Dr Vijay Birchha

School of Computer Science Engineering and Artificial

Intelligence (SCAI)
VIT Bhopal University
Email- vijaybirchha@[Link], Cabin-AB2-203
Module -1 Contents

• Introduction to Fog and Edge Computing: Introduction, Relevant

Technologies, Fog and Edge Computing Completing the Cloud, Advantages
of FEC: SCALE, Security, Cognition, Agility, Latency, Efficiency, How FEC
Achieves These Advantages: SCANC, Storage, Compute, Acceleration,
Networking, Control, Hierarchy of Fog and Edge Computing, Inner-Edge,
Middle-Edge, Local Area Network, Cellular Network, Outer-Edge,
Constraint Devices, Integrated Devices, IP Gateway Devices, Business
Models, X as a Service, Support Service, Application Service, Opportunities
and Challenges, Out-of-Box Experience, OOBE-Based Equipment, OOBE-
Based Software, Open Platforms, OpenStack++, WSO2–IoT Server, Apache
Edgent, System Management, Design, Implementation, Adjustment.
Introduction
• The Internet of Things (IoT) is a comprehensive environment that
interconnects a vast number of heterogeneous physical objects—such
as appliances, vehicles, and sensors—to the Internet to enhance
application efficiency in areas like logistics, manufacturing, and smart
homes.
CIOT model facing challenges-BLURS
• Traditionally, IoT systems follow a
Cloud-centric architecture (CIOT),
where physical objects are
managed as Web resources by
central servers in the global
Internet.
• The CIOT model faces growing
challenges-BLURS
1. Bandwidth
2. Latency
3. Uninterrupted Connectivity
4. Resource Constraints
5. Security
Advantages of FEC: SCALE

The advantages of implementing FEC are categorized as SCALE:

1. Security: Supporting safety and trustworthiness in transactions.
2. Cognition: Enabling autonomous decision-making and self-management for
smart devices.
3. Agility: Allowing individuals and small businesses to participate in providing
services through open interfaces.
4. Latency: Providing rapid responses for time-sensitive applications.
5. Efficiency: Improving performance while reducing unnecessary costs, such
as outgoing bandwidth.
How FEC achieved the advantages-
These advantages are achieved through five basic mechanisms known
as SCANC:
1. Storage (caching),
2. Compute (I/PaaS or SaaS),
3. Acceleration (programmable networking/computing),
4. Networking (vertical and horizontal connectivity),
5. Control (deployment, actuation, mediation, and security)
Relevant Technologies
Key Architectural Paradigms
1. Cloudlets: These are virtual machine (VM)-based "micro-datacenters" co-
located within a Wi-Fi subnet.
2. Multi-access Edge Computing (MEC): Introduced by ETSI as a
telecommunications standard, MEC enables service providers to offer
virtualization-based services at cellular base transceiver stations (BTSs).
3. Mist Computing: This paradigm distributes computing mechanisms to the
"extreme edge" where IoT devices themselves are located.
Virtualization and Compute Technologies

FEC nodes utilize various models to provide "X as a Service" (XaaS):

1. Virtual Machines (VMs) and Containers: These are the primary methods
for hosting applications in a sandbox environment at the edge. Docker
Containers are specifically noted as a focused technology for virtualization
in FEC due to their lightweight nature compared to traditional VMs.
2. Service Models: Edge nodes support Infrastructure or Platform as a Service
(I/PaaS) and Software as a Service (SaaS). Specific SaaS offerings include
On-Demand Data Processing (ODP) and Context as a Service (CaaS).
Networking and Communication Protocols

To support both vertical (device-to-cloud) and horizontal (node-to-

node) connectivity, FEC employs a variety of protocols:
1. Standard IP Protocols: TCP/UDP sockets and HTTP.
2. IoT-Specific Protocols: CoAP (Constraint Application Protocol), MQTT
(Message Queue Telemetry Transport), XMPP, and AMQP.
3. Low-Power Wireless: Bluetooth Low Energy (BLE), ZigBee (IEEE 802.15.4),
and Z-Wave are used for energy-efficient communication with constrained
devices.
4. Software-Defined Networking (SDN) & NFV: Network Function
Virtualization (NFV) and SDN allow FEC nodes to manage multiple routing
tables in parallel and configure optimal transmission paths.
Hardware and Acceleration

• Programmable Hardware: FEC nodes use GPUs (Graphics Processing

Units) and FPGAs (Field Programmable Gate Arrays) to accelerate
complex algorithms and allow for runtime reconfiguration of sensors.
FPGAs are highlighted for being more energy-efficient than GPUs for
certain edge tasks.
• Constraint & Integrated Devices: These include microcontrollers like
the Atmel ATmega328 (used in Arduino) or more powerful ARM-based
processors found in smartphones and tablets.
Open Platforms and Frameworks

Several open-source projects aim to standardize FEC deployment:

• OpenStack++: A framework for providing VM-based cloudlet platforms on
regular x86 computers.
• Apache Edgent: An IBM-contributed runtime platform that allows for dynamic
data processing and migration between the cloud and the outer-edge.
• EdgeX Foundry: A Linux Foundation project providing a complete SDK to
facilitate industrial IoT connectivity.
• WSO2-IoT Server: An extension of the WSO2 integration platform that
provides FEC capabilities to outer-edge devices
Fog and Edge Computing Completing the
Cloud
• FEC acts as a complementary layer that fills the critical gap between
centralized cloud data centers and the physical "things" in an IoT
environment. Rather than replacing the cloud, FEC extends its
capabilities to create a service continuum.
Advantages of FEC:
• SCALE,
• Security,
• Cognition,
• Agility,
• Latency,
• Efficiency,
How FEC Achieves SCANC
1. Storage (Caching)- The storage mechanism in FEC involves temporary data storing and caching at the edge nodes. This improves content delivery
performance by placing data nearer to the end-user. For instance, content service providers can cache multimedia data at roadside FEC nodes to improve
the "quality of experience" for users in connected vehicles.
2. Compute- FEC nodes provide computing power through two primary models:
• Infrastructure or Platform as a Service (I/PaaS): This allows clients to deploy customized software within sandbox environments like virtual machines (VMs) or container
engines.
• Software as a Service (SaaS): This includes On-demand Data Processing (ODP), where nodes process client requests using pre-installed methods, and Context as a Service
(CaaS), where nodes collect and process data to generate meaningful situational information.
3. Acceleration- Acceleration in FEC is built on the concept of programmability and is divided into two aspects:
• Networking Acceleration: Utilizing network virtualization and Software-Defined Networking (SDN), FEC nodes can operate multiple routing tables in parallel to find the most
efficient transmission path for data.
• Computing Acceleration: Nodes use advanced hardware like Graphics Processing Units (GPUs) or Field Programmable Gate Arrays (FPGAs) to speed up complex algorithms.
FPGAs are noted for being more energy-efficient and allowing for the runtime reconfiguration of sensors.
4. Networking- FEC manages both vertical and horizontal connectivity to bridge the gap between devices and the cloud.
• Vertical Networking: Connects things to the cloud using standard IP-based protocols such as TCP/UDP, HTTP, CoAP, MQTT, XMPP, and AMQP. FEC nodes can act as proxies
or message brokers to facilitate this communication.
• Horizontal Networking: Connects nodes and devices to one another using energy-efficient, often non-IP protocols like Bluetooth, ZigBee (IEEE 802.15.4), and Z-Wave.
5. Control- The control mechanism consists of four essential types:
• Deployment Control: Allows for the dynamic (re)deployment of software programs across the network hierarchy.
• Actuation Control: Enables the cloud to delegate decision-making to FEC nodes, which then directly control the behavior of IoT devices.
• Mediation Control: Facilitates interaction and interoperability between devices and entities owned by different service providers.
• Security Control: Manages the authentication, authorization, and identity of users while protecting the virtualized runtime environment on the FEC nodes.
Hierarchy of Fog and Edge Computing
From the perspective of the central cloud in a core network, Fog and
Edge Computing (FEC) is structured into three layers: inner-edge,
middle-edge, and outer-edge. Together, these layers create a
continuous service path from centralized cloud data centres to end IoT
devices.
1. Inner-Edge
• The inner-edge (near-edge) includes large-scale regional or national
Wide Area Networks (WANs) operated by enterprises and ISPs. It
contains core data centers, evolved packet cores (EPC), and
metropolitan networks. While it originally focused on basic
connectivity, it now supports distributed caching and processing to
improve service performance and user experience. WAN-based cloud
data centers at this level act as the “fog” layer.
2. Middle-Edge
• The middle-edge represents the main operational layer of FEC and
includes:
• Local Area Networks (LANs): Ethernet, Wi-Fi, and campus networks
where local servers or gateways provide cloud-like services (often
called cloudlets or local clouds).
• Cellular Networks: Base stations such as macro, micro, pico, and
femtocells enhanced with Multi-access Edge Computing (MEC),
enabling fast real-time processing for applications like connected
vehicles and augmented reality.
3. Outer-Edge
The outer-edge (extreme-edge or mist) is the closest layer to IoT
devices and includes:
• Constrained Devices: Low-power sensors and actuators with limited
processing capability.
• Integrated Devices: More capable devices like smartphones and
tablets that support sensing and cloud interaction.
• IP Gateway Devices: Devices such as Raspberry Pi that connect
constrained devices to higher layers and may run lightweight
virtualization (e.g., Docker) to function as FEC nodes.
Business Models in Fog and Edge Computing
(FEC)
Fog and Edge Computing (FEC) business models define how edge
resources are delivered and monetized.
They enable a decentralized ecosystem where cloud providers,
enterprises, and even individuals can offer services through open
platforms and software tools.
Three main models are identified: X as a Service (XaaS), Support
Service, and Application Service.
1. X as a Service (XaaS)
This model delivers computing resources such as infrastructure,
platforms, software, storage, and networking over the network.
Customers pay to use hardware resources that support Storage,
Compute, Acceleration, Networking, and Control (SCANC).
It includes models like IaaS, PaaS, SaaS, and Storage/Caching as a
Service (S/CaaS).
Examples include distributed computing networks and shared Wi-Fi
platforms that allow individuals to monetize their equipment.
2. Support Service
This model provides technical support and software management for
clients who already own edge hardware.
Services include installation, configuration, maintenance, and updates.
Although some vendors bundle support with their hardware, the
industry trend is toward hardware-independent support services.
3. Application Service
This model offers complete, ready-to-use edge-based solutions without
requiring clients to manage the underlying infrastructure.
It is suitable for complex environments such as industrial systems,
smart cities, transportation, and healthcare.
Examples include digital twin systems, real-time traffic management,
and edge-based patient monitoring solutions.
Out-of-Box Experience
The Out-of-Box Experience (OOBE) in the context of FEC refers to the
ease of setup and immediate usability of FEC-ready products such as
routers, switches, and IP gateways.
As the market for FEC hardware is projected to grow significantly, OOBE
is seen as a key feature to simplify the complex task of system
management.
It is primarily delivered in two forms: OOBE-based equipment and
OOBE-based software.
OOBE-Based Equipment

This refers to hardware products where the vendor has already

integrated the FEC runtime platform into the device itself.
Because the platform is built-in, users can easily configure and deploy
FEC applications using graphical user interfaces, similar to the
experience of setting up a standard commercial router.
This approach removes the need for users to perform difficult, low-
level technical configurations to get their edge computing environment
running
OOBE-Based Software
OOBE-based software is a solution for users who already own FEC-
compatible devices and wish to enable an FEC runtime environment.
Challenges:
• Standardization: Difficulty in deciding which FEC platforms and software
packages should be included for a universal experience.
• Compatibility: Ensuring the software works across heterogeneous hardware
(such as x86 and ARM architectures) is complex and requires vendors to
provide multiple versions.
• Cost: Without common industry standards, developing and maintaining this
type of software is extremely expensive for vendors