Performance monitoring in the distributed cloud: does it matter?

We are at a point in the evolution of our technical understanding of 5G networks. We can safely say that it is a very different ‘ball game’ as compared to earlier mobile network deployments from an architectural and service assurance perspective.

In short, the network-centric approach to SLAs and the generic assurance practices for service assurance are long gone based on the new Service Based Architecture (SBA) defined by 5G.

Architecture requirements for 5G Service-Based Architecture

With 5G, the services define the architecture requirements where the distributed compute, network, and cloud-based service environment is responsible for assuring the resource requirements. This way the SBA enables service and application traffic to flow in the most feasible logical path; and it also opens the door for a dynamic service chain creation where logical paths, called slices, can be created on-demand. This allows for the creation of personalized logical network paths allowing services to traverse the most optimal physical and virtual compute/network entities in the architecture based on their SLAs.

In this environment, these service classes and SLAs are related to well-defined, stringent parameters relevant to individual industry verticals for the associated applications and customer segments.

How to ensure you’re meeting those SLAs

So, how can we do all of this in a scalable and cost-effective manner? And keep in mind that we also need to remember that in this new service model, one of the critical enablers and challenges is the ability to assure the edge cloud/compute instances and service traffic.

Part of this complexity is derived from the need for assurance at all layers of the network to identify areas where improvement is needed and also where new investments should go in the existing network.

One of the main problems lies in meeting the latency budget for services from an end to end perspective. In this type of mesh and segmented environment, the only way to achieve this is to realize that we need to measure the network in a segmented fashion – both horizontally and vertically – to identify where the actual cause of latency impacting issues in the underlay is originating. Measuring the latency segments will not be enough. 

At the same time, we need to be able to understand what is happening in the overlay, service, and application layers and correlate them to identify the impact to the end user, application, developer, or other entities that will be leveraging this network.

Why is this important for wireless and cloud service providers right now?

In short, the initial services and applications are likely to address specific markets to reap the quickest return on investment. Services such as virtual gaming, wireless broadband, and distributed application engineering, for example, will be at the forefront of service offerings. They will be the basis for how end users evaluate and perceive the present and future ‘quality’ and the capabilities of the 5G wireless and cloud service provider offerings.

It is evident at this point that the cloud and the wireless industry are bound together through 5G. They must work together to achieve the level of SLA assurance and quality of experience (QoE) expected and to differentiate their services in this new market.

For example, the ability to differentiate the difference of perhaps 1ms of latency between service layers, the edge compute, and the traffic path, understanding how that may impact the customer, and being able to identify the issue to resolve it before it has a significant impact will be the game changer for everyone. It’s all about the quality of the service offered and assured.

To attain this level of quality and visibility, service providers will need to work together to identify how they can drive value from the edge-cloud and general service layers. One aspect will be to convert raw performance data to enriched data to create high-quality “smart data” that is meaningful for each domain. This ‘smart data’ can then be used to look at the end-to-end architecture and services in a holistic way to gain real-time ‘smart’ visibility into their network, the services, the technologies, the network segments, and the QoE of their subscribers.

Back to the question of “performance monitoring in the distributed cloud: does it matter?” – it is safe to say that service providers for the wireless, cloud and their customers will not survive without a next-generation solution that can address the assurance and performance management requirements for the new 5G SBA. YES, IT MATTERS!