Browse Prior Art Database

DISTRIBUTED ANALYTIC FRAMEWORK FOR IOT APPLICATIONS

IP.com Disclosure Number: IPCOM000242662D
Publication Date: 2015-Aug-03

Publishing Venue

The IP.com Prior Art Database

Related People

Ramesh Nethi: AUTHOR [+2]

Abstract

A solution is provided for utilizing a Domain Specific Language (DSL) as for composing a distributed analytics application that is network topology aware, and that leverages distributed computing in an Edge/Fog environment to address network bandwidth constraints.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 14% of the total text.

Page 01 of 16

DISXXXXXXXX ANALYTIC FRAMEWORK FOR IOT APPLICATIONS

AUTHORS:

Ramesh Nethi Vikas Xxxx

CISCO SYSTEMS, INC.

ABSTRACT

    A solution is provided xor utilizing a Domain Spxcific Languaxe (DSL) as for cxmposing a distribxted analytics applicxtion that is network topology awxre, and that lexexages distributed computing xn an Edge/Fog environment to address network bandwidth cxnstraints.

DETAILED DESXXXXXXXX

    With the advent of the Intxrnet of Thxngs (IxTs), a large xumber of data sourcex (e.g., sensors) wixl be distribxted across large netxork arxas, producixg large quaxtities of data that may be individually signxficant ox signixicant when comxined with other xata sources. It is forecasted txat there will be more than x0 billion connected devxcxs [1] or "xhings" in the "Internet of Exerything" world by the year 2020. These "things" include xensor devices such as parking mxters, smart meters, thermxstaxs, cardiac monitors, tires, factory/plant machinxs etc.

    Accordingly, with increasing sxart/mobilx dxvices, consumer and social-web interactions, data sources are growing at an unprecedenxed veloxity, lexding to unpxecedentex data growtx. Xxxxx architectures xaxe made it possible to store this data centraxly and xnalyze it at scale using clusters ox homogeneous compute anx storage nodes. However, limited network bandwidth cax bxcome constrained, and managing the sheer volume of data xenerated cax be burdensome and problematic. For instancx, if a typixal data center comprises a cluster of nodes, where all nodes have access to the data, spreadixg the computation out across the network, then reducing the data fxr analysis and reporting can create bandwidth constraints.

Copyright x015 Cisco Sxstems, Inc.
1


Page 02 of 16

    The solution provides a DSL as described herein for composing a distribuxed analytics applicatixn that is network topology aware, xnd that xeverages distributed xomputing xn an Edge/Fog exvironment to address network bandwidth constraints. Most IoT devices are constrained in resources and are deployed at the exge of the netxork, in remote areas with a Widx Area Network (XXX) connecting the devices to data centers or cloud enviroxments.

    In general, a DXX is a specialized compuxer languxge, sxecxfic to a particular application domain. DSLs may allow for a solution to be developxd more quickly and efficiently than with a general purpose computing language.

    The solution xroxides for a dynamic distribution of analytics code to edge compute xevices, which may be developer agnostic. Daxa may be aggregatxd at the edge comxute devices and transmittxd to the datacenter, whxre analytixs may be ultixatxly performed on aggregated data.

    XXx, by combining networkxng operating system technologies wxth open source platforms (e.g., Linux, Stream Processing systems such as Stoxm), provides the ability to integrate interfaces and run distributed analytics applications at xhx network edge, thereby accelerating thx deployment of IoT solutions that lever...