Cloud and Edge Computing Systems

Cloud and Edge Computing Systems
Sayed Chhattan Shah
Associate Professor
Department of Information Communications Engineering
Hankuk University of Foreign Studies Korea
www.mgclab.com

Contents
 IoT Systems
 IoT and 5G Applications
 Mobile Cloud Computing
 Edge and Fog Computing
 Mobile Ad hoc Cloud
 Private Edge Cloud
o System and network architecture
o Research challenges and research directions
o Middleware platform for a private edge cloud
 Conclusion

The Internet of Things
 IoT is simply a concept wherein machines and everyday objects are
connected via Internet
o The thing refers to all the things that can be connected to Internet
 Door locks
 Lights
 Household appliances
 Car
 Clothes
IoT refer broadly to extension of network
connectivity and computing capability to
objects, devices, sensors, and items not
ordinarily considered to be computers

 The architecture of IoT is divided into three basic layers
o Perception layer is used to collect data
o Network layer provide data transmission services
o Application layer deliver application specific services to users
Personal area
network
Local area
network
Wide area
network
Bluetooth
Bluetooth Low
Energy
ZigBee
Wi-Fi
Wi-Fi
WiMAX
WiMAX
LoRa
4G (LTE)
5G
Enabling network technologies

Source: GIV 2025 Unfolding the Industry Blueprint of an Intelligent World
 More than 9 billion IoT devices are connected to the Internet
 Number of IoT connected devices is projected to increase to 43 billion by
2023

500 zettabytes of data produced by people and things
Credit Cisco Global Cloud Index
5 Gigabyte data every second 1 Gigabyte data every second
Few bits data every second

 Apple Smart Watch
o GPS
o ECG
o Blood oxygen level
o All-day activity tracking and sleep monitoring
o High and low heart rate notifications
https://www.apple.com/lae/watch/
IoT and 5G Network Applications

 Smart Diapers
o A diaper and smartphone app for monitoring child's health.
o Diapers have patches at the front with several colored squares that change color as
they react to different compounds, such as water content, proteins or bacteria.
o Sensors read the data and send it to a physician.
http://iotlineup.com/

 Smart Home
o Several environmental, video, audio, and bio sensors are deployed to observe the
home environment and physiological health of an individual
o The data collected by sensors are sent to an
application where numerous algorithms for
emotion and sentiment detection, activity
recognition and situation management are
applied to provide healthcare- and
emergency-related services and to manage
resources at the home
Credit CC0 Public Doma

o The collected data is also used to detect and
track mobile and stationary targets.
o The whole process involves sophisticated
image and video processing algorithms.
 Mobile Intelligent Video Surveillance System
o Mobile robots and micro drones equipped with audio-video and environmental
sensors are deployed to collect data which is then processed to construct a three
dimensional map of environment in real time.

Use case Latency Data rate
Factory Automation 0.25 - 10 ms 1 Mbps
Process Automation 50 - 100 ms 1-100 Mbps
Intelligent Transport Systems 10 - 100 ms 10 - 700 Mbps
Robotics and Telepresence 1 ms 100 Mbps
Virtual and Augmented Reality 1 ms 1 Gbps
Health Care 1-10 ms 100 Mbps
Serious Gaming 1 ms 1 Gbps
Smart Grid 1 - 20 ms 10 - 1500 Kbps
Education and Culture 5 - 10 ms 1 Gbps
Latency and data rate requirements of IoT and 5G applications

Resource intensive applications Non-resource intensive applications
Non-real time applications
Real time applications
Applications
Data intensive
Computationally intensive

 Resource intensive and real-time IoT and 5G network applications such as smart
home and intelligent video surveillance have diverse requirements
o Low latencies
o High data rates
o Access to sensors and actuators or IoT devices
o Significant amount of computing and storage resources

 To address the requirements of these applications, several cloud and edge
computing systems, such as cloudlet computing, mobile edge computing, and fog
computing, have been proposed

Cluster Grid Cloud
Distributed System
A collection of interconnected
computers cooperatively work together
as a single integrated computing
resource

Cloud Computing
Everything — from computing power to
computing infrastructure and applications are
delivered as a service
Basic idea
Computing resources should be available on
demand for a fee just like the electrical
power grid

Mobile devices are integrated with cloud computing systems through an
infrastructure-based communication network [1]
Mobile Cloud Computing
Service provider nodes
Service consumer nodes

Enabling Technologies
Image by CNET

Benefits
 Improved data storage capacity and processing power
o Users can execute computationally and data-intensive applications on mobile devices
 Extended battery life
 Improved reliability

Cloud Robotics
 Robots rely on a cloud-computing infrastructure to access vast amounts of
processing power and data
 Execution of computationally intensive
tasks on cloud would result in cheaper
lighter and easy-to-maintain hardware
 Shared library of
 objects
 algorithms
 skills

Sensor Cloud
Sensors are integrated with cloud computing systems

Computation Offloading
Migrating computation to more resourceful computers
Computation offloading = Surrogate computing = Remote execution

 Offloading approaches are classified based on various factors
o Why to offload
 Improve performance or save energy
o What mobile systems use offloading
 Smart phones
 Robots
 Sensors
o Infrastructures for offloading
 Cluster
 Grid
 Cloud

o Offloading decision criteria
 Bandwidths
 Server speeds
 Server loads
 Energy consumption
 Amounts of data exchanged between servers and mobile systems

o Types of offloading
 Partial
 Full
o Application partitioning
 Static
 Dynamic
o When to decide offloading
 Static
 Dynamic
o Offloading data-intensive tasks
 Knowledge on the amount of data to be processed
o Offloading small tasks
 May not improve performance or reduce energy consumption

Edge Computing
 Mobile cloud computing system
o High communication latencies
o Unnecessary traffic to the core Internet network
o Transmission energy consumption when data are transmitted via cellular networks

Edge Computing
 To overcome the drawbacks of the conventional cloud-based approach, a
technology called edge computing is proposed
 The implementation of edge layer can be classified into three types [2] [9] [10]
Mobile Edge Computing
Fog Computing
Cloudlet Computing

Edge Computing
 Mobile Edge Computing
o MEC proposes deployment of intermediate nodes with storage and processing
capabilities in the base stations of cellular networks thus offering Cloud
Computing capabilities inside the Radio Area Network
MEC for video stream analysis

Edge Computing
 Role of Mobile Edge Node or Server
o Computational power
o Storage space
o Data processing
o Data caching

An illustration of a multi-interface BS [4]
Multi-interface BS

Edge Computing
 Fog Computing
o Fog nodes can be deployed anywhere with a network connection
 Any device such as router or a video surveillance camera with computing and
network connectivity can be a fog node
https://internetofthingsagenda.techtarget.com/definition/fog-computing-fogging

Fog Computing
A Fog-based computing system
 Wireless heterogeneous devices
 Battery-powered
 Limited Computing Resource
 Spatially distributed Fog Nodes
 FN a small data center equipped with a
number of physical servers
interconnected by a LAN
 Delay-sensitive demands are processed
by FN
Internet Gateway
Router
 A set of powerful Cloud data centers
 A cloud node serves delay-tolerant
computing requests of FNs

Fog Computing
 Device-Fog communication
 Single-hop short-range
communication technology such as
Wi-Fi or Bluetooth
 Inter-clone communication
 In a same FN via a High speed LAN
such as Gigabit Ethernet
 Inter-Fog communication
 Medium-range wireless back-bone
that exploits broadband
transmission technologies such as
IEEE 802.11n or WiMAX
A Fog-based computing system
 IoT device may offload a task to serving FN
 Each device is associated to a virtual clone
that runs on a physical server hosted by the
serving FN

Cloudlet Computing
Cloudlet is a small cloud data center located close to an end device
https://elijah.cs.cmu.edu/

Mobile Cloud Computing and Edge Computing
 Mobile cloud computing
o High communication latencies
o Unnecessary traffic to the core Internet network
o Transmission energy consumption when data are transmitted via cellular networks
 Edge computing implementations
o Require new infrastructure or upgradation of existing infrastructure
o Does not exploit the capabilities of end devices

Mobile Ad hoc Cloud
Multiple mobile devices interconnected through a mobile ad hoc network are
combined to create a virtual supercomputing node or a small-scale cloud data
center [1] [6] [7] [8]

Cloud and Edge Computing Systems

Mobile Ad hoc Cloud Applications
 Autonomous Threat Detection in Urban Environments [8]
o A group of miniature autonomous mobile robots are deployed in urban
environments to detect and monitor a range of military and non-military threats
 Use sophisticated image and video processing algorithms
 Vision-based navigation algorithms to navigate in the environment

Mobile Ad hoc Cloud Applications
 Construction of 3D-Map and Identification of Static and Mobile Targets
within a Map [7] [8]
o A set of miniature unmanned aerial vehicles or mobile robots can be deployed
in a targeted area
 Broadcast live video streams
 Processed to construct map and identify mobile targets
o Requires huge processing power

Research Challenges
 Compared to traditional parallel and distributed computing systems such as
cloud mobile ad hoc cloud is characterized by
o Node mobility
o Limited battery power
o Low bandwidth and high latency
o Shared and unreliable communication medium
o Infrastructure-less network environment

Research Challenges
 Node Mobility
o Global node mobility
 Task Failure
o Local node mobility
 Increased data transfer times
o Mobility of an intermediate node
 Increased data transfer times
 May disconnect network
o Approaches
 Task migration
 Task reallocation
• In both cases delay due to reallocation or migration of task

Research Challenges
 Node Mobility
o To improve performance and avoid task failure or migration, nodes with long-
term connectivity are required for the allocation of tasks
o An effective and robust two-phase resource allocation scheme
 Exploit the history of user’s mobility patterns in order to select nodes that provide
long-term connectivity
o Location prediction schemes
 Use node’s direction and speed to predict future connectivity

Research Challenges
 Power Management
o Main sources of energy consumption are CPU processing memory and data
transmission in the network
o Key factors that contribute to transmission energy consumption
 Transmission power required to transmit data
 Communication cost induced by data transfers between tasks
o Most of the schemes are focused on the conservation of processing energy
o Energy efficient resource allocation scheme aims to reduce transmission
energy consumption and data transfer cost
 Basic idea is to allocate tasks to nodes that are accessible at minimum
transmission power

Research Challenges
 Constrained communication environment due to shared medium and mobility
o Low bandwidth and unstable connectivity problems
 Data transfer cost is very critical for application and system performance
 A few promising approaches to reduce data transfer costs
• Directional antennas
• Efficient medium access control
• Channel switching
• Multiple radios
 Parallel applications usually consist of a range of tasks
with varying bandwidth and processing constraints

Research Challenges
 Parallel programming model
o The traditional parallel programming models are not suitable for mobile ad
hoc environments due to high communication latencies and link failure and
activation ratios
o Actor-based programming model could be the possible candidate because it
deals quite well with high latencies, offers lightweight migration and can be
easily adopted to deal with node mobility

 Security risks
 Incentive mechanism
 Failure management
 Quality of Service support
 Standards for heterogeneous environments
 Development of simulation environment
 Architecture
Research Challenges

Resource Management System for a Mobile Ad hoc Cloud

 Middleware Layer
o Discovery, monitoring, allocation, distribution and migrations of application tasks
 Network Layer
o Ad hoc network and communication services to a middleware layer and shares collected information
 Development of cost estimation models
o Energy consumption estimation model
 Transmission energy consumption
 Data transfer cost
o Data transfer time estimation model
 Estimation of link quality
 Link lifetime
o Link lifetime prediction model
o Task completion time estimation model
Resource Management System for a Mobile Ad hoc Cloud

Heterogeneous private edge cloud system
 A heterogeneous private edge cloud system is a small-scale cloud data
center at a local physical area such as a home or an office
 It consists of various stationary and mobile devices interconnected
through a single or multiple infrastructure-based or infrastructure-less
wireless local area networks

The role of private mobile edge clouds in the overall system

 A heterogeneous private edge cloud system
o reduces data latency
o provides high throughput
o reduces traffic to the core Internet network
o reduces data privacy risks
o utilizes high end devices

Table. Mobile ad hoc cloud vs Private mobile edge cloud
Mobile
cloud computing
Mobile
edge computing
Fog computing Mobile ad hoc cloud Heterogeneous
private edge cloud
Mobile nodes
consume services
Mobile nodes
consume services
Mobile nodes consume
services
Mobile nodes provide
and consume services
Mobile nodes and
Stationary nodes
Provide and consume services
Infrastructure-based
Network
Network
Network
Ad hoc network Ad hoc networks
Infrastructure-based local area
networks
Plain heterogeneous
environment
Plain heterogeneous
environment
Plain heterogeneous
environment
Homogeneous
environment
Complex multi-network
heterogeneous environment
Interconnected to
central cloud
computing system
through wide area
networks
Interconnected to
cloud computing
system at network
edge through wide
area networks
Interconnected to cloud
computing system at
network edge through
local and wide area
networks
Isolated system Interconnected to edge
computing systems or central
cloud system through WAN
Suitable for
infrastructure-based
environments
Suitable for
environments
Suitable for
environments
Suitable for
infrastructure-less
environments such as
battlefield
Suitable for infrastructure-
based environments

 A private mobile edge cloud is an integration of several computing and
networking systems and technologies
o Cloud computing
o Mobile computing
o Edge computing
o Mobile ad hoc cloud computing
o Mobile ad hoc networking technologies
o Local and wide area networking technologies

Characteristics
 The heterogeneous private edge cloud system is characterized by
o Heterogeneous computing environment
o A complex and dynamic multilayer network infrastructure
https://blog-gn.dronacharya.info/index.php/iot-communication-protocols/

Research Challenges
 Heterogeneous computing environment
o Numerous heterogeneous devices
o The devices differ with respect to processor architecture, operating system,
execution environment, and speed
o Devices may offer from a simple service to a rich set of services

 Heterogeneous computing resources
o Uniform representation and control of heterogeneous devices
o Efficient discovery, registration, and monitoring of wide range of devices
and services
o Allocation of heterogeneous computing, sensing, and actuating resources to
emerging application tasks with a diverse quality of service requirements
o Execution or processing of tasks submitted by another homogenous or
heterogeneous device
o Communication and collaboration of heterogeneous services regardless of
application platforms, programming languages, operating systems, or system
architecture
Research Challenges

 Heterogeneous and multilayer communication and network infrastructure
o Devices with multimode connectivity are common
o 5G devices will also support multiple wireless communication technologies
Research Challenges
BLE Wi-Fi Z-Wave ZigBee LTE-M NB-IoT LoRa
Range 10 m – 1.5 km 15 m – 100 m 30 m – 50 m 30 m – 100 m 1 km – 10 km 1 km – 10 km 2 km – 20 km
Throughput
125 kbps to
2 Mbps
54 Mbps to
1.3 Gbps
10 kbps to
100 kbps
20 kbps to
250 kbps
Up to 1 Mbps Up to 200 kbps 10 kbps to
50 kbps
Power
Consumption
Low Medium Low Low Medium Low Low
Topology P2P
Star
Mesh
Broadcast
Star
Mesh
Mesh Mesh Star Star Star
https://www.bluetooth.com/blog/wireless-connectivity-options-for-iot-applications

 Heterogeneous and multilayer
communication and network
infrastructure
o Numerous opportunities
 Simultaneous transmission of data to
fulfill QoS requirements of real time
applications
 Adaptive protocols to reduce energy
consumption or task completion time
 Dynamic allocation of communication
interfaces to application tasks
Research Challenges
o Research challenges
 Efficient discovery and monitoring
mechanism for a multi-layer
network environment
 Cost estimation models for multiple
wireless communication
technologies
 Adaptive and robust multi-network
management and routing protocol
 Maximum utilization of static and
dynamic links at multiple network
layers

Research Objective
 To develop an intelligent middleware platform that efficiently utilize the
characteristics and address the challenges of heterogeneous computing
and a multilayer network environment to
1) manage heterogeneous computing and network resources efficiently, and
2) provide task processing, data collection, and data storage services to support
emerging smart city and 5G network applications.

 Middleware platforms for IoT systems
o provide access to and control of physical devices
o support data collection, data analysis, or application composition services
o do not provide computing services, do not utilize network-level information, such
link quality and link lifetime, and are not designed for complex multilayer network
environments
Exiting middleware platforms

 Middleware platforms for distributed computing systems, such as edge clouds
and mobile ad hoc clouds
o provide computing or storage services but do not support data collection and
actuation services
o they do not efficiently utilize heterogeneous routes, simultaneous transmission on
multiple communication technologies, and several network-level parameters, such as
link quality and lifetimes
o Most of these platforms do not use end devices as service provider nodes

 Machine learning at network and resource management layers
o [11] uses supervised machine learning technique to predict hidden paths.
o [12] and [13] employs deep learning model that uses traffic patterns in router to predict the next
node in the routing path.
o [14] uses non-linear regression technique to estimate link quality and [15] uses machine
learning to improve multi hop wireless routing.
o [16] uses reinforcement learning based approach to manage resources in distributed computing
environment.
o [17] uses reinforcement learning to reduce application execution times.
o Machine learning based algorithms has been used to address resource allocation problem but
they don’t exploit data generated at heterogeneous and multilayer communication and network
infrastructure

Summary of exiting middleware platforms
Existing middleware
platforms for mobile
computing systems
A proposed middleware
platform for a
heterogeneous private
edge cloud system
Multi-network aware ✓ ✓
Efficient utilization of multi-network environment × ✓
Complex multi-network aware
(A complex multi-network infrastructure integrate ad
hoc and infrastructure-based network technologies)
× ✓
Sensing or actuation services ✓ ✓
Computing and storage services ✓ ✓
Computing, storage, sensing, and actuation services × ✓
Mobility management ✓ ✓
Failure management ✓ ✓
Link quality and lifetime aware ✓ ✓
Energy aware ✓ ✓

A machine learning-based intelligent middleware platform
 A new middleware platform aims to
o provide computing, data collection, data storage, sensing and actuation
services to support emerging resource-intensive and non-resource-intensive
smart city and 5G network applications
o leverage regression analysis and reinforcement learning methods to solve the
problem of efficiently allocating heterogeneous resources to application tasks
o adopts parallel transmission techniques, dynamic interface allocation
techniques, and machine learning-based algorithms in a dynamic multilayer
network infrastructure to improve network and application performance

A Machine Learning-based Intelligent Middleware Platform

 Physical device layer
o This layer includes devices, such as sensors, actuators, personal computers,
and smartphones
o Sensors provide data collection service, and actuators provide device
movement and control services
o High-end devices, such as personal computers and smartphones, provide task
processing and storage services
o A single device can provide multiple services

 Virtual device layer
o A virtual device is a representation of a physical device
o It enables an application or service to access the physical device resources or
its functionality
o A single virtual device may also represent multiple physical devices
Service Physical
Device
Front-end Back-end
Virtual Device
HTTP-REST Request
HTTP-REST Response
Service
Front-end Back-end
Virtual Device
HTTP-REST Request
HTTP-REST Response
Physical Device 1
Physical Device 2

o Simple virtual devices, such as virtual sensors or actuators, provide data
collection or actuation services
o A microservice is used to implement a simple virtual device
o A microservice or simple virtual device either run on physical device they
represent or on another physical device such as a RPi or WiFi router
Service
Physical
Sensor X
Virtual Sensor X
HTTP-REST Request
HTTP-REST Response
Microservice X
Raspberry PI

o A container-based virtual device represent a high-end physical device that
provide task processing, data caching and data storage services
o A container-based virtual device hosts a container engine that executes
containerized applications or microservices
Container-based virtual device
Container
Physical Device
Operating System
Docker Container Engine
Container
Microservice
Container
Application
Data caching microservice
Data storage microservice

 Raspberry Pi hosts a container engine that executes remote containerized
microservice and also executes virtual sensor x microservice that enables
access to physical sensor x
Container
Raspberry Pi
Raspberry Pi OS
Virtual Sensor X
Microservice
Physical
Sensor X
Container
Microservice
Container
PC
Windows OS
Applications
Container
Microservice
Container
Mobile Robot
Linux OS
ocker Container Engine
Microservice
ner
uator
vice
Con
Raspberry Pi
Raspberry Pi OS
Virtual
Micro
Container
Microservice
Container
PC
Windows OS
Applications
Container
Microservice
Container
Mobile Robot
Linux OS
Microservice
Container
Virtual Actuator
Microservice
Physical
Actuator

 The multi-network management layer aims to
o use the capabilities of machine learning and SDN to improve network and application
performance
o provide serial and parallel data transmission services across multiple heterogeneous
networks
o support the dynamic allocation of
network interfaces
o adopt new machine learning-based link
quality and Markov chain-based link
lifetime estimation techniques to reduce
communication and energy consumption
costs.

 The resource management layer aims to
o Neural networks and reinforcement learning methods to efficiently allocate
heterogeneous computing and network resources to application tasks
o use parallel transmission techniques, dynamic interface allocation techniques,
and network-level parameters to address diverse application requirements
Task Queue
Containerized Machine Learning
based Failure Management Service
Containerized Discovery and
Monitoring Service
Containerized Task
Dispatch Service System Data Store
Containerized Machine Learning-based Multi-
network Aware Resource Allocation Service
Virtual Device Registry

 Middleware services are also containerized and execute on Docker
platform
Container
Mobile Robot
Linux OS
Multi-network
Management Layer
Services
Container
Resource Management
Layer Services
Container
Microservice
Container
Microservice
Container
Microservice

 www.mgclab.com
o A Machine Learning-based Middleware Platform for a Heterogeneous Private Edge
Cloud System
 National Research Foundation Korea
o An Energy Efficient Resource Management System for a Mobile Ad hoc Cloud
o Wi-Fi Direct based mobile ad hoc network
o Adaptive network layer for heterogeneous network environment
 HUFS Research Foundation
o Self learning Smart Ageing Service
Research at Mobile Grid and Cloud Computing Lab

Acknowledgements
 Cisco
o Fog Computing and the Internet of Things: Extend the Cloud to Where the Things Are
 Koustabh Dolui and Soumya Kanti Datta
o Comparison of Edge Computing Implementations: Fog Computing, Cloudlet and Mobile
Edge Computing
 5G White Paper by NGMN Alliance

References
[1] Sayed Chhattan Shah, Recent Advances in Mobile Grid and Cloud Computing, Intelligent Automation and Soft Computing, 2017
[2] Sayed Chhattan Shah, Design of a Machine Learning-Based Intelligent Middleware Platform for a Heterogeneous Private Edge Cloud System, Sensors, 2021
[3] E. K. Markakis et al., Computing, Caching, and Communication at the Edge: The Cornerstone for Building a Versatile 5G Ecosystem, IEEE Communications Magazine, 2017
[4] Sayed Chhattan Shah, An Energy-Efficient Resource Management System for a Mobile Ad Hoc Cloud, IEEE Access, 2018
[5] H. Kim et al., Autonomic Management of Application Workflows on Hybrid Computing Infrastructure, Telecommunication System, 2011.
[6] Sayed Chhattan Shah et al., An Energy-Efficient Resource Allocation Scheme for Mobile Ad Hoc Computational Grids. Journal of Grid Computing, 2011.
[7] Sayed Chhattan Shah et al., An Effective and Robust Two-phase Resource Allocation Scheme for Interdependent Tasks in Mobile Ad Hoc Computational Grids, Journal of Parallel
and Distributed Computing, 2012.
[8] Sayed Chhattan Shah, Energy Efficient and Robust Allocation of Interdependent Tasks on Mobile Ad hoc Computational Grid, Concurrency and Computation: Practice and
Experience, 2015.
[9] Bo Li, Yijian Pei, Hao Wu, Bin Shen, Heuristics to allocate high-performance cloudlets for computation offloading in mobile ad hoc clouds, The Journal of Supercomputing, 2015.
[10] Jianbo Du et al., Computation offloading and resource allocation in mixed fog/cloud computing systems with min-max fairness guarantee. IEEE Transactions on Communications,
2018.
[11] A. Boutet et al., C3PO: A network and application framework for spontaneous and ephemeral social networks. Web Information Systems Engineering, 2015.
[12] B. Mao et al., Routing or Computing? The Paradigm Shift Towards Intelligent Computer Network Packet Transmission Based on Deep Learning, IEEE Trans. Computers, 2017
[13] N. Kato et al., The Deep Learning Vision for Heterogeneous Network Traffic Control: Proposal, Challenges, and Future Perspective, IEEE Wireless Commun., 2016.
[14] K. Tang, C. Li, H. Xiong, J. Zou and P. Frossard, Reinforcement learning-based opportunistic routing for live video streaming over multi-hop wireless networks, IEEE 19th
International Workshop on Multimedia Signal Processing, 2017.
[15] S. Karunaratne and H. Gacanin, An Overview of Machine Learning Approaches in Wireless Mesh Networks, IEEE Communications Magazine, 2019
[16] Z. Liu et al., A Reinforcement Learning Based Resource Management Approach for Time-critical Workloads in Distributed Computing Environment, IEEE International
Conference on Big Data, 2018.
[17] Orhean A.I., Pop F., Raicu I., New scheduling approach using reinforcement learning for heterogeneous distributed systems, J. Parallel Distrib. Comput., 2017

Cloud and Edge Computing Systems

Recommended

More Related Content

What's hot (20)

Similar to Cloud and Edge Computing Systems (20)

More from Sayed Chhattan Shah (17)

Recently uploaded (20)

Cloud and Edge Computing Systems