The Downside of "Reliability"

It might seem strange to argue that "reliability" is a potential negative in the telecommunications business, but in some ways "high reliability" can be a problem, rather than an advantage. 

The simple problem is that "high reliability" costs money. And in today's market, it isn't clear that consumers and business buyers are willing to pay a service provider for traditional reliability standards. 

A presentation by Sam Johnston, Equinix director of cloud IT services, nicely captures many of those challenges are faced by the global telecom business.

If you extrapolate from the shift in the way computing is accomplished, and the way software strategies evolve, you will see clearly why the traditional "telco" value-price proposition increasingly is being disrupted.

Consider "over the top" apps such as Skype, cloud-based productivity apps such as Google Drive, Google Docs, messaging apps such as WhatsApp. 

When creating apps and services of any type, designers have choices to make. They can assume or create "reliable software" designed to run over "unreliable hardware." That largely is the position occupied by major consumer and enterprise-facing apps. 

That also is the way major data centers are designed, using racks of cheaper servers instead of a smaller number of reliable servers. 

Conversely, at launch, the strategy might be "unreliable software" running over "unreliable hardware" or "reliable hardware." 

That approach might be taken by a newly-launched over the top app, aimed at the bottom of the functionality scale, or with an initial set of features that grows over time to provide more robustness.

The approach that arguably does not work so well is "reliable software" running over "reliable hardware." The reason is that this approach costs too much. 

You might argue that approach--reliable software running on reliable hardware--is the traditional telco approach to business, in the monopoly era.

That is why so much attention has been paid to "five nines" of reliability, and stringent testing of new apps. "Our name is on it" is one way of describing the traditional telco concern about app or service robustness. 

Today's software apps typically do not use that approach. They often launch in beta, expecting some things to break, and fix those problems on the fly. 

The point is that this approach necessarily involves a different approach to product development and market launches. "Moving fast" requires a willingness to launch products that are not yet fully formed and completely robust. 

Telcos typically do not create or launch products that way, which means telco apps tend to take longer to develop, and imply more cost. 

And that is the problem. It no longer is the case that high-cost approaches such as "reliable software running on reliable hardware" will work, in a competitive and fast-moving market. 

"Good enough" reliability really is the new requirement. Consumers are used to mobile phones that do not work some places, calls that drop occasionally, operating systems that need to be rebooted, perhaps daily. Users understand how to deal with using apps that do not always work perfectly, but work well, most of the time, and feature very low cost. 

No mobile network ever approaches "five nines" levels of availability. Nor do operating systems, browsers or other apps. 

Instead, the approach is "high value" experiences that "mostly work" but which are not as reliable as traditional telco services.