Mobile Computing at the Edge Grace A. Lewis Software Engineering Institute Carnegie Mellon University [email protected] Software Technology Conference (STC 2015) October 14, 2015 © 2015 Carnegie Mellon University Mobile Device Trends Smartphones, tablets, and now phablets, have become for many the preferred way of interacting with the Internet, social media and the enterprise • Mobile devices are increasingly becoming the first go- to device for communications and content consumption [1] [5] [8] • Number of mobile devices will surpass desktops for the first time this year [9] • The time people spend using their smartphone is now exceeding the time spent looking at TV screens [3] • Not uncommon for there to be multiple mobile devices per user and household [7] • Wearable technology is showing a consistent increase in popularity [2] Organizations are pushing out more and more content and functionality to mobile users Mobile Computing at the Edge STC 2015 2 © 2015 Carnegie Mellon University Mobile Traffic Will Keep Increasing By the end of 2015, 4G LTE data use will rise by 59% and mobile video will account for 60% of data traffic [4] Source: Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update 2014–2019 White Paper [15] Wearables market will grow five-fold in the next five years from 109 million devices in 2014 to 578 million devices by 2019. This will result in an 18-fold increase in mobile data traffic [15] Mobile Computing at the Edge STC 2015 3 © 2015 Carnegie Mellon University Internet of Things (IoT): More Than Just “Things” 4.9 billion connected things will be in use in 2015, up 30 percent from 2014, and will reach 25 billion by 2020 [12] 99 percent of physical objects will eventually become part of a network[13] Permission for diagram use by Carnegie Mellon University consistent with its status as a non-profit University for any purpose the institution sees fit by Matt Ceniceros, @mattceni, mattceni.com Mobile Computing at the Edge STC 2015 4 © 2015 Carnegie Mellon University The 3Vs of Data: Volume, Velocity and Variety New Vs: Veracity, Validity, Volatility, Visualization, Vulnerability and Value [14] What to do with all the data? How does it change business processes and business models? Permission for diagram use by Carnegie Mellon University consistent with its status as a non-profit University for any purpose the institution sees fit by Amy Allen, www.qmee.com Mobile Computing at the Edge STC 2015 5 © 2015 Carnegie Mellon University Top Jobs in the Next 7 Years According to Gartner [25] Integration Specialists Digital Business Architects Regulatory Analysts Risk Professionals Mobile Computing at the Edge STC 2015 6 © 2015 Carnegie Mellon University Smartphone Penetration Source: KPCB. 2015 Internet Trends. http://www.kpcb.com/internet-trends [6] Mobile Computing at the Edge STC 2015 7 © 2015 Carnegie Mellon University Low-Cost Smartphones By 2020, 75 Some low-cost percent of smartphones are smartphone expected to reach buyers will pay approximately $35 less than $100 unsubsidized by for a device [1] year-end 2014 [1] As an example, ZTE manufactures several low-end smartphones under $50 and is quickly gaining market share globally [16] Mobile Computing at the Edge STC 2015 8 © 2015 Carnegie Mellon University Therefore … Not unreasonable for users to expect the performance and capabilities of mobile devices to be equal to laptops and desktops However … • Mobile devices will always lag behind their PC counterparts due to size and battery limitations • Large and variable end-to-end latency between mobile device and cloud, and the possibility of disruptions, have a negative effect on user experience • Will only get worse with the amount of network traffic generated by IoT and growing market share of low-cost smartphones Mobile Computing at the Edge STC 2015 9 © 2015 Carnegie Mellon University Edge Computing Idea is to push applications, data and computing power to the logical extremes of a network — closer to where they are being used • As an example, Akamai has servers around the world to distribute web site content from locations close to the user Cyber-foraging moves the edge even closer to the user Mobile Computing at the Edge STC 2015 10 © 2015 Carnegie Mellon University Cyber-Foraging Leverage of external resource-rich surrogates to augment the capabilities of resource-limited devices • Computation Offload – Offload of expensive computation in order to extend battery life and increase computational capability • Data Staging – Improve data transfers between mobile computers and the cloud by Nokia Siemens Networks temporarily staging data in transit on Liquid Applications intermediate surrogates Industry is starting to build on this concept to improve mobile user experience and decrease network traffic [10] [11] Cisco Systems Fog Computing Mobile Computing at the Edge STC 2015 11 © 2015 Carnegie Mellon University Cyber-Foraging to Proximate and Remote Resources If we assume that tsurrogate is less than tcloud, proximate surrogates are a better option from an energy consumption and latency perspective Mobile Computing at the Edge STC 2015 12 © 2015 Carnegie Mellon University Cyber-Foraging: The Present Forward-deployed surrogates located in single-hop proximity of mobile devices • Communication with the central core in many cases is only needed for provisioning Goal is to bring the cloud closer to the user Mobile Computing at the Edge STC 2015 13 © 2015 Carnegie Mellon University Computation Offload (Short Operations) Computation-intensive operations which if executed on a mobile device would take in the order of tens of seconds, but if offloaded could improve response time considerably Typically request-response, synchronous operations • Image, audio and video processing • Face detection and recognition PowerSense: Image Processing for Dengue Detection [17] • Speech recognition • Speech translation PowerSense • Antivirus/Anti-malware Features • Gaming (typically AI-based) • User-Guided Systems can make runtime decisions on Runtime whether or not to offload computation Partitioning PowerSense leverages • Resource- • Equivalent code on mobile device and Microfluidic paper-based surrogate analytical devices (μPADs) [17] Adapted Input Mobile Computing at the Edge STC 2015 14 © 2015 Carnegie Mellon University Computation Offload (Long Operations) Computation-intensive operations 3DMA Features [18] which if executed on a mobile device would take minutes to hours, but if • Offload requests are placed in a offloaded could improve response space, processed on the surrogate, and results placed in time considerably the same space. Typically asynchronous operations to • When a device becomes avoid blocking the application disconnected, it waits until a connection is restored, and then • Service-based applications reads all available messages • Workflow-based applications (results) from the space • Search-based applications Mobile device may lose contact with a surrogate before the operation finishes • Caching data until the mobile device is reconnected • Alternative communication mechanisms to reach the mobile device Mobile Computing at the Edge STC 2015 15 © 2015 Carnegie Mellon University Low Coverage Environments Resource-challenged environments: Less-privileged regions characterized by limited Internet access, limited electricity and network access, and potentially low levels of literacy can leverage surrogates to obtain information to support their communities Field operations: People that spend time away from their main offices or labs, such as researchers, medics, and sales personnel, can leverage portable surrogates to support their computation and data needs AgroTempus Features [18] • Surrogates in villages download and cache data from mobile hub • Surrogates upload field- collected data to the mobile hub which eventually syncs with the cloud AgroTempus: Agricultural Knowledge Exchange in Resource- Challenged Environments [19] Mobile Computing at the Edge STC 2015 16 © 2015 Carnegie Mellon University Hostile Environments Characterized by very dynamic environments in which disconnected operations (or occasionally- connected operations) between surrogates and the cloud, and between mobile devices and surrogates, are highly likely Tactical Cloudlet Packaged [19] Capabilities Central Core Data Features (Service (Enterprise Cloud) Sources VMs) • Pre-Provisioned Cloudlets with App Low-bandwidth, High-bandwidth, intermittent connection stable connection Store for opportunistic data for pre-provisioning synchronization • Standard Packaging of Service VMs Deployment in • Optimal Cloudlet the field Selection Tactical Cloudlet Tactical Cloudlet • Cloudlet Management Single-Hop Component Network Tactical Cloudlets [19] • Cloudlet Handoff/Migration Mobile Devices Mobile Computing at the Edge STC 2015 17 © 2015 Carnegie Mellon University Data-Intensive Mobile Apps 1 Mobile device specifies Rely on large sets of data to interest in changes to specific provide their functionality — parts of web pages data typically resides in data Surrogate polls the web centers or in the enterprise servers involved, and if relevant changes have cloud occurred, it aggregates the • Mobile cloud applications updates as one batch that is sent to the mobile device • Online gaming Edge Proxy: Web Page [20] • Apps in data-rich domains Monitoring Design goals Edge Proxy Features [20] • Display of prioritized/relevant • In-Bound Pre-Processing: Mobile device is information only notified of changes • Query efficiency • Alternate