Service assurance has been around for decades – I built service assurance solutions early in my career when BT launched Frame Relay services in the US over thirty years ago. There have always been challenges in delivering service assurance solutions and these challenges have been getting exponentially more difficult in recent years. Fortunately, technology is also evolving to address these challenges, and this ensures that for every problem there is a solution. The trick as an operator is to ensure you choose the right tools from the toolbox for service assurance, so you address both current and upcoming requirements. In this blog I want to cover real world networks – which are not always pretty with increasing requirements, and we will look at the key characteristics of effective service assurance solutions for these real-world problems.
Let’s start by considering how real networks are built. There has been a huge amount of focus over the past several years on the disaggregated network. This is where solutions such as SDN/NFV help separate the delivery of traffic from its routing, management, and control. Such solutions offer many advantages and are seeing some significant deployment – especially in core networks. However real networks have been built over time and most operators of any size are likely to have to manage services running over several generations of network technology. They may have fiber or DSL or Coax or microwave or wireless in the access domain. They may have DWDM or SONET/SDH or IP/MPLS in the metro and core in addition to any disaggregated network solutions they have deployed. Even more challenging is the fact that this network that was built up over multiple decades is almost certainly multi-vendor and likely multi-generational for several of these vendors.
Let’s talk about how operators deliver multiple services over their networks from voice, video, and broadband services to residential consumers to an extensive range of networking services for businesses. Business services may include high-volume voice, video, and data as well as networking. Business services may also include hosted services; data-centre services, cell tower backhaul, monitoring and probe connectivity. With the evolution of IoT and digital transformation across almost all industrial segments, the number of services is exploding. We are also seeing an explosion in new service delivery models including cloud-based services and edge compute. All these services are delivered over the same multi-generational, multi-technology, multi-vendor networks.
The business of offering services to any audience requires commitment in some form required by regulators, customer SLA’s or set by competitive expectation. At this point I usually question the whole use of the term service management. In many deployments the main focus is network management – that is the monitoring of values and events in the network. In many cases the ability to correlate between what is happening on the network and the service impact is at best manual, which means the response is delayed and at worst does not exist at all. Businesses are waiting for their customers to tell them when they have failed to meet a Service Level Agreement (SLA), and have limited means to independently verify whether they have met or failed to meet such commitments. They may also be over building networks just to try to ensure they meet commitments meaning their infrastructure may be costing more than it needs to for the services carried over it.
Of course, in very small networks some of this can be handled by having very clever network engineers who “just know” how the network is built and how it carries services. But in networks of any size these very smart people cannot handle the scale and complexity – even with spreadsheets and freeware tools.
We also know that network equipment companies can offer very elegant tools, but these tend to either work only for that vendor’s equipment or work best for that vendor’s networks. These tools are not the answer for a multi-vendor network, and they struggle managing anything outside their own portfolio.
In the real network we are on a journey towards fully optimized service assurance. We are not at the start, but we are certainly not at the end. We know that real networks are not ubiquitous – rather they are built using multiple generations of multiple technologies from multiple vendors. The complexity of networks and the types of services they must support are such that individuals with simple tools cannot scale; legacy tools cannot cope, and network equipment vendor solutions work only when that vendor delivers most of the network and operators are willing to accept vendor lock in. Mapping from networks to services requires an understanding of network inventory and topology and an ability to normalize data and understand what it means – including cutting through verbosity.
From my formative years at BT building service assurance solutions for Frame relay to today building OSS solutions at Enghouse networks, the technology stack has changed massively. That said, the need to understand what is happening in the real network and map this to the services running over it remains the same objective. To learn more, take a look at our service assurance solutions to see how you can keep up with the latest innovations that can reflect the real-time reality of real-world networks.
