This paper presents MAUI, a system that enables fine-grained energy-aware offload of mobile code to the infrastructure. Previous approaches to these problems either relied heavily on programmer support to partition an application, or they were coarse-grained requiring full process (or full VM) migration. MAUI uses the benefits of a managed code environment to offer the best of both worlds: it supports fine-grained code offload to maximize energy savings with minimal burden on the programmer. MAUI decides at run-time which methods should be remotely executed, driven by an optimization engine that achieves the best energy savings possible under the mobile device's current connectivity constrains. In our evaluation, we show that MAUI enables: 1) a resource-intensive face recognition application that consumes an order of magnitude less energy, 2) a latency-sensitive arcade game application that doubles its refresh rate, and 3) a voice-based language translation application that bypasses the limitations of the smartphone environment by executing unsupported components remotely.

/pdf/maui-making-smartphones-last-longer-with-code-offload-go3o1uwp1u.pdf

MAUI: making smartphones last longer with code offload

Smartphones have exploded in popularity in recent years, becoming ever more sophisticated and capable. As a result, developers worldwide are building increasingly complex applications that require ever increasing amounts of computational power and energy. In this paper we propose ThinkAir, a framework that makes it simple for developers to migrate their smartphone applications to the cloud. ThinkAir exploits the concept of smartphone virtualization in the cloud and provides method-level computation offloading. Advancing on previous work, it focuses on the elasticity and scalability of the cloud and enhances the power of mobile cloud computing by parallelizing method execution using multiple virtual machine (VM) images. We implement ThinkAir and evaluate it with a range of benchmarks starting from simple micro-benchmarks to more complex applications. First, we show that the execution time and energy consumption decrease two orders of magnitude for a N-queens puzzle application and one order of magnitude for a face detection and a virus scan application. We then show that a parallelizable application can invoke multiple VMs to execute in the cloud in a seamless and on-demand manner such as to achieve greater reduction on execution time and energy consumption. We finally use a memory-hungry image combiner tool to demonstrate that applications can dynamically request VMs with more computational power in order to meet their computational requirements.

/pdf/thinkair-dynamic-resource-allocation-and-parallel-execution-5di1n63wfi.pdf

ThinkAir: Dynamic resource allocation and parallel execution in the cloud for mobile code offloading

With the recent advent of 4G LTE networks, there has been increasing interest to better understand the performance and power characteristics, compared with 3G/WiFi networks. In this paper, we take one of the first steps in this direction.Using a publicly deployed tool we designed for Android called 4GTest attracting more than 3000 users within 2 months and extensive local experiments, we study the network performance of LTE networks and compare with other types of mobile networks. We observe LTE generally has significantly higher downlink and uplink throughput than 3G and even WiFi, with a median value of 13Mbps and 6Mbps, respectively. We develop the first empirically derived comprehensive power model of a commercial LTE network with less than 6% error rate and state transitions matching the specifications. Using a comprehensive data set consisting of 5-month traces of 20 smartphone users, we carefully investigate the energy usage in 3G, LTE, and WiFi networks and evaluate the impact of configuring LTE-related parameters. Despite several new power saving improvements, we find that LTE is as much as 23 times less power efficient compared with WiFi, and even less power efficient than 3G, based on the user traces and the long high power tail is found to be a key contributor. In addition, we perform case studies of several popular applications on Android in LTE and identify that the performance bottleneck for web-based applications lies less in the network, compared to our previous study in 3G [24]. Instead, the device's processing power, despite the significant improvement compared to our analysis two years ago, becomes more of a bottleneck.

/pdf/a-close-examination-of-performance-and-power-characteristics-2g7pjkqkb1.pdf

A close examination of performance and power characteristics of 4G LTE networks

Where is the energy spent inside my app? Despite the immense popularity of smartphones and the fact that energy is the most crucial aspect in smartphone programming, the answer to the above question remains elusive. This paper first presents eprof, the first fine-grained energy profiler for smartphone apps. Compared to profiling the runtime of applications running on conventional computers, profiling energy consumption of applications running on smartphones faces a unique challenge, asynchronous power behavior, where the effect on a component's power state due to a program entity lasts beyond the end of that program entity. We present the design, implementation and evaluation of eprof on two mobile OSes, Android and Windows Mobile.We then present an in-depth case study, the first of its kind, of six popular smartphones apps (including Angry-Birds, Facebook and Browser). Eprof sheds lights on internal energy dissipation of these apps and exposes surprising findings like 65%-75% of energy in free apps is spent in third-party advertisement modules. Eprof also reveals several "wakelock bugs", a family of "energy bugs" in smartphone apps, and effectively pinpoints their location in the source code. The case study highlights the fact that most of the energy in smartphone apps is spent in I/O, and I/O events are clustered, often due to a few routines. Thismotivates us to propose bundles, a new accounting presentation of app I/O energy, which helps the developer to quickly understand and optimize the energy drain of her app. Using the bundle presentation, we reduced the energy consumption of four apps by 20% to 65%.

/pdf/where-is-the-energy-spent-inside-my-app-fine-grained-energy-xumhb9g4z3.pdf

Where is the energy spent inside my app?: fine grained energy accounting on smartphones with Eprof

Smart phones are now capable of supporting a wide range of applications, many of which demand an ever increasing computational power. This poses a challenge because smart phones are resource-constrained devices with limited computation power, memory, storage, and energy. Fortunately, the cloud computing technology offers virtually unlimited dynamic resources for computation, storage, and service provision. Therefore, researchers envision extending cloud computing services to mobile devices to overcome the smartphones constraints. The challenge in doing so is that the traditional smartphone application models do not support the development of applications that can incorporate cloud computing features and requires specialized mobile cloud application models. This article presents mobile cloud architecture, offloading decision affecting entities, application models classification, the latest mobile cloud application models, their critical analysis and future research directions.

A Survey of Mobile Cloud Computing Application Models

This paper describes PowerBooter, an automated power model construction technique that uses built-in battery voltage sensors and knowledge of battery discharge behavior to monitor power consumption while explicitly controlling the power management and activity states of individual components. It requires no external measurement equipment. We also describe PowerTutor, a component power management and activity state introspection based tool that uses the model generated by PowerBooter for online power estimation. PowerBooter is intended to make it quick and easy for application developers and end users to generate power models for new smartphone variants, which each have different power consumption properties and therefore require different power models. PowerTutor is intended to ease the design and selection of power efficient software for embedded systems. Combined, PowerBooter and PowerTutor have the goal of opening power modeling and analysis for more smartphone variants and their users.

http://ziyang.eecs.umich.edu/~lzh228/publications/zhang10oct.pdf

Accurate online power estimation and automatic battery behavior based power model generation for smartphones

The use of cellular data networks is increasingly popular due to the widespread deployment of 3G technologies and the rapid adoption of smartphones, such as iPhone and GPhone. Besides email and web browsing, a variety of network applications are now available, rendering smartphones potentially useful substitutes for their desktop counterparts. Nevertheless, the performance of smartphone applications in the wild is still poorly understood due to a lack of systematic measurement methodology.We identify and study important factors that impact user-perceived performance of network applications on smartphones. We develop a systematic methodology for comparing this performance along several key dimensions such as carrier networks, device capabilities, and server configurations. To ensure a fair and representative comparison, we conduct controlled experiments, informed by data collected through 3GTest, a cross-platform measurement tool we designed, executed by more than 30,000 users from all over the world. Our work is an essential step towards understanding the performance of smartphone applications from the perspective of users, application developers, cellular network operators, and smartphone vendors. Our analysis culminates with a set of recommendations that can lead to better application design and infrastructure support for smartphone users.

/pdf/anatomizing-application-performance-differences-on-2z4u7ig76j.pdf

Anatomizing application performance differences on smartphones

Cloud applications have increasingly come to rely on distributed storage systems that hide the complexity of handling network and node failures behind simple, data-centric interfaces (such as PUTs and GETs on key-value pairs). While these interfaces are very easy to use, the application is completely oblivious to the location of its data in the network; as a result, it has no way to optimize the placement of data or computation. In this paper, we propose exposing the network location of data to applications. The primary challenge is that data does not usually exist at a single point in the network; it can be striped, replicated, cached and coded across different locations, in arbitrary ways that vary across storage systems. For example, an item that is synchronously mirrored in both Seattle and London will appear equally far from both locations for writes, but equally close to both locations for reads. Accordingly, we describe Contour, a system that allows applications to query and manipulate the location of data without requiring them to be aware of the physical machines storing the data, the replication protocols used or the underlying network topology.

/pdf/location-location-location-modeling-data-proximity-in-the-4yy2jwgaoq.pdf

Location, location, location!: modeling data proximity in the cloud

A cloud statistics server generates statistics for a cloud service based on an identified data item and an identified operation. The cloud service may include various computing nodes and storage nodes. The cloud statistics may include expected completion times for the identified operation and the identified data item with respect to each of the computing nodes. A computing node may be selected to execute the identified operation based on the expected completion times. The generated statistics may be generated by the cloud statistics server using a network topology associated with the data item that is based on the latencies or expected transfer times between the various storage nodes and computing nodes, and a replication strategy used by the cloud service. The topology may be implemented as a directed graph with edge weights corresponding to expected transfer times between each node.

Selecting computing nodes in cloud service using replication topologies

Social networks and content publishing platforms have newsfeed applications, which show both organic content to drive engagement, and ads to drive revenue. This paper focuses on the problem of ads allocation in a newsfeed to achieve an optimal balance of revenue and engagement. To the best of our knowledge, we are the first to report practical solutions to this business-critical and popular problem in industry. The paper describes how large-scale recommender system like feed ranking works, and why it is useful to consider ads allocation as a post-operation once the ranking of organic items and (separately) the ranking of ads are done. A set of computationally lightweight algorithms are proposed based on various sets of assumptions in the context of ads on the LinkedIn newsfeed. Through both offline simulation and online A/B tests, benefits of the proposed solutions are demonstrated. The best performing algorithm is currently fully deployed on the LinkedIn newsfeed and is serving all live traffic.

Birjodh Tiwana

Papers

Accurate online power estimation and automatic battery behavior based power model generation for smartphones

Anatomizing application performance differences on smartphones

Location, location, location!: modeling data proximity in the cloud

Selecting computing nodes in cloud service using replication topologies

Ads Allocation in Feed via Constrained Optimization