on research challenges in iot systems...
TRANSCRIPT
![Page 1: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/1.jpg)
On Research Challenges in IoT Systems Engineering
16 August 2018, USC
Schahram Dustdar
Distributed Systems GroupTU Wien
dsg.tuwien.ac.at
![Page 2: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/2.jpg)
Ecosystems of People, Systems, and Things
2
![Page 3: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/3.jpg)
3
![Page 4: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/4.jpg)
4
![Page 5: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/5.jpg)
5
![Page 6: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/6.jpg)
6
![Page 7: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/7.jpg)
Network Neutrality (NN)All traffic on the Internet must be treated equally.
7
![Page 8: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/8.jpg)
Real cases
8
![Page 9: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/9.jpg)
Understanding the context
9
![Page 10: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/10.jpg)
What about the Internet of Things (IoT) ?
10
![Page 11: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/11.jpg)
Think Ecosystems: People, Systems, and Things
Diverse users with complex networked dependencies and intrinsic adaptive behavior –has:
1. Robustness & Resilience mechanisms: achieving stability in the presence of disruption
2. Measures of health: diversity, population trends, other key indicators
11
![Page 12: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/12.jpg)
HVAC (Heating, Ventilation, Air Conditioning) Ecosystem
![Page 13: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/13.jpg)
Water Ecosystem
![Page 14: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/14.jpg)
Air Ecosystem
![Page 15: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/15.jpg)
Fleet Management Systemo Manages fleets of electric vehicles world-
wide (e.g., on golf courses)
Motivating Case StudiesBuilding Management System• Manages building facilities, e.g., HVAC
systems, elevators and emergency alarms
15
IoT Gateways Cloud
![Page 16: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/16.jpg)
Motivation• Lack of systematic support and tools for
developing and operating IoT Cloud systems
• Today IoT Cloud systems are vertically closed and tightly coupled
• Hard to develop and maintain applications• Difficult to operate and reuse existing
infrastructure
16
![Page 17: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/17.jpg)
Programming Model for IoT Cloud Applications
17
![Page 18: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/18.jpg)
Motivation
18
Requirements:• Application: Should be generic
(independent of underlying devices)• Runtime: Dealing with scalability and
elasticity concerns• Developer: Software engineering
expertise
Requirements:• Application: Custom configuration and
behavior of Sens./Act.• Runtime: Dealing with constrained
resources• Developer: Domain expert knowledge
Fleet energy usage management Process energy consumption - Detect energy fault- Notify manager- Stop vehicleActuation steps to stop vehicle
![Page 19: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/19.jpg)
Approach
19
ControlTasks
MonitorTasks
Intents
E.g., sequence of actuation steps to stop a vehicle
• Packaged into domain-specific libraries (e.g., vehicles management)
Task - Encapsulates domain-dependent controls or analytics
• Used by developers to remotely invoke Tasks
• Independent of concrete Task implementation
Stefan Nastic, Sanjin Sehic, Michael Vögler, Hong-Linh Truong, and Schahram Dustdar. PatRICIA - A Novel Programming Model for IoT Applications on Cloud Platforms. SOCA 2013. Hawaii, USA.
Intent - High-level representation of Tasks on Cloud platforms
![Page 20: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/20.jpg)
21
ControlTasks
MonitorTasks
Intents
![Page 21: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/21.jpg)
Intent-based Programming Model
22
o Trade expressiveness for more flexible and easier application development
Intent Structure
• Passive data structure which declaratively describes intended action, e.g., stop vehicle
• Generic applications (What needs to be done instead how to do it)
• Enable developing loosely coupled applications
![Page 22: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/22.jpg)
Provisioning solutions for Smart Cities“abstractions, concepts and processes“
23
![Page 23: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/23.jpg)
Research Challenges Design Principles Main Enablersaddressed by
implemented by
Multi-level provisioning workflows
Cloud-based dependency resolution
Software-defined gateways
Cloud-based controller
Centralized infrastructure API management
Enable fine-grained resource consumption
Enable resource monitoringDevice profiler
Support API encapsulation of the infrastructure resources
Enable managed configuration models
Enable automation of provisioning processes
Support elastically scalable provisioning processes
Provide more autonomy to the edge resources
Enable dynamic feature composition
Edge-compatible provisioning agents
Flexible delpoyment and provisioning models
Software-defined IoT topology (complex units)
Configurations container
On-demand, self-service usage models
Unified representation of heterogeneous resources
Cost-awareness
Efficient provisioning models
Flexible customization of tightly coupled resources
Logically centralized point of operation
Support for elasiticty concerns
Research Challenges Overview
24
![Page 24: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/24.jpg)
25
Motivation
Current provisioning solutions:
o Require on site presenceo Require manual interaction with deviceso Not suitable for resource constrained
devices
How to efficiently provision IoT Cloud applications?
Application
Consider provisioning a simple application for monitoring environmental conditions in Smart City buildings
![Page 25: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/25.jpg)
Approach: Core principles
• From physically isolated, rigid Edge/IoTinfrastructure to virtualized, elastic IoT Cloud, by utilizing “software-defined“ principles.
• From task-specific solutions to fully-fledged ecosystem and management processes, based on DevOps best practices.
26
![Page 26: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/26.jpg)
Approach: Design Principles and Main Enablers
Uniform representation of IoT infrastructure
API encapsulation of infrastructure resources
Software-Defined Gateways
SDG SDG SDGLogically centralized operation
Provisioning Middleware
Automation of provisioning processes
27
![Page 27: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/27.jpg)
28
Approach
Provisioning Middleware
SDG SDG SDG
![Page 28: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/28.jpg)
Software-Defined Gateways – Overview
29
Stefan Nastic, Sanjin Sehic, Le-Duc Hung, Hong-Linh Truong, and Schahram Dustdar. Provisioning Software-defined IoT Cloud Systems. FiCloud 2014. Barcelona, Spain.
Physical Gateway
SDGSDG
Application …• Execute atop physical gateways• Virtualize gateways compute and memory
resources
• Act as isolated containers for applications => lightweight execution environment
• Enable on-demand provisioning of application, libraries and configuration models
![Page 29: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/29.jpg)
Software-Defined Gateways – Provisioning Model
30
Provisioning Model (partial view)
• Independent of the underlying virtualization technology
• Built from SDG prototypes• Based on kernel-supported virtualization: LXCs,
libvirt-sandbox, chroot, etc. • Preconfigured with different functionalities, e.g.,
monitoring mechanisms
• Expose provisioning APIs used to deliver complex functionality
• SDG IoT Units enable encapsulating application components, libraries and configuration models
Provisioning Model (partial view)
![Page 30: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/30.jpg)
SDGs Ecosystem• Hierarchical structure of SDG components and capabilities.• Enables distributing SDGs & SDG IoT Units in a market-like
fashion, e.g., via SDG AppStore.
31
Atomic software-defined IoT units
Custom proc. logic
IoT datastorage
Communication
In-memory image
VPNMessaging
Sand box
Networkoverlay ProtocolVolatile
History
Key/Value store
SecurityData
quality
Outliersfilter
IoT compute
GW runtime
Data pointcontroller
CEPComponent
-model
Elasticity
Auto scaling group controller
Functionalcapabilities
Non-functionalcapabilities
...
...
Monitor.
Config.
Hierarchy of basic SDG components (partial view)
![Page 31: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/31.jpg)
Software-Defined Gateways – Architectural View
32
• Provisioning Agent• Handling remote provisioning
requests• Dynamically downloading
application components• Local installation of application
components
![Page 32: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/32.jpg)
SDG Prototype(e.g., LXC based)
Software-defined Gateway - Example
33
Sensor Filter
SDG API
Com.Client
API
API
ConfigurationModel
Costfunction
API
Software-Defined Gateway
Ready to be deployed on IoT devices such as physical gateways or cloudlets
![Page 33: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/33.jpg)
SDG SDG SDG
34
C2: Approach
Provisioning Middleware
![Page 34: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/34.jpg)
Provisioning Middleware
35
Cloud
Edge device/IoT Gateway
![Page 35: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/35.jpg)
Example of SDG-driven Provisioning Process
36
Gateway Gateway Gateway Gateway
Cloud
IoT
Gateway
Provisioning Controller
API Call
SDG
API Call
![Page 36: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/36.jpg)
SDG Delivery Models
37
GatewayDevice
SDG
CDN
GatewayDevice
SDG
Cloudlet
Download
Script
1.2.
3.SDG‘
GatewayDevice
SDG Well knownRepository
SDG‘
Download
GatewayDevice
SDG
Push Pull
![Page 37: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/37.jpg)
IoT & Data Science – Research Challenges
38
Ranjan R., Rana O., Nepal S., Yousif M., James P., Wen Z., Barr S., Watson P., Jayaraman P. P., Georgakopoulos D., VillariM., Fazio M., Garg S., Buyya R., Wang L., Zomaya A. Y., Dustdar S. (2018).The Next Grand Challenges: Integrating the Internet of Things and Data Science,IEEE Cloud Computing, Volume 5, Issue 3, pp. 12-26
![Page 38: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/38.jpg)
IoT-driven ecosystems
![Page 39: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/39.jpg)
IoT/Data/Application Orchestration
![Page 40: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/40.jpg)
Osmotic Computing
Dynamic management of (micro)services across cloud and edge datacenters deployment, networking, and
security, … providing reliable IoT support with
specified levels of QoS.
In chemistry, “osmosis” represents the seamless diffusion of molecules from a higher to a lower concentration solution.
Villari M., Fazio M., Dustdar S., Rana O., Ranjan R. (2016). Osmotic Computing: A New Paradigm for Edge/Cloud Integration. IEEE Cloud Computing, Volume 3, Issue 6, pp. 76-83
![Page 41: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/41.jpg)
Osmotic movement of MELs in Clouds, Edge, Things
Legend:MEL...Micro Element
![Page 42: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/42.jpg)
IoT Data Sources1. Representation: Structure and represent the data to facilitate multiple modalities,
exploiting the complementarity and redundancy of different data sources.
2. Translation: Interpret data from one modality to another, i.e., provide a translator that allows the modalities to interact with each other for enabling data exchange.
3. Alignment: Identify the relation among modalities. This requires identifying links between different types of data.
4. Fusion: Fuse information from different modalities (e.g., to predict).
5. Co-learning: Transfer knowledge among modalities. This explores the field of how the knowledge of a modality can help or enhance a computational model trained on a different modality.
![Page 43: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/43.jpg)
IoT Computational Unit1. MicroServices (MS), which implement specific functionalities and can be deployed and migrated across
different virtualized and/or containerized infrastructures (e.g., Docker) available across Cloud, Edge, and Things layers
2. MicroData (MD), encodes the contextual information about (a) the sensors, actuators, edge devices, and cloud resources it needs to collect data from or send data to, (b) the specific type of data (e.g., temperature, vibration, pollution, pH, humidity) it needs to process, and (c) other data manipulation operations such as where to store data, where to forward data, and where to store results
3. MicroComputing (MC), executing specific types of computational tasks (machine learning, aggregation, statistical analysis, error checking, and format translation) based on a mix of historic and real-time MD data in heterogeneous formats. These MCs could be realized using a variety of data storage and analytics programming models (SQL, NoSQL, stream processing, batch processing, etc.)
4. MicroActuator (MA), implementing programming interfaces (e.g., for sending commands) with actuator devices for changing or controlling object states in the IoT environment
![Page 44: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/44.jpg)
IoT Programming Patterns needed
1. Decomposing IoT data analysis activities into fine-grained activities (e.g., statistics, clustering, classification, anomaly detection, accumulation, filtering), each of which may impose different planning and run-time orchestration requirements;
2. Identifying and integrating real-time data from IoT devices and historical IoT data distributed across Cloud and Edge resources;
3. Identifying data and control flow dependencies between data analysis activities focusing on coordination and data flow variables, as well as the handling of dynamic system updates and re-configuration;
4. Defining and tagging each data analysis activity with runtime deployment constraints(QoS, security and privacy).
![Page 45: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/45.jpg)
Grand Challenges Manifesto
Ranjan R., Rana O., Nepal S., Yousif M., James P., Wen Z., Barr S., Watson P., Jayaraman P. P., Georgakopoulos D., VillariM., Fazio M., Garg S., Buyya R., Wang L., Zomaya A. Y., Dustdar S. (2018).The Next Grand Challenges: Integrating the Internet of Things and Data Science,IEEE Cloud Computing, Volume 5, Issue 3, pp. 12-26
![Page 46: On Research Challenges in IoT Systems Engineeringcci.usc.edu/wp-content/uploads/2018/08/On-Research... · Research Challenges addressed Design Principles Main Enablers by implemented](https://reader034.vdocuments.site/reader034/viewer/2022042310/5ed7aae35ab86c3b0273a141/html5/thumbnails/46.jpg)
Thanks for your attention
Prof. Schahram Dustdar Member of Academia Europaea IBM Faculty awardACM Distinguished ScientistIEEE Fellow
Distributed Systems GroupTU Wien
dsg.tuwien.ac.at
47