arXiv subjects

Yogesh Simmhan

Publications and source records attributed to Yogesh Simmhan.

At least 91 records · Page 5Linked to original sources

arXiv2017

ARM Wrestling with Big Data: A Study of Commodity ARM64 Server for Big Data Workloads

ARM processors have dominated the mobile device market in the last decade due to their favorable computing to energy ratio. In this age of Cloud data centers and Big Data analytics, the focus is increasingly on power efficient processing, rather than just high throughput computing. ARM's first commodity server-grade processor is the recent AMD A1100-series processor, based on a 64-bit ARM Cortex A57 architecture. In this paper, we study the performance and energy efficiency of a server based on this ARM64 CPU, relative to a comparable server running an AMD Opteron 3300-series x64 CPU, for Big Data workloads. Specifically, we study these for Intel's HiBench suite of web, query and machine learning benchmarks on Apache Hadoop v2.7 in a pseudo-distributed setup, for data sizes up to $20GB$ files, $5M$ web pages and $500M$ tuples. Our results show that the ARM64 server's runtime performance is comparable to the x64 server for integer-based workloads like Sort and Hive queries, and only lags behind for floating-point intensive benchmarks like PageRank, when they do not exploit data parallelism adequately. We also see that the ARM64 server takes $\frac{1}{3}^{rd}$ the energy, and has an Energy Delay Product (EDP) that is $50-71\%$ lower than the x64 server. These results hold promise for ARM64 data centers hosting Big Data workloads to reduce their operational costs, while opening up opportunities for further analysis.

cs.DC↗

arXiv2017

ECHO: An Adaptive Orchestration Platform for Hybrid Dataflows across Cloud and Edge

The Internet of Things (IoT) is offering unprecedented observational data that are used for managing Smart City utilities. Edge and Fog gateway devices are an integral part of IoT deployments to acquire real-time data and enact controls. Recently, Edge-computing is emerging as first-class paradigm to complement Cloud-centric analytics. But a key limitation is the lack of a platform-as-a-service for applications spanning Edge and Cloud. Here, we propose ECHO, an orchestration platform for dataflows across distributed resources. ECHO's hybrid dataflow composition can operate on diverse data models -- streams, micro-batches and files, and interface with native runtime engines like TensorFlow and Storm to execute them. It manages the application's lifecycle, including container-based deployment and a registry for state management. ECHO can schedule the dataflow on different Edge, Fog and Cloud resources, and also perform dynamic task migration between resources. We validate the ECHO platform for executing video analytics and sensor streams for Smart Traffic and Smart Utility applications on Raspberry Pi, NVidia TX1, ARM64 and Azure Cloud VM resources, and present our results.

cs.DC↗

arXiv2017

Demystifying Fog Computing: Characterizing Architectures, Applications and Abstractions

Internet of Things (IoT) has accelerated the deployment of millions of sensors at the edge of the network, through Smart City infrastructure and lifestyle devices. Cloud computing platforms are often tasked with handling these large volumes and fast streams of data from the edge. Recently, Fog computing has emerged as a concept for low-latency and resource-rich processing of these observation streams, to complement Edge and Cloud computing. In this paper, we review various dimensions of system architecture, application characteristics and platform abstractions that are manifest in this Edge, Fog and Cloud eco-system. We highlight novel capabilities of the Edge and Fog layers, such as physical and application mobility, privacy sensitivity, and a nascent runtime environment. IoT application case studies based on first-hand experiences across diverse domains drive this categorization. We also highlight the gap between the potential and the reality of Fog computing, and identify challenges that need to be overcome for the solution to be sustainable. Together, our article can help platform and application developers bridge the gap that remains in making Fog computing viable.

cs.DC↗

arXiv2017

Model-driven Scheduling for Distributed Stream Processing Systems

Distributed Stream Processing frameworks are being commonly used with the evolution of Internet of Things(IoT). These frameworks are designed to adapt to the dynamic input message rate by scaling in/out.Apache Storm, originally developed by Twitter is a widely used stream processing engine while others includes Flink, Spark streaming. For running the streaming applications successfully there is need to know the optimal resource requirement, as over-estimation of resources adds extra cost.So we need some strategy to come up with the optimal resource requirement for a given streaming application. In this article, we propose a model-driven approach for scheduling streaming applications that effectively utilizes a priori knowledge of the applications to provide predictable scheduling behavior. Specifically, we use application performance models to offer reliable estimates of the resource allocation required. Further, this intuition also drives resource mapping, and helps narrow the estimated and actual dataflow performance and resource utilization. Together, this model-driven scheduling approach gives a predictable application performance and resource utilization behavior for executing a given DSPS application at a target input stream rate on distributed resources.

cs.DC↗

arXiv2017

RIoTBench: A Real-time IoT Benchmark for Distributed Stream Processing Platforms

The Internet of Things (IoT) is an emerging technology paradigm where millions of sensors and actuators help monitor and manage, physical, environmental and human systems in real-time. The inherent closedloop responsiveness and decision making of IoT applications make them ideal candidates for using low latency and scalable stream processing platforms. Distributed Stream Processing Systems (DSPS) hosted on Cloud data-centers are becoming the vital engine for real-time data processing and analytics in any IoT software architecture. But the efficacy and performance of contemporary DSPS have not been rigorously studied for IoT applications and data streams. Here, we develop RIoTBench, a Realtime IoT Benchmark suite, along with performance metrics, to evaluate DSPS for streaming IoT applications. The benchmark includes 27 common IoT tasks classified across various functional categories and implemented as reusable micro-benchmarks. Further, we propose four IoT application benchmarks composed from these tasks, and that leverage various dataflow semantics of DSPS. The applications are based on common IoT patterns for data pre-processing, statistical summarization and predictive analytics. These are coupled with four stream workloads sourced from real IoT observations on smart cities and fitness, with peak streams rates that range from 500 to 10000 messages/sec and diverse frequency distributions. We validate the RIoTBench suite for the popular Apache Storm DSPS on the Microsoft Azure public Cloud, and present empirical observations. This suite can be used by DSPS researchers for performance analysis and resource scheduling, and by IoT practitioners to evaluate DSPS platforms.

cs.DC↗

arXiv2016

Knowledge-infused and Consistent Complex Event Processing over Real-time and Persistent Streams

Emerging applications in Internet of Things (IoT) and Cyber-Physical Systems (CPS) present novel challenges to Big Data platforms for performing online analytics. Ubiquitous sensors from IoT deployments are able to generate data streams at high velocity, that include information from a variety of domains, and accumulate to large volumes on disk. Complex Event Processing (CEP) is recognized as an important real-time computing paradigm for analyzing continuous data streams. However, existing work on CEP is largely limited to relational query processing, exposing two distinctive gaps for query specification and execution: (1) infusing the relational query model with higher level knowledge semantics, and (2) seamless query evaluation across temporal spaces that span past, present and future events. These allow accessible analytics over data streams having properties from different disciplines, and help span the velocity (real-time) and volume (persistent) dimensions. In this article, we introduce a Knowledge-infused CEP (X-CEP) framework that provides domain-aware knowledge query constructs along with temporal operators that allow end-to-end queries to span across real-time and persistent streams. We translate this query model to efficient query execution over online and offline data streams, proposing several optimizations to mitigate the overheads introduced by evaluating semantic predicates and in accessing high-volume historic data streams. The proposed X-CEP query model and execution approaches are implemented in our prototype semantic CEP engine, SCEPter. We validate our query model using domain-aware CEP queries from a real-world Smart Power Grid application, and experimentally analyze the benefits of our optimizations for executing these queries, using event streams from a campus-microgrid IoT deployment.

cs.DB↗

arXiv2016

A Meta-graph Approach to Analyze Subgraph-centric Distributed Programming Models

Component-centric distributed graph processing platforms that use a bulk synchronous parallel (BSP) programming model have gained traction. These address the short-comings of Big Data abstractions/platforms like MapReduce/Hadoop for large-scale graph processing. However, there is limited literature on foundational aspects of the behavior of these component-centric abstractions for different graphs, graph partitioning, and graph algorithms. Here, we propose a analytical approach based on a meta-graph sketch to examine the characteristics of component-centric graph programming models at a coarse granularity. In particular, we apply this sketch to subgraph- and block-centric abstractions, and draw a comparison with vertex-centric models like Google's Pregel. First, we explore the impact of various graph partitioning techniques on the meta-graph, and next consider the impact of the meta-graph on graph algorithms. This decouples the unwieldy large graph and their partitioning specific artifacts from their algorithmic analysis. We use 5 spatial and powerlaw graphs as exemplars, four different partitioning strategies, and PageRank and Breadth First Search as canonical algorithms. These analysis over the meta-graphs provide a reliable measure of the expected number of supersteps, and the communication and computational complexity of the algorithms for various graphs, and the relative merits of subgraph-centric models over vertex-centric ones.

cs.DC↗

arXiv2016

Introducing Distributed Dynamic Data-intensive (D3) Science: Understanding Applications and Infrastructure

A common feature across many science and engineering applications is the amount and diversity of data and computation that must be integrated to yield insights. Data sets are growing larger and becoming distributed; and their location, availability and properties are often time-dependent. Collectively, these characteristics give rise to dynamic distributed data-intensive applications. While "static" data applications have received significant attention, the characteristics, requirements, and software systems for the analysis of large volumes of dynamic, distributed data, and data-intensive applications have received relatively less attention. This paper surveys several representative dynamic distributed data-intensive application scenarios, provides a common conceptual framework to understand them, and examines the infrastructure used in support of applications.

cs.DC↗

arXiv2016

Benchmarking Distributed Stream Processing Platforms for IoT Applications

Internet of Things (IoT) is a technology paradigm where millions of sensors monitor, and help inform or manage, physical, envi- ronmental and human systems in real-time. The inherent closed-loop re- sponsiveness and decision making of IoT applications makes them ideal candidates for using low latency and scalable stream processing plat- forms. Distributed Stream Processing Systems (DSPS) are becoming es- sential components of any IoT stack, but the efficacy and performance of contemporary DSPS have not been rigorously studied for IoT data streams and applications. Here, we develop a benchmark suite and per- formance metrics to evaluate DSPS for streaming IoT applications. The benchmark includes 13 common IoT tasks classified across various func- tional categories and forming micro-benchmarks, and two IoT applica- tions for statistical summarization and predictive analytics that leverage various dataflow compositional features of DSPS. These are coupled with stream workloads sourced from real IoT observations from smart cities. We validate the IoT benchmark for the popular Apache Storm DSPS, and present empirical results.

cs.DC↗

arXiv2016

Towards a Practical Architecture for the Next Generation Internet of Things

The Internet of Things, or the IoT is a vision for a ubiquitous society wherein people and "Things" are connected in an immersively networked computing environment, with the connected "Things" providing utility to people/enterprises and their digital shadows, through intelligent social and commercial services. Translating this idea to a conceivable reality is a work in progress for more than a decade. Current IoT architectures are predicated on optimistic assumptions on the evolution and deployment of IoT technologies. We believe many of these assumptions will not be met, consequently impeding the practical and sustainable deployment of IoT. In this article, we explore use-cases across different applications domains that can potentially benefit from an IoT infrastructure, and analyze them in the context of an alternative world-view that is more grounded in reality. Despite this more conservative approach, we argue that adopting certain design paradigms when architecting an IoT ecosystem can achieve much of the promised benefits in a practical and sustainable manner.

cs.CY↗

arXiv2016

Benchmarking Fast Data Platforms for the Aadhaar Biometric Database

Aadhaar is the world's largest biometric database with a billion records, being compiled as an identity platform to deliver social services to residents of India.Aadhaar processes streams of biometric data as residents are enrolled and updated.Besides 1 million enrolments and updates per day,up to 100 million daily biometric authentications are expected during delivery of various public services.These form critical Big Data applications,with large volumes and high velocity of data.

cs.DC↗

arXiv2015

Elastic Resource Allocation for Distributed Graph Processing Platforms

Distributed graph platforms like Pregel have used vertex- centric programming models to process the growing corpus of graph datasets using commodity clusters. The irregular structure of graphs cause load imbalances across machines operating on graph partitions, and this is exacerbated for non-stationary graph algorithms such as traversals, where not all parts of the graph are active at the same time. As a result, such graph platforms, even as they scale, do not make efficient use of distributed resources. Clouds offer elastic virtual machines that can be leveraged to improve the resource utilization for such platforms and hence reduce the monetary cost for their execution. In this paper, we propose strategies for elastic placement of graph partitions on Cloud VMs for subgraphcentric programming model to reduce the cost of execution compared to a static placement, even as we minimize the increase in makespan. These strategies are innovative in modeling the graph algorithms behavior. We validate our strategies for several graphs, using runtime tra- ces for their distributed execution of a Breadth First Search (BFS) algorithms on our subgraph-centric GoFFish graph platform. Our strategies are able to reduce the cost of exe- cution by up to 42%, compared to a static placement, while achieving a makespan that is within 29% of the optimal

cs.DC↗

arXiv2015

An Interoperable Realization of Smart Cities with Plug and Play based Device Management

The primal problem with Internet of Things (IoT) solutions for smart cities is the lack of interoperability at various levels, and more predominately at the device level. While there exist multitude of platforms from multiple manufacturers, the existing ecosystem still remains highly closed. In this paper, we propose SNaaS or Sensor/Network as a Service: a service layer that enables the creation of the plug-n-play infrastructure, across platforms from multiple vendors, necessary for interoperability and successful deployment of large-scale city wide systems. In order to correctly position the new service layer, we present a high level reference IoT architecture for smart city implementations, and follow it up with the workflow details of SNaaS along with preliminary microbenchmarks.

eess.SY↗

arXiv2014

Towards a Practical Architecture for India Centric Internet of Things

An effective architecture for the Internet of Things (IoT), particularly for an emerging nation like India with limited technology penetration at the national scale, should be based on tangible technology advances in the present, practical application scenarios of social and entrepreneurial value, and ubiquitous capabilities that make the realization of IoT affordable and sustainable. Humans, data, communication and devices play key roles in the IoT ecosystem that we perceive. In a push towards this sustainable and practical IoT Architecture for India, we synthesize ten design paradigms to consider.

cs.HC↗

arXiv2014

Floe: A Continuous Dataflow Framework for Dynamic Cloud Applications

Applications in cyber-physical systems are increasingly coupled with online instruments to perform long running, continuous data processing. Such "always on" dataflow applications are dynamic, where they need to change the applications logic and performance at runtime, in response to external operational needs. Floe is a continuous dataflow framework that is designed to be adaptive for dynamic applications on Cloud infrastructure. It offers advanced dataflow patterns like BSP and MapReduce for flexible and holistic composition of streams and files, and supports dynamic recomposition at runtime with minimal impact on the execution. Adaptive resource allocation strategies allow our framework to effectively use elastic Cloud resources to meet varying data rates. We illustrate the design patterns of Floe by running an integration pipeline and a tweet clustering application from the Smart Power Grids domain on a private Eucalyptus Cloud. The responsiveness of our resource adaptation is validated through simulations for periodic, bursty and random workloads.

cs.DC↗

arXiv2014

Scalable Analytics over Distributed Time-series Graphs using GoFFish

Graphs are a key form of Big Data, and performing scalable analytics over them is invaluable to many domains. As our ability to collect data grows, there is an emerging class of inter-connected data which accumulates or varies over time, and on which novel analytics - both over the network structure and across the time-variant attribute values - is necessary. We introduce the notion of time-series graph analytics and propose Gopher, a scalable programming abstraction to develop algorithms and analytics on such datasets. Our abstraction leverages a sub-graph centric programming model and extends it to the temporal dimension using an iterative BSP (Bulk Synchronous Parallel) approach. Gopher is co-designed with GoFS, a distributed storage specialized for time-series graphs, as part of the GoFFish distributed analytics platform. We examine storage optimizations for GoFS, design patterns in Gopher to leverage the distributed data layout, and evaluate the GoFFish platform using time-series graph data and applications on a commodity cluster.

cs.DC↗

arXiv2014

Holistic Measures for Evaluating Prediction Models in Smart Grids

The performance of prediction models is often based on "abstract metrics" that estimate the model's ability to limit residual errors between the observed and predicted values. However, meaningful evaluation and selection of prediction models for end-user domains requires holistic and application-sensitive performance measures. Inspired by energy consumption prediction models used in the emerging "big data" domain of Smart Power Grids, we propose a suite of performance measures to rationally compare models along the dimensions of scale independence, reliability, volatility and cost. We include both application independent and dependent measures, the latter parameterized to allow customization by domain experts to fit their scenario. While our measures are generalizable to other domains, we offer an empirical analysis using real energy use data for three Smart Grid applications: planning, customer education and demand response, which are relevant for energy sustainability. Our results underscore the value of the proposed measures to offer a deeper insight into models' behavior and their impact on real applications, which benefit both data mining researchers and practitioners.

cs.LG↗

arXiv2013

On Using Complex Event Processing for Dynamic Demand Response Optimization in Microgrid

Demand-side load reduction is a key benefit of Smart Grids. However, existing demand response optimization (DR) programs fail to effectively leverage the near-realtime information available from smart meters and Building Area Networks to respond dynamically to changing energy use profiles. We investigate the use of semantic Complex Event Processing (CEP) patterns to model and detect dynamic situations in a campus microgrid to facilitate adaptive DR. Our focus is on demand-side management rather than supply-side constraints. Continuous data from information sources like smart meters and building sensors are abstracted as event streams. Event patterns for situations that assist with DR are detected from them. Specifically, we offer a taxonomy of event patterns that can guide operators to define situations of interest and we illustrate its efficacy for DR by applying these patterns on realtime events in the USC Campus microgrid using our CEP framework

cs.DC↗