Showing papers by "Samir R. Das published in 2017"

Improving User Perceived Page Load Times Using Gaze

Abstract: We take a fresh look at Web page load performance from the point of view of user experience. Our user study shows that perceptual performance, defined as user-perceived page load time (uPLT) poorly correlates with traditional page load time (PLT) metrics. However, most page load optimizations are designed to improve the traditional PLT metrics, rendering their impact on user experience uncertain. Instead, we present WebGaze, a system that specifically optimizes for the uPLT metric. The key insight in WebGaze is that user attention and interest can be captured using a user’s eye gaze and can inturn be used to improve uPLT. We collect eye gaze data from 50 users across 45 Web pages and find that there is commonality in user attention across users. Specifically, users are drawn to certain regions on the page, that we call regions of high collective fixation. WebGaze prioritizes loading objects that exhibit a high degree of collective fixation to improve user-perceived latencies. We compare WebGaze with three alternate strategies, one of which is the state-of-the-art system that also uses prioritization to improve user experience. Our evaluation based on a user study shows that WebGaze improves median uPLT for 73% of the Web pages compared to all three alternate strategies.

SpecSense: Crowdsensing for efficient querying of spectrum occupancy

Abstract: We describe an end-to-end platform called SpecSense to support large scale spectrum monitoring. SpecSense crowdsources spectrum monitoring to low-cost, low-power commodity SDR/embedded platforms and provides necessary analytics support in a central spectrum server. In this work, we describe SpecSense and address specific challenges related to accurately estimate spectrum occupancy on demand with low overhead. To address the accuracy question, we augment state-of-the-art spatial interpolation techniques to accommodate scenarios where RF propagation characteristics change across space. To address the overhead question, we solve the sensor selection problem to select the minimum number of spectrum sensors that can best estimate the spectrum at the requested locations.

Design of a backscatter-based Tag-to-Tag system

Abstract: Practical technologies for the Internet of Things (IoT) must provide connectivity to all objects under a common framework irrespective of their size or value. Power requirement, cost of wireless devices and scalability have proved critical bottlenecks for the universal deployment of the IoT. One approach to address these issues is the use of a communication paradigm where the devices communicate via backscattering and exploit harvested power from an external RF source. In a Backscattering Tag-to-Tag Network (BTTN), the tags themselves are able to read and interpret the backscattered communications from other neighboring tags. In the tag-to-tag link, the BTTN tag has to demodulate a receiving signal with a low modulation index. In order to improve the link range, we propose a power-efficient demodulator design that enables the receiving tag to quantify the amplitude-shift keying (ASK) modulated signal with a modulation index as low as 0.6%. The demodulator consumes 1.21 µW at 1.1 V supply voltage at a data rate of 10 kbps.

FSONet: A Wireless Backhaul for Multi-Gigabit Picocells Using Steerable Free Space Optics

Abstract: Expected increase in cellular demand has pushed recent interest in picocell networks which have reduced cell sizes (100-200m or less). For ease of deployment of such networks, a wireless backhaul network is highly desired. Since RF-based technologies are unlikely to provide the desired multi-gigabit data rates, we motivate and explore use of free space optics (FSO) for picocell backhaul. In particular, we present a novel network architecture based on steerable links and sufficiently many robust short-range links, to help circumvent the key challenge of outdoor effects in reliable operation of outdoor FSO links. Our architecture is motivated by the fact that, due to the high density of picocells, many short-range links will occur naturally in a picocell backhaul. Moreover, use of steerable FSO links facilitates networks with sufficient redundancy while using only a small number of interfaces per node. We address the key problems that arise in the context of such a backhaul architecture, viz., an FSO link design with desired characteristics, and related network design and management problems. We develop and evaluate a robust 100m FSO link prototype, and simulate the proposed architecture in many metro US cities while show its viability via evaluation of key performance metrics.

FireFly: reconfigurable optical wireless networking data centers

Abstract: We explore a novel, free-space optics based approach for building data center interconnects. Data centers (DCs) are a critical piece of today’s networked applications in both private and public sectors. The key factors that have driven this trend are economies of scale, reduced management costs, better utilization of hardware via statistical multiplexing, and the ability to elastically scale applications in response to changing workload patterns. A robust DC network fabric is fundamental to the success of DCs and to ensure that the network does not become a bottleneck for high-performance applications. In this context, DC network design must satisfy several goals: high performance (e.g., high throughput and low latency), low equipment and management cost, robustness to dynamic traffic patterns, incremental expandability to add new servers or racks, and other practical concerns such as cabling complexity, and power and cooling costs. Current DC network architectures do not seem to provide a satisfactory solution, with respect to the above requirements. In particular, traditional static (wired) networks are either overprovisioned or oversubscribed. Recent works have tried to overcome the above limitations by augmenting a static (wired) “core” with some flexible links (RF-wireless or optical). These augmented architectures show promise, but offer only incremental improvement in performance. Specifically, RFwireless based augmented solutions also offer only limited performance improvement, due to inherent interference and range constraints of RF links. This paper explores an alternative design point—a fully flexible and all-wireless DC interrack network based on free-space optical (FSO) links. We call this FireFly as in; Free-space optical Inter-Rack nEtwork with high FLexibilitY. We will present our designs and tests using various configurations that can help the performance and reliability of the FSO links.

Demystifying Hardware Bottlenecks in Mobile Web Quality of Experience

Abstract: Mobile web page load time depends on three key factors: (1) the complexity of the Webpage, (2) the underlying network conditions, and (3) the processing capability of the device. While there are several works focusing on the Web complexity and the network, there is a little work in understanding the hardware bottlenecks in the page load process. In this poster, we analyze the effect of hardware bottlenecks of Web pages. We also analyze the effect of GPU offloading, a commonly used solution to speed up Web page loads.

A first look at performance of TV streaming sticks

Abstract: Recent measurements show that more than half of the peak time Internet traffic is due to video streaming. Recent trends also suggest that consumers are increasingly receiving their TV content over the Internet via streaming appliances that are connected to the TV. We present the first systematic measurement study of a popular class of such devices that have the ‘stick’ form factor. The study covers streaming and network related performance using a widely used content server on the Internet (Netflix) and a local instrumented media server. On the user-end, we use three widely available mediasticks in the US – Chromecast , Amazon Fire and Roku . We observe that there are significant performance differentials across the streaming sticks. Our experiments show that Amazon Fire and Chromecast provide better user experience in the presence of varying network conditions, whereas Roku performs best at high stable bandwidth.