Clouds as Enablers for M2M Service Provisioning Roch Glitho, PhD Associate Professor and Canada Research Chair Concordia University, Montreal, Canada and Monique Morrow Distinguished Consulting Engineer, Cisco
Agenda 1. Framing the Problem: The Big Picture 1. M2M service provisioning problem and why cloud computing a powerful enabler for solving it 2. Early research on cloud based approaches for M2M service provisioning and limitations 3. Proposed cloud based architecture for M2M service provisioning
Vision of IoT and M2M: Trillion of End Points Converging Evolution of IP and Mobile Networks The Evolution of IP Networks Trillion End Points The Evolution of Mobile Networks Million PC / Laptop Billion IP Device Office & Home From RFID to Smart Sensors to Machines to More Complex Objects (e.g., a car) Billion Mobile Voice, Internet, SMS One The Common Evolution IT Infrastructure: of The Ultimate Convergence IP Networks of Enterprise, Service Provider, and Data Center Technologies Million Voice Source: Cisco (maalam@cisco.com)
IoT Evolution Beyond Connectivity Complex Applications Simple Applications Compound Applications Source: Harbor Research, 2012
M2M Service provisioning problem and why cloud computing is a powerful enabler On M2M Service Provisioning Current approaches to M2M service provisioning Advantages expected from cloud based approaches
M2M On M2M Service Provisioning: Basics Interactions between devices (e.g. wireless sensors, robots) with little or no human intervention M2M Service provisioning Process that brings to life the numerous applications expected from the interactions between M2M devices (e.g. demand/response in smart grids, health monitoring, surveillance) M2M service provisioning: M2M applications development, deployment and management
On M2M Service Provisioning: The Problem
Current Approaches to M2M Service Provisioning 1. Proprietary silo solutions No interoperability No re-use 2. A few early standards (e.g. OGC, IETF CoAP) that do not yet 2. A few early standards (e.g. OGC, IETF CoAP) that do not yet allow plug-and-play at all stages of the provisioning process
Cloud Computing as Enabler Cloud Computing could help move from the current proprietary silo approach to an horizontal plug and play approach building on the early standards (e.g. CoAP) and developing new standards!!! Expected benefits: efficient use of M2M resources, easy M2M applications - development, management and access.
Early Research on Cloud Based M2M Service Provisioning A Use Case The drawbacks
A Use Case: Managing Wearable Sensor Data through Cloud C. Doukas and I. Maglogiannis, Managing Wearable Sensor Data through Cloud Computing, 2011 Third International Conference on Cloud Computing Technology and Science
Managing Wearable Data Through Cloud
Managing Wearable Data Through Cloud
Managing Wearable Data Through Cloud
The Drawbacks Early research on cloud based M2M service provisioning focuses on how M2M applications could re-use resources that reside in clouds Drawbacks Do not generally enable An efficient use of M2M resources through virtualization (IaaS) An easy development and management of M2M applications (PaaS) An access anywhere, anytime of the developed M2M applications
Proposed cloud based M2M service provisioning architecture General IoT/M2M End to End Architecture Envisioned Cloud Based Service Provisioning architecture Open research issues and issues for standardization
IoT/M2M End-to-End Architecture Business Applications Applications mhealth/ Fleet Management / Industrial Automation Business Services SAP/ Oracle Systems Integ grator Ca arrier SP M2M Cl loud Services Networks Apps Development Service Maintenance & Customer Support Compliance Mgr Network Intelligence SDK & Open API Service Delivery Platform Asset Mgmt Transport Optimization EPC Connectivity Service Device Mgmt Big Data Mgmt Identity Mgmt EPC Integration Service Security Service Billing Service Policy Manager HTTP RESTful SOAP XML XMPP OneAPI Network, subscriber and Services Abstraction and Orchestration Layer EPC / BNG Remote Monitoring Service Personalization Manager Data Analytics Real-Time Subscriber Mgmt Network- Aware VAS Intelligent M2M Network Services Asset (Sensor) & M2M GW (Edge Computing 2G/3G Modem Sensor # 1 2G/3G Modem Sensor # 2 2G/3G Modem Sensor # n Data Analytics & Apps Orchestration Layer 3G/4G/WiFi/GPS M2M GW Wireless/Wireline M2M GW
Envisioned cloud based - architecture
Envisioned cloud based architecture M2M applications re-used functionality / applications offered as SaaS. However, unlike the current state of the art M2M applications are offered as SaaS, developed an managed using a PaaS, an furthermore their resource are used optimally thanks to virtualization.
A Use Case Let us consider a robotic application offered as SaaS, used by doctors, nurses and other hospital staff for goods / materials transportation and delivery within the hospital locate and (i.e goods / material are transported by robots) Orders are placed using smart phones, computers, tablets or even the keyboards that may be on some robots - Users can submit requests using text of voice Some examples Kitchen staff load robots with meals and order the transportation of those meals to various patients rooms A doctor orders through a smart phones, transportation of the radiology results of her patients from the radiology department to her office, ready for her to check the next day.
A Use Case Since orders can be placed by voice, a speech recognition application can be used by the robots and this application might reside in a separate cloud where it is offered by a third party as SaaS A knowledge data base offered by a third party as a SaaS in another cloud might also be used by the robots to learn how to load and unload The robotic application that enables transportation and delivery is offered as SaaS and developed and managed using a PaaS The fleet of robots that are used are virtualized to enable efficiency
Summary: Open Research Issues and Items for Standardization Research issues M2M devices / network virtualization Specialized M2M apps engine Modelling of M2M applications as SaaS Inter-cloud interactions (i.e. interactions between M2M applications and resources residing in a different cloud) Examples of items for standardization M2M PaaS and M2M IaaS interactions Programmatic interfaces for exposing M2M applications Inter-cloud interfaces