Updated: Sep 3, 2021
Intent-based networking (IBN) is a form of network infrastructure and innovative concept that uses artificial intelligence (AI), network orchestration, and machine learning (ML) to automate administrative tasks across a computer network.
In other words, it has been years that network engineers are deploying networks manually with no unified tool. Each network architect is using their own "standards" and best practices to come up with a network design for businesses that are targeting to be profitable in the future.
This means, network engineers and architects should be able to forecast what business will look like in the future, so they can come up with a good design to address those growth points. Normally, this does not happen.
The goal of Intent-based networking (IBN) is to reduce the complexity of creating, managing, and enforcing network policies and reduce the manual labor associated with traditional configuration management.
For example, an IBN command may look like this:
Allow accounting applications to access server ABC, but do not allow manufacturing applications to access.
The IBN management application will then determine which devices and routes match the business intention and make the appropriate configuration changes automatically.
Intent-based networking and software-defined networking (SDN) are similar in many aspects. Both approaches rely on a centralized controller to manage distributed devices on the network instead of individually managing each device from its own management console. Both approaches have the ability to understand network configuration and interaction across multiple devices.
Where the two approaches differ, however, is in how they are addressed at the administrator level. SDN continues to have a device-centric view of the network and commands are primarily about how devices should operate, but intent-based networking commands are issued from a business perspective. This second-level abstraction is the primary difference when it comes to intent-based networking vs. SDN.
Intent-Based Networking will indeed transform the way organizations understand enterprise networking. Intent-Based Networking will disrupt the networking infrastructure market as smart self-drive autonomous cars are disrupting transportation.
It has been about more than 30 years that network engineers and network architects are designing and deploying networks. The process is very straight forward. A team of network engineers sit down together and come up with different ways of connecting internal networks and branches to headquarter. If they are lucky enough and the IT organization is embedded in business processes and strategy, there will be a more coherent strategy around connecting what to wear. But again, even when IT strategy is aligned with business strategy, the way that network engineers are doing their job is exactly the same as 30 years ago. Deploying a general purposed data center may take about months to be completed because of the hardware nature of all solutions.
Just to be clear, when I speak about enterprise networking, I am referring to infrastructure devices such as a switch, router, firewall, wireless access point, load balancer, storage, and server in a data center or campus environment. Of course, there are always some services that will be deployed on top of these hardware and software-based devices as well. Cloud meant to change this approach. Cloud could transform the hardware-based view of networking devices and could bring the same result in a software environment that engineers — we can call them architects now — could design, deploy and manage enterprise networks in a matter of minutes instead of days and months. Plus, the PAYG consumption-based model, gives organizations the flexibility to start simulating some of the solutions in a test environment isolated from the production with a high level of confidence. In addition, organizations can now think of different scenarios in terms of having disaster recovery sites in the cloud, to achieve business resiliency if the main data centers go down. DevOps was another shift in the industry. The way organizations structure development and operations team has been a big improvement in reducing the life cycle of applications and indeed networking environment. Automation and orchestration shifted our views from having pieces of software devices (the equivalent of traditional hardware 30 years ago) into a way of deploying and monitoring a network with code. Continues Implementation (CI) and shorter sprints of application release, cause architects to think differently about networking devices in a way that they started to deploy virtual data centers in a matter of minutes in the cloud. If there was a problem with the performance or functionality, no problem. Kill the whole data center and redeploy the fixed version again. Only minutes of downtime. Intent-Based Networking (IBN) will change this approach again. Or better to say will transform this mindset. Let’s have an analogy about cars and transportation. Think of DevOps as “automatic transmission” and Intent-Based Networking as a self-driving autonomous car. Let’s have a look at the following figure. This figure explains different levels of automation in the driving technology which basically level 0 is no automation and level 5 the highest level of automation which means autonomous driver-less car. In the enterprise networking world, we have been reached level 4 of automation with DevOps and Cloud technologies. However, level 5 is what has been known as Intent-Based Networking. This is the ultimate weapon of business leaders to use networking technologies integrated with business strategy. The way that Intent-Based Networking works is actually very easy to understand. As the input, the operator which is normally a developer, enters the networking requirements in plain English language, and then the Intent engine will translate that into a vendor-agnostic high level and low-level design. For the next step, the network will be deployed automatically in a multi-cloud environment and then continuously been monitored to make sure that it meets all SLAs and still is fit for purpose. I give you some examples. The operator can enter his intent as follows: “I need a centralized one site data center in AWS to deliver Microsoft Active Directory services for 10,000 users.” “I need two data centers, one in Azure and one in AWS, to deliver IoT applications with 4,000 edge routers.” “I need an AI-ready data center with 0% packet loss and low latency less than mil second to cater for user profiling AI application with the throughput of 1 TB.” In all these examples, the Intent engine translates these business requirements into a high-level and low-level design with a vendor-agnostic approach consisting of networking infrastructure devices from e.g. Cisco, Juniper, Mellanox etc. Then the intent engine will start to deploy the network as requested and after deployment, it will monitor the performance continuously. This way, we are dealing with zero-touch provisioning of network solutions and an effective way of enabling business strategy at the networking layer. Which was a challenge till now. There are leading vendors in this space that are moving towards Intent-Driven Networking technology including Cisco, Juniper and Huawei. There are several startups that are active in this space, but most of their focus is on monitoring Intent-Based Networks. It is still early to talk about market size and the opportunity here. Even Gartner is still hesitant of naming IBN a market. In my view, IBN is the future of networking and it is inevitable. Just like Tesla Motors which started by level 5 automation in-car space, IBN will be the dominant force in the networking space in 3 years' time. Network Engineers should watch this space closely as this job title may not exist in 3–5 years time . . .
Characteristics of Intent-Based Network IBN
Every intent-based networking system (IBNS) incorporates the following four aspects:
Translation and validation: The system can translate a given command or business intent into actions that the software can perform. Additionally, it verifies that the intent can be executed successfully in the first place.
Automated implementation: Once the intent or desired state is defined, the system will allocate network resources and enforce policies to meet the goal.
State awareness: The system will continuously gather and monitor data to reflect the current state of the network.
Assurance and dynamic optimization/remediation: Using machine learning, the system will implement and maintain the desired state of the network, applying automated corrective action if necessary. ML gives the network the ability to analyze, extract and learn from data dynamically.
Benefits of IBN
A few benefits of intent-based networking include:
Reduces the complexity of the management and maintenance of network policies.
Simplifies the deployment of additional network services.
Reduces labor associated with the traditional configuration of switches and routers.
Strengthens network security capabilities.
Improves agility of the entire network system.
Eliminates repetitive or error-prone coding associated with manual inputs.