Growing demand for network data
CSPs are facing challenges on many fronts – rolling out new network technologies such as FTTH (Fiber to the Home) and 5G, improving customer satisfaction, while at the same time trying to reduce operating costs. The one link between these disparate goals is the need for network data. The variety of these data and their volume (often 10’s or 100’s of TB per day), make them valuable to many operational processes, and enable CSPs to:
- build personal, one-to-one interactions with the customer, relying on a deep understanding of her behavior and experience;
- improve service assurance, gaining insight into the quality of service that customers experience, without the need for surveys or complaints;
- Help Engineering make smart network investments through insights into the nature and quality of services, value of customers, bandwidth, and latency appetite.
In these cases and many others, readily available network data would help a lot. The question is: what’s the best way for CSPs to gather, collate and use their network data?
An outdated dilemma: buying a platform or building your own?
To most CSPs this is nothing new; indeed, they’ve been amongst the first adventurers in the deployment of big data. But up until recently they’ve been choosing between two strategies, neither of which fully meets their needs:
- Buying an off-the-shelf, proprietary solution. On the plus side, this option is pre-built with integrated telecom expertise. Unfortunately, data and derived insights are too often confined within the limits of a proprietary environment – tied to a specific data model, which is owned by the vendor and not the CSP. This makes it difficult or impossible to use your data in new ways, which may not have been considered when the platform was designed. It can also be very difficult to find the expertise necessary to manage these platforms, further limiting the potential uses of valuable data.
- Building one’s own big data platform with open-source technologies. This has proven quite effective at capturing massive amounts of data, but it demands significant resources. It’s difficult to evolve; it relies on dedicated expertise, and like the off-the-shelf option, it’s difficult to scale. It ties up valuable resources which you would prefer to allocate on solving business problems!
All in all, both options have proven to be more expensive than initially expected, especially in the long run. They both lack flexibility and fail to exploit the full potential of data across the many dimensions of the organization.
Why cloud-native technologies are taking the lead
CSPs have already embraced cloud-native technologies to support their data transformations, with the first initiatives focusing on the corporate and customer domains. More recently, many have embraced the cloud for their network data initiatives, and for good reasons:
- AI – artificial intelligence is becoming a ubiquitous tool, using data to improve operational processes as well as quality of service and relationships with clients. To implement and operationalize AI, CSPs and their data scientists need a wide choice of advanced tools and techniques, plus access to large datasets and large computational power at an economical cost when needed for training models. Cloud-native data platforms deliver these advanced AI capabilities, and with a much lower price tag than most in-house solutions;
- Value – open source and cloud provide a wide range of other advanced capabilities as well, in a way that’s easy to use and highly cost-effective. Under constant pressure to improve their operations and invest wisely, CSPs can rely on cloud to make sure every penny counts;
- Ownership – many CSPs have recently announced plans to become more “software-oriented” – in a sense more like the big Internet companies – developing software on their own, within their own control. They realize the criticality of their data and associated software, which manages, extracts value from and activates those data. It benefits CSPs greatly to own their code and make it a core asset. Cloud aids massively in the creation of agile software solutions, and in later adaptations and scaling.
- Uniqueness – behind a façade of high normalization, networks and service platforms actually differ substantially from CSP to CSP, due to the histories of the particular organizations, the choices of architecture, and the combination of vendors and technologies. Thus, any off-the-shelf solution requires significant amounts of adaptation to meet the individual needs of a given CSP (due to the innate limitations of its design). Cloud-native data platforms make it possible to collect, process and massage data according to the unique needs of each operator. One size does not fit all when managing complex networks and extracting information out of them!
In a word, cloud offers agility. It makes it possible to experiment, adjust, pivot, personalize and scale with a freedom that’s simply not practical with off-the-shelf or in-house platforms. And with the increasingly central role of data, this freedom helps to enable Agility and DevOps throughout your organization.
Functions expected from the Network Data Platform
Cloud-native data technologies are making the dream of network data democratization come true, while helping CSPs advance many of the challenges they’re facing, including the three mentioned at the outset. How?
- One-to-one interactions – a cloud data platform renders network and service usage data more accessible – breaking down silos while ensuring security and privacy. This makes it possible to share the data necessary to drive the personalized, one-to-one customer interactions CSPs strive for.
- Service assurance automation – a combined edge / cloud data platform collects CPE (Customer Premises Equipment) and network telemetry data efficiently – filtering, aggregating and correlating to obtain real-time insights into the operation of services. CSPs can thus spot potential problems early and identify root causes.
- Smart investments – a cloud data platform collects high volumes of data on network, usage and quality of service (typically from cell towers and transmission equipment), aggregates them and applies advanced analytics and machine learning to anticipate future consumption patterns, identifies revenue opportunities, and prepares optimal investment allocation.
Capgemini as partner for building cloud network data platforms
At Capgemini we have experience deploying network data platforms for our clients, starting with on-premises big data platforms and then migrating them to the cloud, or starting with a cloud-native use case and expanding. We expect to see more CSPs taking advantage of the possibilities of cloud network data platforms to unleash the power of their data, remain in control, extract value and become data masters in combining network and customer data. Contact us to learn more.
Interested in cloud network data platforms?
Read about Capgemini’s Cloud Platform
TelcoInsights is a series of posts about the latest trends and opportunities in the telecommunications industry – powered by a community of global industry experts and thought leaders.
|Yannick Martel, Vice President, Artificial Intelligence & Analytics at Capgemini
Yannick is Group Offer Lead for Data & AI in the Telecom industry. Yannick has more nearly 30 years of experience working in the Telecom and Banking industries, helping organizations implement their digital and data transformation and better serve their customers.