A trend has been building in the telco sector over recent years – modernising to cloud. While some communication service providers (CSPs) have gone as far as providing their own cloud services, they all have relationships with the hyperscalers.
In fact, our recent research shows that just a quarter of telco companies rely solely on on-premise data centres to store data.
This shift started with a realisation from CSPs that they need to move operations to the public cloud – whether to optimise their B2B, network or IT operations. Typically, they chose one of the hyperscalers and announced a big-money deal.
But then the process began to stall once data volumes became so enormous that the cloud bill was much higher than expected. Concerns over compliance with data protection regulations, security and portability to other clouds soon followed.
For cloud to be worthwhile, CSPs must be able to operate cost effectively and be able to drive value from their data, while minimising regulatory risk and vendor lock-in. There’s only one effective way to do that: with the help of a modern, hybrid data architecture, built around a unified data framework.
Sometimes the siren song of IT cloud modernisation is so strong it drowns out legitimate concerns over cost, complexity and compliance. Take cost. Telcos manage huge volumes of technical data, pulling sometimes as much as three petabytes per day from their networks to check everything’s functioning as intended.
That’s a huge amount of data – to put it into context, a single petabyte is equivalent to more than 500 billion pages of nothing but text. Ingesting, processing and storing this data in the public cloud can incur enormous cost, and many organisations used to low-cost, on-prem hardware don’t have the culture or mindset to scale down, which requires a change in both thinking and practice.
There are also architectural challenges that may drive up spend. Some proponents of the public cloud claim “it’s just like the datacentre except you rent it”. This isn’t quite true. In security and networking there are different designs and architectures in the public cloud. The handshake between the telecom network (WAN) and datacentre may be well established in telecom operations, but where should it happen between the WAN and the public cloud? What are the rollback procedures? Answering these questions can require architectures that demand more units of capacity in different areas than would be expected in a pure data-centre operation.
Costs are further driven up by the need to rearchitect applications to specialised cloud-native platforms – this is also where complexity comes in. Often forgotten is the fact that all three major providers run architecturally different platforms. Even open source-based systems like Kubernetes vary between Microsoft, AWS and Google, as well as on-premises environments. This complexity drives up the cost of migration – both from on-premises to public cloud, and between clouds.
That presents a stark choice to CSPs coming to the end of a contract. Either stick with the incumbent – which impacts agility – or move to a different provider and spend huge amounts of time, money and resource rewriting applications.
Cost is not the only headache. Also consider the huge volumes of customer data telcos manage, including payment details, addresses and other personally identifiable information (PII). CSPs must be confident they’re taking the appropriate steps to stay compliant as PII is migrated to third-party cloud datacentres.
They may be confident in their primary hyperscale providers, but what about disaster recovery? Or administrators located in a different jurisdiction? Regulators don’t help by sometimes being slow to update policies and/or publishing contradictory decisions.
The compliance landscape is moving so fast that telcos need an architecture that’s fit for purpose now, and future proof to mitigate whatever risks may come down the line. There’s already a Schrems challenge to the new EU-US agreement on data flows in the pipeline, for example.
These costs, compliance headaches and complexity concerns are why an increasing number of CSPs are partnering with third-party providers capable of delivering a hybrid data platform. No matter how attractive the benefits of public cloud, there will always be data that needs to reside on-premises, and the factors influencing such determinations will change over time.
In fact, two thirds (66%) of telcos have adopted a hybrid cloud approach. That makes it critical for CSPs to be able to access, move, use and extract value from that data seamlessly.
This is the promise of a hybrid data architecture. Having the capacity to execute workloads on-prem or on various public clouds provides the operator with the flexibility to decide where to run what parts of the job.
The same analytics could help telcos calculate where data should be stored – whether on public cloud or on-prem – via workload analytics, to generate maximum value for the organisation. Full governance and observability allows consistent enterprise data policies to be applied irrespective of the underlying infrastructure.
These security and governance guardrails manage compliance risk, applying consistent policies to data, wherever it resides. And a unified data platform means it operates as a single logical platform across all public and private clouds and the edge, meaning data and workloads can move freely without the need to rearchitect and retrain staff.
With vast swathes of data being generated on a daily – or even hourly – basis, CSPs know they need to modernise their IT systems if they’re to reap the benefits of these large, complex data sets. The public cloud is a great choice for many use cases, but not all. And it can add cost, complexity and compliance challenges that often fly in under the radar.
By running their data and workloads from a unified data platform, telcos can manage these headaches today, and set themselves up for success tomorrow.
Subscribe to our Editor's weekly newsletter
Share
