SlideShare a Scribd company logo

Accelerated adoption of Internet of Things (IoT) with In-network computing and Cloud

•
•
Infosys
Infosys

In-network computing gives you the ability to compute at a particular point in the network where it can deliver maximum value. This opens new avenues of how applications and services are conceptualized or implemented, harvesting the benefits of distributed computing. In-network computing has significant benefits for the network infrastructure as it improves latency for end user/ devices while it also reduces the network traffic to a great extent. Emerging technologies like IoT and its application can immensely benefit by using In-network computing technology in conjunction with cloud technologies.

1 of 15
Download to read offline
Accelerated adoption of IoT with In-network computing and Cloud Mitesh Patel Group Project Manager, Head – Internet of Things (IoT) Manufacturing
InternetofThings Network connectivity | Low latency | High availability | Low cost mobility Everyonewantstojumponthebandwagon andthisisjustthebeginning 2
Time Challenges Tolerance to disruption in service /connectivity Internet traffic Low High …Computinginfrastructureisgettingstretched Volume Variety Velocity 3
Real time Apps Big Data Apps UX …. In-network computing device Compute Storage Network Healthcare Devices Edge network Internet Data Center In- network computing Manufacturing Buildings MiningLogistics OilResidencesRetailAgriculture Cloud computing Manufacturing Buildings MiningLogistics HealthcareOilResidencesRetailAgriculture X ms <<X ms In-networkcomputing&cloudcomputing 4
• Compute near the edge of the network / close to actual devices • Smaller, less powerful devices can do complex jobs • Opportunity to convert compute intensive application to internet services • Storage capability at the edge along with computing makes it a self sustained ecosystem reducing external dependencies Data Center Network Core Edge of the network In-networkcomputingwilloffloaddatacentercomputing 5
Data Center / Cloud Core Multi-Service Edge Embedded Systems and Sensors Centralized Intelligence Network Fabric End Point Intelligence Easingnetworkdemandwillcreatespaceformoredevices 6 • Reduced traffic volume in higher layers of the network • Improved latency response • Reduce availability stress on primary data center • Network load shaping lot easier • Resilient network, no single point failure except for last leg of network • Build alternate paths • Foundation for SDN (Software Defined Networks)
Advantages: • Network as a virtualized service • Extremely light edge / application devices • End devices can leverage compute, storage, network as a service • Multiple service provider tenants on a single device Networking services Wired / wireless front end Compute Storage Real time OS Real time Applications Device Management Services Up link Down link Application n Application 2 Application 1 Virtualization In-networkcomputingdeviceswillbehighlyvirtualized 7
In-network computing can be leveraged for several applications Predictive Maintenance Enable New Knowledge Agriculture Energy Saving Transportation and Connected Vehicles Smart City Intelligent Buildings Enhance Safety and Security Healthcare Industrial Automation Smart HomeSmart Grid 8
In-network computing will increase IoT Adoption Focus will shift from device capability to network capability Multi tenant devices for out of box industry specific solutions will be possible Enhanced network security as ingest points become smarter Foundation for SDN (Software Defined Networks) implementation Smarter, smaller & cheaper end devices as they leverage In-network computing Enable communication without dependency on core systems, will create more natural ways of communication Systems tolerant to disruption in connectivity 9
Expectations from technology • High availability & low latency networks, even with large consumer base • Long term technology • Infrastructure to enable regulatory compliance • Tamper proof & IT security Peak Clipping Conversation Load Building Valley Filling Flexible Load Shape Load Shifting Smartgridshavelottoofferbuttherearechallenges Demand management is key for energy companies 10
Renewable energy is changing the grid dynamics with more and more end users becoming power producers (Client = Consumers + Producers). Further proliferation in renewable energy sources will cause accelerated change in this phenomenon Renewable power needs to be handled differently, its dependency on external / environmental factors, consumption, generation, storage will generate a demand for computing infrastructure to manage such users / producers. In-network computing has the ability to address this Smartgrids:Newbreedofclients(consumers+producers) 11
• Connectivity, communication, and computation with low latency • Ability to deploy knowledge & algorithm based solution and upgrade it on the fly in the field • Client = consumer + producers, will make contracting, billing complex and will require close loop and advanced monitoring of distribution • Real time analytics to match demand and supply will be key success factor for distribution companies to succeed In-networkcomputingcanaddressSmartgridsneeds 12
In-network computing can simplify Smart Factory implementation Un-attended Production Cell CNC Machine QC center In ~ . Industrial Network Make -> Check -> Go Availability & Capability based dynamic Job routing Smart Factory Shop floor • PLC / DCS / SCADA • Safety systems • Process control • Motion control • Telemetry • End user selectable prioritization of network traffic • Intense computing application like condition monitoring can compute & store data locally / on edge devices • Ready to use generic template based on specific domains deployable on edge devices • Resource and energy consumption monitoring and control • Ability to offer out of box features 13
Ethernet / Wireless Enterprise Network Business PlantPlant Control Network • Event processing • Time synchronization & triggering • Segregation of priority traffic like safety, compute and act locally • Workflow & control system • Mission critical applications Ethernet / Wireless single network connectivity for all needs • Motion control • Quality & operations • Safety • Condition monitoring • Workflow In-networkcomputing canreducenetworkcomplexityfor SmartFactory 14
© 2014 Infosys Limited, Bangalore, India. All Rights Reserved. Infosys believes the information in this document is accurate as of its publication date; such information is subject to change without notice. Infosys acknowledges the proprietary rights of other companies to the trademarks, product names and such other intellectual property rights mentioned in this document. Except as expressly permitted, neither this documentation nor any part of it may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, printing, photocopying, recording or otherwise, withoutthe prior permissionof Infosys Limitedand/ or any named intellectual property rights holders under this document. Thank You

More Related Content

Accelerated adoption of Internet of Things (IoT) with In-network computing and Cloud

  • 1. Accelerated adoption of IoT with In-network computing and Cloud Mitesh Patel Group Project Manager, Head – Internet of Things (IoT) Manufacturing
  • 2. InternetofThings Network connectivity | Low latency | High availability | Low cost mobility Everyonewantstojumponthebandwagon andthisisjustthebeginning 2
  • 3. Time Challenges Tolerance to disruption in service /connectivity Internet traffic Low High …Computinginfrastructureisgettingstretched Volume Variety Velocity 3
  • 4. Real time Apps Big Data Apps UX …. In-network computing device Compute Storage Network Healthcare Devices Edge network Internet Data Center In- network computing Manufacturing Buildings MiningLogistics OilResidencesRetailAgriculture Cloud computing Manufacturing Buildings MiningLogistics HealthcareOilResidencesRetailAgriculture X ms <<X ms In-networkcomputing&cloudcomputing 4
  • 5. • Compute near the edge of the network / close to actual devices • Smaller, less powerful devices can do complex jobs • Opportunity to convert compute intensive application to internet services • Storage capability at the edge along with computing makes it a self sustained ecosystem reducing external dependencies Data Center Network Core Edge of the network In-networkcomputingwilloffloaddatacentercomputing 5
  • 6. Data Center / Cloud Core Multi-Service Edge Embedded Systems and Sensors Centralized Intelligence Network Fabric End Point Intelligence Easingnetworkdemandwillcreatespaceformoredevices 6 • Reduced traffic volume in higher layers of the network • Improved latency response • Reduce availability stress on primary data center • Network load shaping lot easier • Resilient network, no single point failure except for last leg of network • Build alternate paths • Foundation for SDN (Software Defined Networks)
  • 7. Advantages: • Network as a virtualized service • Extremely light edge / application devices • End devices can leverage compute, storage, network as a service • Multiple service provider tenants on a single device Networking services Wired / wireless front end Compute Storage Real time OS Real time Applications Device Management Services Up link Down link Application n Application 2 Application 1 Virtualization In-networkcomputingdeviceswillbehighlyvirtualized 7
  • 8. In-network computing can be leveraged for several applications Predictive Maintenance Enable New Knowledge Agriculture Energy Saving Transportation and Connected Vehicles Smart City Intelligent Buildings Enhance Safety and Security Healthcare Industrial Automation Smart HomeSmart Grid 8
  • 9. In-network computing will increase IoT Adoption Focus will shift from device capability to network capability Multi tenant devices for out of box industry specific solutions will be possible Enhanced network security as ingest points become smarter Foundation for SDN (Software Defined Networks) implementation Smarter, smaller & cheaper end devices as they leverage In-network computing Enable communication without dependency on core systems, will create more natural ways of communication Systems tolerant to disruption in connectivity 9
  • 10. Expectations from technology • High availability & low latency networks, even with large consumer base • Long term technology • Infrastructure to enable regulatory compliance • Tamper proof & IT security Peak Clipping Conversation Load Building Valley Filling Flexible Load Shape Load Shifting Smartgridshavelottoofferbuttherearechallenges Demand management is key for energy companies 10
  • 11. Renewable energy is changing the grid dynamics with more and more end users becoming power producers (Client = Consumers + Producers). Further proliferation in renewable energy sources will cause accelerated change in this phenomenon Renewable power needs to be handled differently, its dependency on external / environmental factors, consumption, generation, storage will generate a demand for computing infrastructure to manage such users / producers. In-network computing has the ability to address this Smartgrids:Newbreedofclients(consumers+producers) 11
  • 12. • Connectivity, communication, and computation with low latency • Ability to deploy knowledge & algorithm based solution and upgrade it on the fly in the field • Client = consumer + producers, will make contracting, billing complex and will require close loop and advanced monitoring of distribution • Real time analytics to match demand and supply will be key success factor for distribution companies to succeed In-networkcomputingcanaddressSmartgridsneeds 12
  • 13. In-network computing can simplify Smart Factory implementation Un-attended Production Cell CNC Machine QC center In ~ . Industrial Network Make -> Check -> Go Availability & Capability based dynamic Job routing Smart Factory Shop floor • PLC / DCS / SCADA • Safety systems • Process control • Motion control • Telemetry • End user selectable prioritization of network traffic • Intense computing application like condition monitoring can compute & store data locally / on edge devices • Ready to use generic template based on specific domains deployable on edge devices • Resource and energy consumption monitoring and control • Ability to offer out of box features 13
  • 14. Ethernet / Wireless Enterprise Network Business PlantPlant Control Network • Event processing • Time synchronization & triggering • Segregation of priority traffic like safety, compute and act locally • Workflow & control system • Mission critical applications Ethernet / Wireless single network connectivity for all needs • Motion control • Quality & operations • Safety • Condition monitoring • Workflow In-networkcomputing canreducenetworkcomplexityfor SmartFactory 14
  • 15. © 2014 Infosys Limited, Bangalore, India. All Rights Reserved. Infosys believes the information in this document is accurate as of its publication date; such information is subject to change without notice. Infosys acknowledges the proprietary rights of other companies to the trademarks, product names and such other intellectual property rights mentioned in this document. Except as expressly permitted, neither this documentation nor any part of it may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, printing, photocopying, recording or otherwise, withoutthe prior permissionof Infosys Limitedand/ or any named intellectual property rights holders under this document. Thank You