Autonomous Systems

The Internet is a network of networks and Autonomous Systems are the big networks that make up the Internet. More specifically, an autonomous system (AS) is a large network or group of networks that has a unified routing policy. Every computer or device that connects to the Internet is connected to an AS.

Imagine an AS as being like a town's post office. Mail goes from post office to post office until it reaches the right town, and that town's post office will then deliver the mail within that town. Similarly, data packets cross the Internet by hopping from AS to AS until they reach the AS that contains their destination Internet Protocol (IP) address. Routers within that AS send the packet to the IP address.

Every AS controls a specific set of IP addresses, just as every town's post office is responsible for delivering mail to all the addresses within that town. The range of IP addresses that a given AS has control over is called their IP address space.

Autonomy requires that the system be able to do the following:

• Sense the environment and keep track of the system’s current state and location.
• Perceive and understand disparate data sources.
• Determine what action to take next and make a plan.

Act only when it is safe to do so, avoiding situations that pose a risk to human safety, property or the autonomous system itself.

Examples of Autonomous Systems

Autonomous Robots

Autonomous robots vary from simple robot floor cleaners to complex autonomous helicopters. Otto, the first autonomous snowplow in North America, keeps runways clear at an airport in Manitoba. 

Autonomous Warehouse and Factory Systems

From mail sorting systems to material conveyors to assembly robots, a diverse array of autonomous systems performs routine and repetitive tasks, enabling better use of human labor. One type of warehouse autonomous system is a robot forklift that moves products around an ecommerce giant’s automated distribution center. On assembly lines, autonomous factory robot arms perform many heavy and precision tasks such as arc welding, painting, finishing and packaging.

Autonomous Drones

Unmanned aerial vehicles, known as UAVs or drones, are small self-piloting autonomous aircraft. Drones have long been used for reconnaissance, surveying, asset inspection and environmental studies. Two common uses for drones are agriculture and oil well inspection.

Sensors and Sensor Fusion

Sensors and sensor fusion play a vital role in autonomous systems. They enable such systems to gather data from sources in the environment and make use of the data to plan and take action. In this section, you’ll learn about the diverse types of sensors used in autonomous systems and how sensor fusion helps an autonomous system acquire and develop a more accurate assessment of its environment.

Edge AI

Edge AI involves the deployment of artificial intelligence (AI) algorithms and models directly on edge devices. An edge device is a physical, remote computing device that’s connected to the network edge, such as smartphones, IoT devices, and embedded systems. This approach enables smarter, faster, and more secure processing on the devices closest to the data source, and without relying on cloud-based processing.

Edge AI allows responses to be delivered almost instantly. With edge AI, data is processed within milliseconds providing real-time feedback with or without internet connection because AI algorithms can process data closer to the location of the device. This process can be more secure when it comes to data because sensitive data never leaves the edge.

Use cases of Edge AI

Smart Homes, Cities and Infrastructure: Edge AI plays a crucial role in building smarter and more efficient homes and cities, enabling analysis and processing of vast amounts of data from sensors, cameras, and other IoT devices in real time.

Industrial IoT: By embedding AI capabilities into edge devices, such as robots and machines, tasks that require real-time processing and decision-making can be performed locally, resulting in improved productivity, increased safety, and better overall performance.

Autonomous Vehicles: By using real-time processing of data from sensors like cameras, LiDAR, and radar, edge AI enables AI-powered vehicles to make decisions critical for safety and efficiency.

Importance of Edge AI

Edge AI is revolutionizing various industries by bringing advanced computing capabilities directly to the edge. With the increased demand for edge devices to think for themselves, edge AI brings intelligence and real time analytics to even the smallest edge devices.

Edge AI offers several advantages over traditional AI approaches:

• Minimize latency by reducing the time delay involved in sending data to the cloud, crucial for real-time applications.
• Improve overall system performance with real-time data processing for discission critical applications.
• Reduce the power budget and increase battery life to maximize device operation.
• Reduce reliance on cloud connectivity and increase autonomy in remote or network-constrained use cases.
• Enhances privacy and security by avoiding the transmission of data between systems.
• Reduce cost and network congestion by using less bandwidth.

Benefits of edge AI

Less power use: Save energy cost with local data processes and lower power requirements for running AI at the edge compared to cloud data centers

Reduced bandwidth: Reduce the amount of data needed to be sent and decrease costs with more data processed, analyzed, and stored locally instead of being sent to the cloud

Privacy: Lower the risk of  sensitive data getting out with data being processed on edge devices from edge AI

Security: Prioritize important data transfer by processing and storing data in an edge network or filtering redundant and unneeded data

Scalability: Easily scale systems with cloud-based platforms and native edge capability on original equipment manufacturer (OEM) equipment

Reduced latency: Decrease the time it takes to process data on a cloud platform and analyze it locally to allow other tasks

Network Function Virtualization (NFV)

Network Function Virtualization (NFV) is the replacement of network appliance hardware with virtual machines. The virtual machines use a hypervisor to run networking software and processes such as routing and load balancing.

With the help of NFV, it becomes possible to separate communication services from specialized hardware like routers and firewalls. This eliminates the need for buying new hardware and network operations can offer new services on demand. With this, it is possible to deploy network components in a matter of hours as opposed to months as with conventional networking. Furthermore, the virtualized services can run on less expensive generic servers.

Additional reasons to use network functions virtualization include:

• Pay-as-you-go: Pay-as-you-go NFV models can reduce costs because businesses pay only for what they need.
• Fewer appliances: Because NFV runs on virtual machines instead of physical machines, fewer appliances are necessary and operational costs are lower.
• Scalability: Scaling the network architecture with virtual machines is faster and easier, and it does not require purchasing additional hardware.

Risks of network functions virtualization

Physical security controls are not effective: Virtualizing network components increases their vulnerability to new kinds of attacks compared to physical equipment that is locked in a data center.

Malware is difficult to isolate and contain: It is easier for malware to travel among virtual components that are all running off of one virtual machine than between hardware components that can be isolated or physically separated.

Network traffic is less transparent: Traditional traffic monitoring tools have a hard time spotting potentially malicious anomalies within network traffic that is traveling east-west between virtual machines, so NFV requires more fine-grained security solutions.

Complex layers require multiple forms of security: Network functions virtualization environments are inherently complex, with multiple layers that are hard to secure with blanket security policies.

Advantages of network functions virtualization

• Lower expenses as it follows Pay as you go which implies companies only pay for what they require.
• Less equipment as it works on virtual machines rather than actual machines which leads to fewer appliances, which lowers operating expenses as well.
• Scalability of network architecture is quite quick and simple using virtual functions in NFV. As a result, it does not call for the purchase of more hardware.

Benefits of network functions virtualization

• Many service providers believe that advantages outweigh the issues of NFV.  
• Traditional hardware-based networks are time-consuming as these require network administrators to buy specialized hardware units, manually configure them, then join them to form a network. For this skilled or well-equipped worker is required.
• It costs less as it works under the management of a hypervisor, which is significantly less expensive than buying specialized hardware that serves the same purpose.
• Easy to configure and administer the network because of a virtualized network. As a result, network capabilities may be updated or added instantly.

Software-Defined Networking (SDN)

Software defined networking (SDN) is an approach to network management that enables dynamic, programmatically efficient network configuration to improve network performance and monitoring. It is a new way of managing computer networks that makes them easier and more flexible to control.

In traditional networks, the hardware (like routers and switches) decides how data moves through the network, but SDN changes this by moving the decision-making to a central software system. This is done by separating the control plane (which decides where traffic is sent) from the data plane (which moves packets to the selected destination).

Importance of Software-Defined Networking

Increased control with greater speed and flexibility: Instead of manually programming multiple vendor-specific hardware devices, developers can control the flow of traffic over a network simply by programming an open standard software-based controller. Networking administrators also have more flexibility in choosing networking equipment, since they can choose a single protocol to communicate with any number of hardware devices through a central controller.

Customizable network infrastructure: With a software-defined network, administrators can configure network services and allocate virtual resources to change the network infrastructure in real time through one centralized location. This allows network administrators to optimize the flow of data through the network and prioritize applications that require more availability.

Robust security: A software-defined network delivers visibility into the entire network, providing a more holistic view of security threats. With the proliferation of smart devices that connect to the internet, SDN offers clear advantages over traditional networking. Operators can create separate zones for devices that require different levels of security, or immediately quarantine compromised devices so that they cannot infect the rest of the network.

Benefits of software-defined networking

Simplified network management and control

SDN helps to simplify network management for IT teams. A network administrator needs to deal with only one centralized controller to configure and manage all connected devices. This approach is a radical departure from traditional networking, where configuring multiple devices individually is the norm.

End-to-end visibility into networks

This makes device configurations, resource provisioning and management easier. It also enables IT teams to easily monitor network health and act quickly to increase network capacity as business requirements change.

Stronger network security

Centralized, software-defined network also provides a security advantage for organizations. The SDN controller can monitor traffic and deploy security policies. If the controller deems traffic suspicious, for example, it can reroute or drop the packets. Also, admins can easily implement security policies across the entire network to increase its ability to withstand threats.

Simplified policy changes

With SDN, an administrator can change any network switch's rules when necessary -- prioritizing, deprioritizing or even blocking specific types of packets with a granular level of control and security.

This capability is especially helpful in a cloud computing multi-tenant architecture, as it enables the administrator to manage traffic loads in a flexible and efficient manner. Essentially, this enables administrators to use less expensive commodity switches and have more control over network traffic flows.

Reduced hardware footprint and Opex

SDN virtualizes hardware and services that were previously carried out by dedicated hardware. Also, administrators can use open source controllers instead of costly vendor-specific devices. This reduces the organization's hardware footprint and lowers operational costs.

Networking innovations

SDN also contributed to the emergence of software-defined wide area network (SD-WAN) technology. SD-WAN employs the virtual overlay aspect of SDN technology. SD-WAN abstracts an organization's connectivity links throughout its WAN, creating a virtual network that can use whichever connection the controller deems fit to send traffic to. By adopting this technology, organizations can programmatically configure their network topology in a WAN. Also, SD-WAN can better handle large amounts of traffic and multiple connectivity types compared to traditional WANs.

Challenges of SDN

Security

Security is both a benefit and a concern with SDN technology. The centralized SDN controller presents a single point of failure and, if targeted by an attacker, can prove detrimental to the network.

Controller redundancy costs

Implementing controller redundancy is one way to minimize the risk of a single point of failure. However, this can entail an additional cost.

Unclear definition

Another challenge with SDN is the industry really has no established definition of software-defined networking. Different vendors offer various approaches to SDN, ranging from hardware-centric models and virtualization platforms to hyperconverged networking designs and controllerless methods.

Market confusion

Some networking initiatives are often mistaken for SDN, including white box networking, network disaggregation, network automation and programmable networking. While SDN can benefit and work with these technologies and processes, it remains a separate technology.

Slow adoption and costs

SDN technology emerged with a lot of hype around 2011 when it was introduced alongside the OpenFlow protocol. Since then, adoption has been relatively slow, especially among enterprises with smaller networks and fewer resources. Many enterprises cite the cost of SDN deployment to be a deterring factor.