Blog

How to bring mass customization to AIOps

Some of you may be considering embarking on an AIOps journey for your network or IT infrastructure. There are three different paths you can choose from. While all three possibilities are valid, it’s important to note that the convenience and advantages of one or the other will depend on your team’s specific goals and abilities. Let’s explore these options in more detail.

The first path is to use existing commercial AIOps tools that will provide your teams with a basic level of configurability, usually in the form of rules. These tools address very common out-of-the-box use cases but are limited by rigid semantic definitions. If this is compatible with your specific use case, you will likely accomplish what you need in a few clicks in the GUI. For simplicity’s sake, let’s name this option ‘Build Your Own Rules’ (BYOR).

A second possibility is to build your platform leveraging open-source technologies and components. This path will give you the greatest level of flexibility and customizability for your specific use case. One drawback however is the need for highly skilled teams which can be costly and labor-intensive. While the offer of open-source projects for data management, data storage, analytics, and machine learning (ML) frameworks is very broad and extensive, the combined effort required to make them function as needed can provide a lot of value. In particular, the availability of many open-source ML frameworks creates the perception that it’s somewhat easy to build a system that uses machine learning. However, the key is not on the model, framework, or the parameters you use, but rather on all the others that were discarded in the research process. It’s not to say that building your platform is conceptually wrong, but the effort and investment required will only pay off if there is a level of custom behavior needed that no other alternative could provide. The ongoing investment to operate, maintain and evolve a custom-built platform will be considerable, and the commercial licenses for some of those open-source components will have to be considered to take full advantage of the hardened, secure, and scalable versions of them.

So, this begs the question: what is the third option? At Selector, we believe there is an alternative path to the options mentioned above. Selector has designed a data-centric network and IT operation platform that provides a level of programmability allowing you ‘Build Your Own Use Case’ for increased flexibility and scalability.

The Selector Analytics platform is built using a declarative architecture that allows us or you to create new use cases simply by configuration and using a human-friendly modeling language expressed in YAML. By configuration, it’s possible to activate and customize the behavior of all the different components of the platform, as if they were Lego blocks, to build the use case you need for your unique infrastructure. At Selector, we take care of all the heavy lifting to build and integrate all those Lego blocks in the most cost-efficient and effective way. With a ‘Build Your Own Use Case’, Selector provides the maximum level of customization with minimal effort and investment required.

For most of you familiar with the term Mass Customization, you may now have realized that this is the value Selector brings to AIOps. Mass Customization is a manufacturing technique commonly applied in the car industry, among others. The idea is to combine the flexibility and personalization of custom-made products with the low unit costs associated with mass production. So, you may be asking yourself, how is this relevant to AIOps? It depends on the “product” you need. For example, if you think you need “what everyone else needs” (basic car, same options, choice of color, or a basic observability tool where you can create some rules), then you fit in the bottom-right corner of the diagram below, ‘Build Your Own Rules’. However, if you require something highly customized where cost, maintenance, and effort don’t play a big factor in the overall decision-making process (i.e., purchasing a Rolls Royce), then you fit at the top-left of the diagram, ‘Build Your Own Platform’.

With Selector, you can leverage the benefits of customization, low cost, and agile delivery by applying the platform, reusability, and customizability principles of mass customization. This is the value Selector can provide to help you move along your AIOps journey. With the ‘Build Your Own Use Case’, you still get the behavior you need but minimize the delivery and recurring costs.

Every network and IT infrastructure is unique, and this is especially true for larger, more complex infrastructures. Each organization has different constraints, networks, and IT design, and as a result, deployment and operational decisions should be made accordingly. With careful consideration and planning, you can strike the right balance between customizability, cost, effort, and scalability that will help you successfully embark on your AIOps journey.