Network Analytics and AI: How to choose a solution that best aligns with your adoption strategy
Artificial intelligence (AI) is undoubtedly a very hot topic across many industry segments and verticals, as one analyst even points out “it will empower the fourth industrial revolution”. In the telco world, AI when used together with analytics, is now associated with powering intelligent and “self-aware” digital networks (which, Ciena and Blue Planet refer to as the Adaptive Network). Presently, there are substantial discussions and write-ups around the topic of AI-driven analytics, however, our VP of Blue Planet Solutions & Engineering, Kailem Anderson made a very important point in one of his recent blogs – that the term “AI” must avoid becoming a cliché.
Kailem points out that service providers would be understandably apprehensive in “giving control of a complex network to a machine” but that there are clear opportunities. Further he points out that coming up with a clear adoption strategy is one of the most important first steps to reaping the full benefits of the associated technologies. So what should this strategy should look like?
First, one must define what AI really means from a practical standpoint and place deep thought into how it can be used to solve real business problems. Second, one must understand where the most benefit and ROI lies in the short term and take that initial step of deployment. And third, formulate plans to expand from there as one gains more experience, knowledge and confidence about the network “running” on its own, so to speak.
So far, so good. The natural question then is: how do you pick the right vendor and solution that’s capable of helping you align with this strategy? To answer this question, we’ll elaborate a bit more on each of these strategic aspects, then discuss the factors you should consider when choosing a solution:
1. Gaining an understanding of AI and what it means to your business:
As Kailem mentions, AI is really more about its objective rather than its precise technical definition. Generally speaking, the main objective of AI when combined with analytics is to allow “machines” to perceive what is happening around them, “think” about the situation and make decisions --- essentially performing like humans, but at “ superhuman” speed and scale.
The value AI provides can be measured in terms such as increased revenue, higher net promoter scores, and cost savings. So what are the specific areas of service provider business and network operations that would benefit from superhuman capabilities? Fortunately, there is already a wealth of information that points to a number of these areas.
For example, a recent analyst report from Ovum1 states that analytics and AI are not just "nice-to-haves," but are "must-haves" in the digital market, and that big data is essential to enabling the insights that allow for efficient operations. It also describes how managing today’s hybrid-network environment requires the consolidation of data from all data sources – both physical and virtual – and how the cross analysis of these datasets help rapidly identify issues so operators can effectively respond to them. Even though business and network scenarios may differ from provider to provider, these reports can help service providers jumpstart their individual assessments.
Further, many software and hardware vendors have already released and announced AI based network analytics solutions; however, when choosing a vendor to help you clarify the exact role of AI for your particular business, it is prudent to consider these following questions:
- Does the vendor have a true understanding of the service provider network and its multiple parts, which is the core medium through which you operate your business and transport services?
- How much experience and/or credibility do they have in software itself? How about in analytics and AI – the technologies that will serve as the “brain” that drives intelligence?
- Given the significance of the project and potential risks involved (new technology, new application) -- does the vendor have the capability and resources to stand behind you throughout the entire project lifecycle?
2. Projecting tangible business value and benefits
As a service provider, your readiness to adopt new technologies probably depends on a variety of factors, one of which would be the financial investment required. This is all the more reason that it would make more sense for some providers to start out small before venturing into more expansive domains - especially those that do experience feelings of “apprehension” around the term AI as mentioned earlier. But no matter how big or small of a project, the ability to calculate its ROI is crucial. It is important then to ask these questions:
- Does the vendor/solution provide the ability and flexibility for you to take a controlled, manageable and phased approach to deployment?
- Do they provide financial modeling services to help ensure you are making well-informed investment decisions along the way?
- Can they provide any ROI examples and/or models or have current customers successfully benefiting from their solutions and/or services that are specific to analytics and AI?
3. Beyond initial deployment, expanding into new domains as business needs evolve
We all know that making changes to the network is a lengthy process involving multiple phases of integration, testing and collaborative efforts. When it comes to incorporating a technology like AI, the initial process may feel even more intensive and burdensome, as this is a project that over time is meant to give your network increasing power and capability to make decisions and take actions on its own.
It is important however, to remember that the real objective here is not about building a fully autonomous network, but rather a network that is intelligent, automated and adaptive, with AI and machine learning applied in areas where it provides the most value, while giving humans the ability to take full control at any time. With this in mind, once the initial project is successfully deployed and benefits are being realized, the natural next step is to expand its application with the goal of reaping further benefits. In this light, the following questions should be taken into consideration:
- Does the solution provide a firm foundation for expansion and is it flexible enough to grow and adapt to my changing business priorities without significant disruption?
- Will the vendor/solution enable me to acquire deep knowledge and expertise so I can program my own network and take control of its transformation?
- Will I be provided with the technical and expert resources I need, where and when I need them?
While these points are general guidelines, they are intended to help you get started on your journey towards digital transformation and adaptive networking that can help shape the growth of your business for decades to come. With this, we invite you to explore our white paper, Making Intelligent Automation a Reality with Advanced Analytics and Machine Learning. This paper offers great insight into the role of analytics, AI and machine learning in telecom networks, and the ultimate value you can gain from them. It also provides an overview of Blue Planet’s direction and vision for the future of intelligent networking. For questions, or if you require further information, you can always visit our website at www.blueplanet.com or contact the Blue Planet team.
1 "Digital Transformation World highlights the role AI and analytics will play in CSPs' network operations", Ovum, June 2018