Fundamentals

of
Fog and Edge Computing
BCSE313L
WINTER 2024 - 2025

Module 1 - Internet of Things (IoT) and New

Computing Paradigms
Introducing Internet of Things (IoT)
 IoT as a Network of Physical Objects: The IoT connects
physical devices like appliances, vehicles, and sensors to the
internet.
 Enhanced Efficiency: IoT improves the efficiency of
various applications such as logistics, manufacturing, and
agriculture.
 Cloud-Centric IoT (CIoT) Architecture: IoT systems often
follow a cloud-centric architecture where physical objects are
represented as web resources managed by cloud servers.
CIoT – Traditional Approach
CIoT – Traditional Approach
Introducing Internet of Things (IoT)
 Key Components of IoT Systems:
Front-end Devices: Sensors, actuators, and readers
interact with physical objects.
 Gateways: Devices like routers and modems connect
front-end devices to the internet.
Introducing Internet of Things (IoT)
 Core Technologies of IoT:
 Embedded Systems: Provide intelligence to front-end
devices.
 Middleware: Connects heterogeneous devices to the
cloud.
 Cloud Services: Offer storage, processing, and
management capabilities.
Introduction - Limitation of CIoT - BLURS
B • Bandwidth – (Data Generated)

• Latency – (Delay or Time

L between)

U • Uninterrupted Connectivity

• Resource Constraints (Compute

R Power)

S • Security (Attacks on edge)

Introduction - Limitation of CIoT - BLURS
Bandwidth: High-frequency data generation exceeds cloud
bandwidth.
 Connected cars and self-driving vehicles generate massive amounts of data.

 Latency: Cloud-based processing can introduce significant delays.

 Industrial control systems, autonomous vehicles, and real-time applications
require low latency.
 Uninterrupted Connectivity: Distance and network instability can
disrupt service.
 Network outages can disrupt IoT services, especially for remote devices.
Introduction - Limitation of CIoT
Resource Constraints: Front-end devices may lack
processing power and energy.
 Battery-powered devices may struggle with continuous data
transmission.
 Security: Remote devices are vulnerable to attacks.
 Edge devices are susceptible to attacks and data manipulation.
Introduction - Emerging Paradigms: Overcome Limitations of CIoT

 Geo-Distributed Cloud Computing: Distributes

processing to data centers closer to the edge.
Introduction - Emerging Paradigms: Overcome Limitations of CIoT

 Geo-Distributed Cloud Computing: Distributes

processing to data centers closer to the edge.
Introduction - Emerging Paradigms: Overcome Limitations of CIoT

Mobile Cloud Computing: Leverages local wireless access

points for cloud resources.
Introduction - Emerging Paradigms: Overcome Limitations of CIoT

MANET-Based Cloud Computing: Utilizes mobile devices

to form ad-hoc networks for cloud services.
Introduction - Emerging Paradigms: Overcome Limitations of CIoT

Fog
Computing: A
hierarchical
approach that
distributes
processing to
devices at the
edge of the
network.
Introduction - Emerging Paradigms: Overcome Limitations of CIoT

Fog
Computing: A
hierarchical
approach that
distributes
processing to
devices at the
edge of the
network.
Introduction – Fog Computing - Architecture
Introduction – Fog Computing - Architecture
Cloud

Management and
Fog Gateways

Orchestration
Fog
Nodes/Servers
Edge
Gateways

Edge Device
Introduction – Fog Computing - Architecture
Introduction – Fog Computing - Architecture
Hierarchical Model: Involves multiple layers of devices,
from edge devices to fog nodes and cloud servers.
Scalability: Supports horizontal scaling to accommodate
increasing workloads.
Flexibility: Can adapt to different IoT scenarios and
requirements.
Introduction - Emerging Paradigms: Overcome Limitations of CIoT

 Geo-Distributed Cloud Computing: Distributes

processing to data centers closer to the edge.
Mobile Cloud Computing: Leverages local wireless access
points for cloud resources.
MANET-Based Cloud Computing: Utilizes mobile devices
to form ad-hoc networks for cloud services.
Fog Computing: A hierarchical approach that distributes
processing to devices at the edge of the network.
Introduction – Fog Computing - Benefits
Reduced Latency: Processing data closer to the source
reduces response time.
Improved Bandwidth Utilization: Offloads data
processing from the cloud.
Enhanced Security: Sensitive data can be processed
locally, reducing security risks.
Increased Reliability: Distributed architecture improves
system resilience.
IoT applications and environments with supporting computing paradigms
Why Fog Computing for IoT
 Distributes tasks across different tiers based on
complexity and resource requirements.
 Offloads simple tasks to edge devices and complex
tasks to cloud or core network.
 Enables timely decision-making at the edge for
critical applications like healthcare, security, and
industrial automation.
Relevant Technologies

Fog Computing

Edge Computing

Multi-access Edge
Computing

Mist Computing

Cloudlets

Virtualisation
Relevant Technologies
A distributed computing architecture that brings computation and
Edge data storage closer to the sources of data.
Computing Reduced Latency: By processing data locally, edge computing significantly
reduces latency, making it ideal for real-time applications like autonomous
vehicles, AR/VR, and IoT
Improved Performance: Edge computing can handle large
volumes of data and complex computations more efficiently than
traditional cloud computing
Enhanced Security: Sensitive data can be processed and stored
locally, reducing the risk of data breaches and cyberattacks.

Increased Reliability: Distributed architecture makes edge

computing more resilient to failures and disruptions..

Reduced Bandwidth Consumption: Less data needs to

be transmitted to the cloud, saving bandwidth and
reducing costs
Relevant Technologies
Fog A specific implementation of edge computing, often focused
on IoT applications
Computing

Involves a hierarchical architecture with multiple layers of

computing resources.

Aims to address the limitations of cloud-centric IoT by offloading

processing tasks to edge devices.
Relevant Technologies
Small-scale cloud data centres located at the edge of the internet. They are
Cloudlets designed to provide computing resources to mobile devices, reducing
latency and improving the performance of resource-intensive and
interactive mobile applications.
Small-scale: Cloudlets are smaller than traditional data centres,
often housed in racks or containers.

Edge location: They are located closer to the end-users, typically

within wireless networks or cellular base stations.

Resource-constrained: While they provide significant computing power

compared to mobile devices; cloudlets have limited resources compared to
large-scale data centres.

Focus on mobile devices: Cloudlets are optimized for mobile

applications, providing low-latency access to computing resources
Relevant Technologies – Cloudlet Computing
Cloudlet computing is a direction in mobile computing that shares traits with
Mobile Cloud Computing (MCC) and Multi-Access Edge Computing (MEC).
A cloudlet is a trusted, resource-rich computer or cluster of computers with a
strong internet connection, utilized by nearby mobile devices.
Cloudlets are small data centers (miniature clouds) typically one hop away
from mobile devices.
Cloudlet computing involves offloading computation from mobile devices to
Virtual Machine (VM)-based cloudlets located on the network edge.
Cloudlet is the middle tier of a 3-tier continuum: mobile device-cloudlet-
cloud.
Relevant Technologies – Cloudlet Computing
 Cloudlet computing offers lower latency and energy consumption compared to cloud
computing.
 It is intended to serve devices in the local area.
 Cloudlet computing has high potential in domains such as wearable cognitive
assistance and web applications.
 Cloudlet computing differs from Mobile Access Cloud Computing (MACC) in that it
requires infrastructure with virtualization and VM capability.
 Cloudlet computing supports local services for mobile clients by dividing tasks among
cloudlet nodes in proximity to mobile devices.
 Fog computing offers a more generic alternative that natively supports large amounts
of traffic and allows resources to be anywhere along the thing-to-cloud continuum.
Relevant Technologies – Cloudlet Computing
Relevant Technologies
Virtualizati Technology that allows multiple operating systems to run
on a single physical machine.
on

Enables efficient utilization of resources and flexible

deployment of cloud services
Relevant Technologies
A network architecture that brings cloud computing
Multi-
capabilities closer to the edge of the network.
access
Instead of processing data in a centralized cloud data center,
Edge processing is done at the edge of the network, like cellular
base stations or Wi-Fi access points
Computing
Introduced by ETSI as a telecommunications standard.
(MEC)

Focuses on providing cloud-like services at the edge of

the network.

Leverages existing network infrastructure like cellular

base stations.
Relevant Technologies
MEC is an extension of mobile computing through edge computing.
Multi-
access ETSI defines MEC as a platform providing IT and cloud-computing
capabilities within the Radio Access Network (RAN) near mobile
Edge subscribers.
Computing MEC extends edge computing by providing compute and storage
resources near low-energy, low-resource mobile devices.
(MEC)
MEC operates from the edge of the Internet and can function with
little to no Internet connectivity.

MEC establishes connectivity through WAN, WiFi, and cellular

connections.

MEC incorporates Software-Defined Networking (SDN) and Network

Function Virtualization (NFV) capabilities.
Relevant Technologies – MEC Multi Access Edge Computing
Relevant Technologies
A subset of fog computing. It aims to minimize latency
Mist
and enable real-time decision-making at the edge.
Computing

A specific implementation of edge computing that

focuses on bringing computing capabilities to the
extreme edge of the network, closer to IoT devices.

Enables IoT devices to have self-awareness and self-

management capabilities.

Employs extreme edge computing to minimize latency.

Relevant Technologies – Mist Computing
Fog Computing Vs Cloudlets
Criteria Fog Computing Cloudlet
Scope Fog computing encompasses a Cloudlets are more focused on
broader range of edge providing cloud services to
computing scenarios, including mobile devices.
IoT devices and network
gateways.

Architectu Fog computing involves a Cloudlets are typically deployed

re hierarchical architecture with as standalone servers or virtual
multiple layers of edge devices machines.
and fog nodes.

Goals Fog computing aims to Cloudlets primarily focus on

improve overall system enhancing mobile application
performance, efficiency, and performance and user
reliability. experience.
Essence of Integration
Reduced
Latency

Enhance
d
Security

Improved
Reliability

Increase
d
Flexibility
Fog and Edge Computing Completing the Cloud

Security

Efficienc Cognitio
y
SCAL n

E
Latency Agility
Fog and Edge Computing Completing the Cloud
Advantages of FEC - SCALE

Security
 Enables faster and more reliable security updates for IoT devices.

 Reduces vulnerability to cyberattacks by allowing for more timely

security patches.
Fog and Edge Computing Completing the Cloud
Advantages of FEC - SCALE

Cognition
 Enables autonomous decision-making at the edge.

 Allows IoT devices to adapt to changing conditions and respond to

events without relying on cloud-based decisions.
 Promotes self-awareness, self-organization, and self-healing
capabilities in IoT devices.
Fog and Edge Computing Completing the Cloud
Advantages of FEC - SCALE

Agility
 Enables faster and more flexible deployment of IoT systems.

 Empowers individuals and small businesses to participate in

providing FEC services.
 Promotes the adoption of open standards and open-source
software.
Fog and Edge Computing Completing the Cloud
Advantages of FEC - SCALE

Latency:
 Reduces latency by processing data closer to the source.

 Enables real-time processing and decision-making.

 Supports time-sensitive applications like industrial automation and

autonomous vehicles.
Fog and Edge Computing Completing the Cloud
Advantages of FEC - SCALE

Efficiency:
 Improves system performance by offloading tasks to edge devices.

 Reduces bandwidth consumption and operational costs.

 Enables efficient utilization of resources.

Fog and Edge Computing Completing the Cloud
 How does FEC provide SCALE advantages?
Fog and Edge Computing Completing the Cloud
 How does FEC provide SCALE advantages?

Storage Compute Accelerati Networkin Control

• FEC nodes • FEC nodes on g • Deployment
store data provide • Networking • Vertical Control
temporarily computing Acceleration: Networking: • Actuation
to improve resources • Computing • Horizontal Control
performance. through Acceleration: Networking • Mediation
I/PaaS and Control
SaaS models. • Security
Control
Fog and Edge Computing Completing the Cloud
 How does FEC provide SCALE advantages?

Storag FEC nodes store data temporarily to

improve performance.
e:
Content providers can cache multimedia
content closer to customers for better
quality of experience.
Connected vehicles can use FEC nodes to
store and share information.
Fog and Edge Computing Completing the Cloud
 How does FEC provide SCALE advantages?
Comput FEC nodes provide
computing resources
e: through I/PaaS and SaaS
models.

I/PaaS offers flexible

platforms using VMs or CEs
for customized software
deployment.
On-demand data processing (ODP): Pre-
SaaS provides: installed methods for processing client
data.
Context as a service (CaaS): Customized
data collection, processing, and information
generation.
Fog and Edge Computing Completing the Cloud
 How does FEC provide SCALE advantages?
Accelerati Networking Traditional networks have fixed
routing paths.
on
Acceleration FEC supports network virtualization
for flexible routing.

Clients can configure custom routing

paths for optimal speed.

FEC nodes can use advanced

Computing hardware like GPUs and FPGAs.
Acceleration GPUs enhance complex algorithms.

FPGAs allow for reconfigurable

hardware, improving efficiency and
customization.
Fog and Edge Computing Completing the Cloud
 How does FEC provide SCALE advantages?
Networki Connects IoT devices and cloud
Vertical Networking using IP-based protocols (TCP/UDP,
ng: HTTP, CoAP, XMPP, AMQP, MQTT).

FEC nodes act as proxies or

message brokers for efficient data
transfer.

Connects IoT devices using various

Horizontal protocols (Bluetooth, ZigBee, Z-
Networking Wave).
IP network gateways are ideal for
hosting FEC servers to bridge the gap
between IoT devices and cloud.
FEC servers can control communication
between IoT devices and cloud, optimizing
efficiency and customization.
Fog and Edge Computing Completing the Cloud
 How does FEC provide SCALE advantages? - Control
• Clients can dynamically deploy and control software execution on FEC
nodes.
Deployment Control: • FEC providers can offer network topology as a service for flexible
program deployment.
• Clients can control multiple FEC nodes for optimal performance.

• Cloud can delegate control of IoT devices to FEC nodes for

Actuation Control: efficient decision-making.

• FEC nodes can facilitate communication between devices from

Mediation Control: different providers.
• On-demand software updates enhance interoperability.

• FEC nodes ensure secure authentication, authorization, identity,

Security Control: and protection of the runtime environment.
Fog and Edge Computing Completing the Cloud
Deployment Control - Example
• Automatic software updates: Fog edge devices can
automatically update software and firmware to ensure
they have the latest security patches and features.
• Container orchestration: Fog edge devices can use
container orchestration tools like Kubernetes to manage
and deploy containerized applications.
• Device provisioning: Fog edge devices can be
provisioned remotely, allowing administrators to
configure and deploy devices without physical access.
Fog and Edge Computing Completing the Cloud
Actuation Control - Example
• Real-time control of industrial equipment: Fog
edge devices can control industrial equipment in real-
time, allowing for faster response times and improved
efficiency.
• Smart lighting systems: Fog edge devices can
control smart lighting systems, adjusting lighting
levels based on ambient light, occupancy, and time of
day.
• Autonomous vehicles: Fog edge devices can control
autonomous vehicles, making real-time decisions
Fog and Edge Computing Completing the Cloud

Mediation Control - Example

Data filtering and processing: Fog edge devices
can filter and process data in real-time, reducing the
amount of data that needs to be transmitted to the
cloud.
Protocol conversion: Fog edge devices can convert
between different protocols, allowing devices with
different communication protocols to communicate
with each other.
• Data aggregation: Fog edge devices can
aggregate data from multiple sources, providing a
Fog and Edge Computing Completing the Cloud

Security Control - Example

Intrusion detection and prevention: Fog edge devices

can detect and prevent intrusions in real-time, reducing
the risk of security breaches.
Encryption: Fog edge devices can encrypt data in real-
time, protecting it from unauthorized access.
• Access control: Fog edge devices can control access to
devices and data, ensuring that only authorized
personnel have access.
Hierarchy of Fog and Edge Computing
 CIoT systems can deploy FEC servers at three locations
relative to the central cloud:
 Inner-edge
 Middle-edge
 Outer-edge
Hierarchy of Fog and Edge Computing
Hierarchy of Fog and Edge Computing
Hierarchy of Fog and Edge Computing
Hierarchy of Fog and Edge Computing
Factor INNER EDGE MIDDLE EDGE OUTER EDGE
Location Large-scale networks like WANs Local area networks (LANs) and IoT devices & IP gateway
and MANs. cellular networks. devices.
Role: Provides fundamental Offers computing and storage Provides computing and
infrastructure and geo- resources closer to the edge storage capabilities at the
distributed caching/processing devices. very edge of the network.
for improved performance.
Work Leverage existing data centers Utilize network devices like • For constraint devices:
and network infrastructure to routers and base stations as Focus on self-
deploy FEC servers for FEC hosts. management and
enhanced web services and collaboration through
local cloud services. mist computing.
• For integrated devices:
Explore virtualization-
based FEC solutions as
technology advances.
• For IP gateway devices:
Utilize their computing
power and networking
capabilities to host FEC
services.
Hierarchy of Fog and Edge Computing
 Inner Edge
 Location
 Countrywide, statewide, or regional networks of enterprises, ISPs, data
centers of evolved packet core (EPC), and metropolitan area networks (MAN).
 Initial Function
 Provides basic infrastructure to connect local networks to the global internet.

 Recent Trend
 Geo-distributed caching and processing for improved Quality of Experience
(QoE) for web services.
Hierarchy of Fog and Edge Computing
 Inner Edge
 Examples:
 Google Edge Network collaborating with ISPs for data server
distribution at ISP data centers for faster response times.
 Local cloud services offered by ISPs (AT&T, Telstra, Vodafone,
Deutsche Telekom) to meet low-latency needs of local
businesses.
 Fog Computing Reference: WAN-based cloud data centers can be
considered as the "fog" of inner-edge.
Hierarchy of Fog and Edge Computing
Middle Edge
 Local Area Network (LAN)
 Fog Computing Architecture: Cisco's research introduced the concept
of using internet gateway devices (like Cisco IR829) to provide cloud-
like services.
 Virtualization: These devices can support FEC mechanisms through
virtualization technologies.
 Local Cloud/Data Center/Cloudlet: Virtualization-enabled server
computers within the same subnet (one-hop range) can act as FEC
nodes.
Hierarchy of Fog and Edge Computing
Middle Edge
 Cellular Network
 Network Virtualization: Leveraging existing network virtualization
technologies used in cellular networks.
 Diverse BTS Types: Macrocells, microcells, picocells, and femtocells can
serve as roadside FEC hosts for various mobile IoT use cases.
 Equipment Providers: Companies like Nokia, ADLink, and Huawei are
developing MEC-enabled hardware and infrastructure solutions.
 Future Trend: Widespread availability of cellular network-based FEC in
various network equipment.
Hierarchy of Fog and Edge Computing
Outer Edge
 Location: Front-end of the IoT network
 Device Types:
 Constraint Devices (Sensors, Actuators)
 Limited processing power and memory (example: Atmel
ATmega328 microcontroller)
 Field-programmable: Can be updated or reconfigured remotely.
 Mist Computing: Enables self-management and collaboration
among devices for M2M communication.
Hierarchy of Fog and Edge Computing
Outer Edge
 Integrated Devices (Smartphones, Tablets)
 Decent processing power, networking capabilities, embedded sensors, storage
 Limited OS environment might hinder virtualization
 Promising solutions: Apache Edgent, Termux
 Future Trend: Virtualization-based FEC on integrated devices
 IP Gateway Devices (Hubs)
 Mediate between constrained and middle-edge devices
 Use low-energy protocols (Bluetooth Low Energy, ZigBee, Z-Wave)
 Relay communication between constrained devices and internet
 Can host virtual environments (Docker Containers Engine)
 Often used as FEC nodes in research projects
Business Models

X as a Service

Application Service

Support Service
Business Models
X as Covers services like infrastructure, platform, software,
and networking as a service.
a
Servi Allows clients to pay for hardware usage and
additional service fees for software updates and
ce technical support.

(XaaS Service providers may become independent of

hardware vendors, similar to companies like RedHat,
): IBM, or Microsoft.

The model extends to individuals or small businesses

offering IndieFog solutions under the consumer-as-
provider (CaP) model, e.g., MQL5 Cloud Network and
Fon.
Business Models
Applicati Provides solutions to process data efficiently,
on within or outside client environments.
Service:

Examples include Digital Twinning for

industrial optimization, real-time traffic
control for autonomous vehicles, and ambient
assisted living (AAL) analytics.
IndieFog providers can offer specific services,
like data stream classification using platforms
like Apache Edgent, for applications in local and
mobile contexts.
Business Models
Suppo Involves software management,
rt including installation, configuration,
updates, and ongoing technical support
Servic for FEC systems.
e: Clients may subscribe to monthly or
annual plans for customized
maintenance solutions.

Example: Cisco’s fog computing solution

with IOX-enabled equipment that
supports customized operations.
Opportunities
Pre- Allow users to easily configure and deploy applications
integrated via user-friendly interfaces.
FEC runtime
platforms Simplified application deployment, reduced
complexity, and improved user experience.
Cisco Edge, HPE Edgeline

Open Open platforms for FEC are in early stages with limited
Platforms deployment support.
Increased flexibility, reduced vendor lock-in, and
improved interoperability.
OpenStack++, WSO2–IoT Server, Apache Edgent
Opportunities
Design Placement of FEC Servers:

Identify optimal server locations across edge tiers (inner-edge,

middle-edge, outer-edge).
Resource Use or develop abstract models to define resource requirements
Modelling and interactions for FEC servers.

Implementatio Nodes at the middle and outer edge vary in hardware, protocols,
n - and operating systems.
Heterogeneity Vendor-specific isolated platforms (e.g., Cisco, Dell) complicate
implementation, requiring custom solutions for each platform.
Time & Cost Open platform projects exist but are not yet unified, adding to
Concerns development complexity and resource investment.
Challenges
Security Pre-integrated platforms may introduce security
Risks risks if not properly secured.

Limited Pre-integrated platforms may limit customization

Customiza options, making it difficult to meet specific use case
tion requirements.

Vendor Pre-integrated platforms may lead to vendor lock-in,

Lock-in making it challenging to switch to alternative
solutions.
Challenges
Complexity Open platforms can be complex to deploy and manage,
requiring specialized skills and expertise.

Limited Open platforms may have limited support and

Support documentation, making it challenging to troubleshoot
issues.

Interoperabi Open platforms may face interoperability challenges,

lity making it difficult to integrate with other systems and
devices.

Device Managing a large number of IoT devices can be

Managemen challenging, requiring specialized device management
t solutions.
Opportunities & Challenges
Market Growth:
The market value for FEC hardware is expected to reach $7,659
million by 2022, indicating increased availability of FEC-ready
equipment like routers, switches, IP gateways, and hubs.
Opportunities
Out-of-Box Experience (OOBE): Two forms:
OOBE-Based Equipment:
Pre-integrated FEC runtime platforms in devices (e.g., routers and
gateways) allow users to easily configure and deploy applications
via user-friendly interfaces.
OOBE-Based Software:
Like Microsoft Windows, enables FEC runtime environments and
SCANC mechanisms on FEC-compatible devices without requiring
complex configurations.
Challenges
Standardization: OOBE-Based Equipment:
 Vendors face decisions on what FEC platforms and software packages
to pre-integrate.
OOBE-Based Software:
 Compatibility issues arise from heterogeneous device specifications
and architectures (e.g., x86, ARM), requiring multiple software versions.
Development Costs:
 Creating and maintaining OOBE-based software is costly unless a
unified standard for FEC-compatible hardware is established.
Opportunities
Open platforms for FEC are in early stages with limited
deployment support.
A major challenge is the inflexibility of commercial devices in
supporting FEC configurations, presenting opportunities for
improved device designs.
Opportunities
OpenStack++:
Developed by Carnegie Mellon University.
Features IoT-related mechanisms for connecting devices (e.g.,
Arduino, Raspberry Pi, Android, iOS) to the cloud via protocols like
MQTT and XMPP.
Includes WSO2–IoT Server with embedded Siddhi 3.0 for real-time
streaming on embedded devices.
Less suitable for lightweight virtualization like Docker, which is more
relevant for FEC use cases.
Opportunities
WSO2–IoT Server:
Extension of the WSO2 open-source enterprise integration platform.
Provides FEC computing capabilities to outer-edge devices by
enabling real-time streaming and analytics.
Opportunities
Apache Edgent:
Formerly known as Quarks, contributed by IBM (edgent.apache.org).
Supports distributed stream data processing between cloud and
edge devices.
Cloud-side integrates with major platforms like Apache Spark,
Storm, and Flink.
Edge-side supports Linux and Android OS.
Enables dynamic migration of stream data processing between
cloud and edge for edge analytics.
Challenges
Current platforms lack end-to-end support across all layers
of edge networks.
Commercial devices need enhancement to support FEC
configurations effectively.
The shift toward containerization (e.g., Docker) indicates a
need for platforms optimized for lightweight virtualization in
FEC deployments.
Opportunities – Life Cycle Phases
Design:
Placement of FEC Servers:
Identify optimal server locations across edge tiers (inner-edge,
middle-edge, outer-edge).

Resource Modeling:
Use or develop abstract models to define resource requirements
and interactions for FEC servers.
Opportunities – Life Cycle Phases
Implementation:
Heterogeneity Challenges:
Nodes at the middle and outer edge vary in hardware, protocols,
and operating systems.
Vendor-specific isolated platforms (e.g., Cisco, Dell) complicate
implementation, requiring custom solutions for each platform.
Time and Cost Concerns:
Open platform projects exist but are not yet unified, adding to
development complexity and resource investment.
Opportunities – Life Cycle Phases
Adjustment:
Runtime Adaptability:
 Systems must dynamically schedule, deploy, and terminate runtime
environments (e.g., VMs, containers) to optimize performance.
Dynamic Migration:
 Runtime environments and application methods should move between FEC
nodes based on context, raising challenges in:
 Reliability:
 Ensuring successful migration with minimal failures.
 Latency:
 Reducing delays caused by wireless and mobile communication interruptions, particularly at the
far-edge.
Challenges – Life Cycle Phases
Supporting reliable software migration across FEC nodes.
Minimizing latency during runtime adjustments, especially in
dynamic outdoor environments.
Addressing signal interruptions in wireless communication
for code deployment and reconfiguration.