique benefits. For instance, when the throughput required is low, moving to a narrower channel increases range and reduces power consumption; in fixed-width systems, these two quantities are always in conflict. We then present a channel width adaptation algorithm, called SampleWidth, for the base case of two communicating nodes. This algorithm is based on a simple search process that builds on top of existing techniques for adapting modulation. Per specified policy, it can maximize throughput or minimize power consumption. Evaluation using a prototype implementation shows that SampleWidth correctly identities the optimal width under a range of scenarios. In our experiments with mobility, it increases throughput by more than 60% compared to the best fixed-width configuration.

Dynamic graph query primitives for SDN-based cloudnetwork management

The need to provide customers with the ability to configure the network in current cloud computing environments has motivated the Networking-as-a-Service (NaaS) systems designed for the cloud. Such systems can provide cloud customers access to virtual network functions, such as network-aware VM placement, real time network monitoring, diagnostics and management, all while supporting multiple device management protocols. These network management functionalities depend on a set of underlying graph primitives. In this paper, we present the design and implementation of the software architecture including a shared graph library that can support network management operations. Using the illustrative case of all pair shortest path algorithm, we demonstrate how scalable lightweight dynamic graph query mechanisms can be implemented to enable practical computation times, in presence of network dynamism.

IPAC: IP-based Adaptive Packet Concatenation for Multihop Wireless Networks

Because medium contention occurs for each packet that is transmitted in a IEEE 802.11 wireless network, transmission of a large number of small packets can be particularly detrimental to performance. As a result of contention overhead, end-to-end delay and energy dissipation increase and the medium utilization decreases. In this paper, our goal is to reduce contention through concatenation of several small packets into a single large packet, and subsequently transmit this large packet. We propose IPAC, an IP-based packet concatenation protocol that adaptively selects an appropriate packet size based on the route quality. Simulation results show that with IPAC, contention is reduced by a factor of two, resulting in a throughput increase by a factor of two to three.

Unwanted Link Layer Traffic in Large IEEE 802.11 Wireless Networks

Wireless networks have evolved into an important technology for connecting users to the Internet. As the utility of wireless technology grows, wireless networks are being deployed in more widely varying conditions. The monitoring of wireless networks continues to reveal key implementation deficiencies that need to be corrected in order to improve protocol operation and end-to-end network performance. In wireless networks, where the medium is shared, unwanted traffic can pose significant overhead and lead to suboptimal network performance. Much of the previous analyses of unwanted traffic in wireless networks focus on malicious traffic. However, another major contributor of unwanted traffic is incorrect link layer behavior. Using data we collected from the 67th Internet Engineering Task Force (IETF) meeting held in November 2006, we show that a significant portion of link layer traffic stems from mechanisms that initiate, maintain, and change client-AP associations. We further show that under conditions of high medium utilization and packet loss rate, handoffs are initiated incorrectly. We analyze the traffic to understand when handoffs occur and whether the handoffs were beneficial or should have been avoided.

An information-centric architecture for data center networks

We propose a new Data Center Network (DCN) architecture, based on the principles of Information-Centric Networking (ICN). Our Info-Centric Data Center Network (IC-DCN) addresses many of the pain-points in current DCNs, such as network scalability, host mobility, etc. At the same time, IC-DCN introduces a number of new features to the current information-centric network architectures. We achieve this goal by decoupling the control-plane and data-plane functionalities. The control-plane is implemented in a centralized manner, while the data-plane is fully distributed. We show that IC-DCN effectively addresses many of the ICN challenges in the data center, such as routing scalability, full network utilization, and name space management.

Understanding handoffs in large ieee 802.11 wireless networks

As the utility of wireless technology grows, wireless networks are being deployed in more widely varying conditions. The monitoring of these networks continues to reveal key implementation deficiencies that need to be corrected in order to improve protocol operation and end-to-end performance. Using data we collected from the 67th Internet Engineering Task Force (IETF) meeting held in November 2006, we show that under conditions of high medium utilization and packet loss, handoffs can be incorrectly initiated. Using the notion of persistence and prevalence for the association of a client to an Access Point (AP), we show that although the clients were predominantly static, the handoff rate is surprisingly high. Through the analysis of the data set, we show that unnecessary handoff events not only increase the amount of management traffic in the network, but also severely impact client performance.

Analysis of Data from a Taxi Cab Participatory Sensor Network

Mobile participatory sensing applications are becoming quite popular, where individuals with mobile sensing devices such as smartphones, music players, and in-car GPS devices collect sensor data and share it with an external entity to compute statistics of mutual interest or map common phenomena. In this paper, we present an analysis of the data from a real-world city-scale mobile participatory sensor network comprised of about two thousand taxi cabs. Our analysis spans data collected from the taxi cab sensor network over the course of a year and we use it to make inferences about life in the city. The large scale data collection (size and time) from these taxi cabs allows us to examine various aspects about life in a city such as busy “party” times in the city, peak taxi usage (space and time), most traveled streets, and travel patterns on holidays. We also provide a summary of lessons learned from our analysis that can aid similar city-scale deployments and their analyses in the future.

Cloud service placement via subgraph matching

Fast service placement, finding a set of nodes with enough free capacity of computation, storage, and network connectivity, is a routine task in daily cloud administration. In this work, we formulate this as a subgraph matching problem. Different from the traditional setting, including approximate and probabilistic graphs, subgraph matching on data-center networks has two unique properties. (1) Node/edge labels representing vacant CPU cycles and network bandwidth change rapidly, while the network topology varies little. (2) There is a partial order on node/edge labels. Basically, one needs to place service in nodes with enough free capacity. Existing graph indexing techniques have not considered very frequent label updates, and none of them supports partial order on numeric labels. Therefore, we resort to a new graph index framework, Gradin, to address both challenges. Gradin encodes subgraphs into multi-dimensional vectors and organizes them with indices such that it can efficiently search the matches of a querys subgraphs and combine them to form a full match. In particular, we analyze how the index parameters affect update and search performance with theoretical results. Moreover, a revised pruning algorithm is introduced to reduce unnecessary search during the combination of partial matches. Using both real and synthetic datasets, we demonstrate that Gradin outperforms the baseline approaches up to 10 times.

Antler: A multi-tiered approach to automated wireless network management

Management of a large scale wireless network, be it an infrastructured WLAN or a metro-scale mesh network, presents several challenges. Troubleshooting problems related to wireless access in these networks requires a comprehensive set of metrics and network monitoring data. Current solutions gather large amounts of data and require significant bandwidth and processing to offload and analyze this management traffic. As a result, these solutions are typically not scalable or real-time. To this end, we propose a multi-tiered approach to wireless network monitoring that dynamically controls the granularity of data collection based on observed events in the network. Our approach can achieve significant bandwidth savings and enable real-time automated management of a wireless network. Our initial analysis using traces from a large WLAN shows a significant reduction in the amount of data collected to diagnose problems in a WLAN.