What is a key feature of the assurance element of the intent based networking mode?

being specifically programmed for that task. This is especially useful when designing and programming specific algorithms is difficult or infeasible. Examples of such tasks in computer science include malicious code detection, network intruder detection, optical character recognition, computer speech recognition, and computer vision. One objective of learning is to be able to generalize based on experience. For machines, this involves the ability to perform accurately on new, previously unseen tasks after gaining experience with a learning data set. The training data set must come from data that is representative of the larger data pool. This data pool enables the machine to build a general model about this data, which would help it make accurate predictions. ML in the IoT One of the features of the IoT is that it enables the collection of extremely large pools of data that can “teach” programs how to respond in certain conditions. Some of the more common uses of ML technology include:  Speech Recognition - Many different companies now offer digital assistants which allow you to use speech to communicate with a computer system. Apple, Microsoft, Google and Amazon all offer this service. These companies not only allow commands to be given verbally, but offer speech-to-text capabilities.  Product Recommendation - Systems build up a customer profile and recommend products or services based on previous patterns. Users of Amazon and eBay receive recommendations on products. Organizations such as LinkedIn, Facebook, and GooglePlus recommend users you may wish to connect with.  Shape Recognition - Programs exist that allow crude hand-drawn diagrams and notes to be converted to more formal diagrams and text. This allows the shapes and lines of hand writing to be converted to more formal text which can then be searched and analyzed.  Credit Card Fraud Detection - A profile is constructed about the purchasing patterns of a client. Any deviation from these patterns triggers an alert and the system automatically takes action. This action ranges from denying the transaction to notifying the authorities. Some of the events that are detected and could indicate a fraudulent transaction include purchasing products not normally purchased, purchases in a different geographic area, rapidly purchasing many different products, and purchasing large- ticket items.  Facial Recognition - Security cameras are everywhere, from stores and streets to airports and transportation hubs. These cameras continually scan the crowds, normally watching for dangerous or illegal activities, but they can also be used to identify and track individuals. The system builds a pattern of specific facial features and then watches for a match to these facial patterns triggering some action. Think about your interactions with online and offline systems over the past week. How many ML applications have you interacted with? What Is Intent-Based Networking (IBN) For a business to survive, it must be agile and respond quickly to the needs and demands of its customers. Businesses are increasingly dependent on their digital resources to meet customer demands, so the underlying IT network must also be responsive enough to quickly adapt to these requirements. This normally involves adjustments to many systems and processes. These adjustments may include changes to security policies and procedures, business services and applications, and operational policies. With traditional networks, many different components must be manually adjusted to meet ever-changing business requirements. This requires different technicians and engineers to ensure that the systems are changed in a manner that allows them to work together to accomplish their goal. This sometimes results in errors and delays, and often in sub-optimal network performance. The new business network must seamlessly and securely integrate IoT devices, cloud-based services, and remote offices in an agile, responsive, and business-relevant manner. Additionally, the network must secure these new digital initiatives from the ever-changing threat landscape. To address this need, the IT industry has initiated an effort to create a systematic approach to tie infrastructure management to business intent. This approach is known as intent-based networking. The figure illustrates the general idea behind intent-based networking. With this new paradigm, business needs are automatically and continually translated into IT infrastructure execution. How are ML, AI, and IBN Linked? Intent-based networking harnesses the power of automation, AI, and ML to control the function of a network to accomplish a specific purpose, or intent. Intent-based networking allows the IT team to specify, in plain language, exactly what they want the network to accomplish and the network makes it happen. The network is able to translate the intent into policies and then use automation to deploy the appropriate configurations required across the network. The intent-based network uses AI and ML to ensure that any services that are deployed meet the required service level. If they do not meet the service level, the intent-based network can make alerts and provide suggestions for improvement. In some cases, the intent- based network can automatically reconfigure the network to comply with the service levels. The intent-based networking model shown in the figure consists of three key elements:  Assurance - The assurance element is end-to-end verification of network-wide behavior. It predicts the results of any changes, tracks compliance with the original intent, and makes recommendations or adjustments when there is a misalignment between the intent and the outcome. This stage relies heavily on AI and ML. Systems are part of a closed-loop that continually monitors performance and security of the network, and reconfigures the network to ensure compliance.  Translation - The translation element is the ability to apply business intent to network configuration. The intent is what you wish to accomplish, not how it is accomplished. This intent is specified in plain language and used by the system to create policies across the system. For example, an intent might be to segment guest traffic from corporate traffic, or to enable access for remote users.  Activation - The activation element occurs after the intent has been specified and the policies created. This is when individual devices are provisioned to match the intent-based policies. This can be an automated or semi-automated mode that allows the network team to verify configuration before the devices are deployed. An intent-based network creates an agile, responsive network that scales easily and adapts to meet business requirements. It makes efficient use of highly-skilled resources and allows man and machine to work together to optimize the customer experience. Additionally, intent-based networking provides a more secure digital experience by automating time consuming or complicated processes. This makes deploying security policies much easier. Use Cases for Intent-Based Networking Intent-based networking allows the company to focus on business goals. It provides an automated system that understands what the organization needs and then makes it happen. The Cisco Digital Network Architecture (Cisco DNA) is an example of an intent-based network. It is an open, extensible, software- driven architecture. It accelerates and simplifies enterprise network operations, while lowering costs and reducing risks. Cisco DNA automation and assurance are built on a software-defined networking (SDN) controller, rich contextual analytics, network virtualization, and the limitless scalability of the cloud. For more information on how Cisco Digital Network Architecture is being used across different industry verticals visit here. Summary This chapter began by discussing automation.