We've all heard of the network effect. The network effect is a phenomenon whereby the value of good or service to a single user rises or falls relative to the number of other users in its network.
What is Network Effect?
We often cite how fax machines were only useful once most/all other organisations we dealt with also had a fax machine. Conversely, we can also plot the decline in the use of faxes against the network effect. We use the network effect to describe the benefits of a globally connected telephone network. We also use it to describe the rapid uptake in user / subscriber growth after a certain inflection point. The more users join the network, the more useful the network becomes, thus enticing further users to join. This is sometimes described using a hockey-stick growth chart.
Benefits and threats of the Network Effect
The success stories from the Internet era have been driven by the network effect. If you look at the largest companies in the world today by market capitalisation, you'll see that many tech companies have ridden the wave of the network effect. The primary benefit of the network effect is the increase in value to all other users when each incremental user is added to the network.
But what's often lost in the positive aura of the network effect is that it also has significant potential pitfalls. This is not just in diminishing relevance, such as the fax machine example cited above when people incrementally stop using faxes (how many of us still have a fax machine on our desks or in our offices?).
Instead our focus here is on situations when failures happen in highly inter-connected networks. In these cases, the network effect has the potential to bring us all down. Catastrophic failures in a network are just that - we’re massively impacted together instead of just being individual services breaking in isolation. One software glitch or one cascading network event can bring the whole network, or large parts of it, down.
We intrinsically know this though. It's the reason we build so much resilience into our networks. It's the reason telco networks are measured against, and expected to exceed, five-nines (99.999% availability, which equates to 5 mins and 15 seconds of downtime per year).
Part of this up-time responsibility falls upon the high-availability design of the network and related systems. But networks are never infallible. All hardware componentry have failure rates. All software has bugs and/or incomplete test coverage (ie unexpected or unforeseen scenarios).
How to protect against failures?
That's where OSS/BSS comes in. We are the insurance policy. We're needed when the network inevitably fails. We are the monitor, collecting telemetry on the health of the network, the services it carries and the data streaming across it. We are the emergency response unit, coordinating triage and rectification activities. We are the forensic analysis tools, supporting us to understand what happened in the event of a failure / degradation and then ensuring future remedies.
Think of us also like the emergency department of a hospital that is seeing, in some cases, millions of patients streaming through the door every day. Each event needs to be analysed across the many different domains of a modern network and their various different symptoms. Is it cancer, heart attack, auto-immune conditions, pandemic or just a case of some innocuous sniffles behind each patient's symptoms?
More importantly, we’re the contact tracers. Are symptoms being caused by a combination of these conditions in each patient or infections from other patients? This is representative of the many cross-domain effects that impact our networks every day, many in complex situations that our tools may have never seen before.
To perform diagnosis of every event at the scale and speed presented by our complex, layered, modern networks requires exceptional skill. But it also needs exceptional awareness of the relationships, the network effect, that underpins these networks. To perform Root-Cause Analysis (RCA) or Customer / Service Impact Analysis (SIA) requires a knowledge of how the network connects. Not just at a physical level, but across the many logical connections that our networks consist of - Virtual Private Networks (VPNs), Virtual Circuits (VCs), application end-points, QoS / port / protocol groupings, etc.
Summary
As described in our previous article, SunVizion's pre-integrated tools such as Network Inventory, Service Order Management (SOM) and Network Configuration Management provide our customers with the ability to trace through the various layers of network relationships. These tools manage connections across domains, physical and logical networks as well as establishing customer / service associations. The question that remains though, is whether your insurance policy (OSS/BSS) provides you with adequate coverage?