The Telco Data Cloud: Hybrid Cloud, Governance and Ecosystems

The telecom industry is powering forward as the world stabilizes after the shock of the pandemic, and 5G networks roll out everywhere. Connectivity has never been more important for people and for businesses, and digital transformation in every industry is accelerating. As service providers execute on these programs, thoughts are turning again to the role of data, and data architectures in an increasingly virtualized environment.

The old mantra is still the same: volume, velocity and variety are the primary vectors in designing an enterprise telco data architecture. In a virtualized 5G world, ‘volume’ and ‘velocity’ are even more pronounced.

When it comes to volume, 5G networks have more physical elements due to increased density supporting more persistent connectivity, and virtualized network functions dramatically increase the complexity of the environment. 5G is all about IoT, and the number of edge devices further compounds the amount of data being generated.

On the demand side, AI and automation are insatiable: more and more data is required, often data sets that would previously have been seen as ephemeral, and not valuable for much more than real-time monitoring. Small network events that might have been useful in real-time diagnosis of network issues are now predictors or probabilistic indicators that can trigger actuation in automated systems. This means that more of the data that would have been discarded needs to be collected and made available for training and analysis.

Instant gratification architecture

Meanwhile, real-time requirements are mounting as networks are expected to recognize important events within the Internet of Things on behalf of enterprise clients. Use cases related to security, fraud and safety are springing up in different industries, demanding streaming analytics on the fly. How different industries interpret ‘real-time’ can vary, but it is increasingly the case that user expectations are now measured in seconds and minutes rather than hours and days: this is ‘instant gratification’ architecture.

With higher volumes and higher velocity than ever before, the underlying infrastructure comes into focus. All service providers are virtualizing IT operations, and where possible are seeking to exploit the public cloud. On the face of it, the flexibility and agility of public cloud infrastructure, combined with the cost control and shift from capital to operational expenditures are attractive features, and for certain workloads, quite lucrative. Avoiding IT and management costs, along with depreciation, maintenance, upgrade and other aspects is very desirable.

For much of the last decade, telcos outsourced their IT in order to achieve stability in costs, but as those costs climbed, public cloud appeared as a more acceptable option. For enterprise data workloads, however, there are other considerations for those considering cloud. First, while the public cloud allows scaling up and down as required, the experience has been that there is very little scaling down. Early adopters are finding that as their workloads mature on the public cloud, the costs begin to climb.

Data protection

Telcos considering the cloud also face unique set of challenges around security, privacy and governance. The rapid spread of data protection regimes led by Europe, but quickly followed in California, Brazil, India, South Africa and other places has meant that data jurisdiction is important, and moving personally identifiable information outside the jurisdiction may not be compliant. Some service providers have moved back from public cloud to private / on-prem options having spent 12-18 months working on public architectures, only to be stymied by internal security and privacy compliance. The message there is to get security and privacy teams engaged early in the journey.

Many service providers now recognize that in a telco data architecture, decisions need to be made early in the data flow as to how and where the data should be processed and stored. Should it be a cheap cloud for high volume / low value data? Should it be a flexible feature-rich cloud for low volume / high value data? Should it be a private cloud for PII? Should it be a secure cloud for specific client data – like military or medical data? Can the data be processed at the edge and not persisted at all, or aggregated at the edge in sensible ways?

Hybrid cloud

That’s why leading telcos have opted for a hybrid data cloud architecture to provide maximum enterprise flexibility and agility when designing next generation data architectures. Furthermore, integrated data governance is something that should be carefully considered: with a hybrid cloud environment, and the inevitability of a certain amount of data redundancy, consistent data lineage, catalogue and stewardship is critical. Transparency, auditability and unified data policy design and enforcement allow data operations with confidence, as regulations and data transfer laws continue to evolve.

Ecosystems

A final thing to consider is the rising importance of ecosystems. Managing an integrated next generation telco data environment needs to enable core operations in a compliant and effective manner. Applications that depend on that platform are both internal and external: IoT ecosystems, advertising ecosystems, smart cities and government / security demands have multiple requirements for data access, all of which need to be facilitated under consistent compliance posture. Similarly internal ‘customers’ in marketing, finance, networks, retail, enterprise and other domains will have increasing demands for access for data scientists, business analysts, and executives. It’s important to make sure that the API layer and other interfaces are open, extensible and secure. The opportunity for the telecommunications industry is growing, both in enterprise and consumer businesses, as virtualization and 5G catalyze digital transformation. Data is at the core of that transformation – and the successful next generation enterprise data cloud will be hybrid cloud, comprehensively governed, and open for business.