Telcos can become a highly data-driven enterprise by leveraging the Digital Integration Hub (DIH) Architecture built on GridGain’s in-memory computing platform. In this blog post, I will discuss how the DIH architecture can help telcos develop better customer insights, generate new revenue streams and be ready to ride the 5G wave. This easy-to-adopt, no rip-replace architecture can meet the needs of telcos through elegant design and implementation patterns.
Digital Integration Hub Architecture
We’ve all heard that data is the new gold or oil or <insert your favorite asset>. Despite this, a recent HBR survey shows that nearly 70% of companies haven’t created a data-driven organization. This is likely because of the challenges associated in integrating data distributed across multiple legacy and modern information storage systems, in different formats and access control requirements. To solve this problem, more and more companies are adopting the DIH architecture – also known as a Smart Data Hub or Smart Operational Datastore.
Telco DIH Use Cases
Customer 360’ View
Bundled services are not a new concept in the telco world. With consolidation in the market, providers deliver a variety of media connectivity and delivery services to consumers. The services include cellular services (data, calling, pre-paid options) and media delivered over broadband or otherwise (cable television, internet, phone lines, etc). Understanding a customer’s complete profile, service subscriptions, and consumption patterns are important for optimal customer service.
The challenge that telcos run into is that they have many legacy systems, some home grown but mostly packaged, monolithic COTS applications that allow for a very limited ability to access and manipulate data.
GridGain’s DIH architecture is well suited to connect such legacy systems by leveraging GridGain’s ability to integrate with a variety of data stores, change data capture technologies and use of open APIs and standards based communication protocols.
New Revenue Streams
By being able to get that 360` view of its customer, organizations can easily leverage customer behavior. Telcos can anticipate demand patterns for individual customers and identify services that they might be likely to be interested in. This ability can create a new revenue generation opportunity. A simple example is servicing a customer on an “unhinged” (where the customer does not have a contract with a cell phone carrier and can switch any time) pay-as-you-go data plan. Carriers can offer promotional data refills and customized bundles on the user profile. A more complex example is sending a customer a coupon to a nearby restaurant, based on the time of the day and their location. By applying a bit of additional intelligence, a carrier could actually predict the customer’s food preferences and send them a coupon that they are more likely to use.
While these use-cases are not new, a DIH architecture makes the use cases more easily achievable. The architecture allows telcos to access and aggregate customer data quickly and apply some real-time intelligence on it.
Quality of Service for 5G Networks
As we move from 4G & 4GLTE to 5G, telcos need better management of stateful/stateless subscriber data. They need this data in a highly reliable and available way with varying access speeds (from microseconds to many seconds, depending on the telco functional domain). This data management ability will enable telcos to support the quality of service requirements across the major 5G service classes (eMBB, mMTC & uRLLC). The shared unstructured data store (UDSF) with services to access this data from a variety of network functions (NFs) is a key tenet of the 5G architecture.
DIH architectures can help by providing a low-latency aggregation layer for data from multiple telco systems.
The Telco Data Problem Statement
The need for disparate data at high speeds and scale is clear from the above three areas. While the specific data problem in each of these three use cases is somewhat different, there is commonality around availability, durability, and reliability of the data layer. Across these three scenarios we see the following data characteristics:
- High availability, reliability, and ultra-low latency
- Variety of data types, structures, and profiles (static, transient, frequently accessed, frequently updated, real-time changes)
- Real-time intelligent decisioning and other complex computations
- ACID support
- Security and access controls
A data platform that supports these requirements should be able to:
- Unidirectionally integrate with different proprietary data stores as sources and pull data from them as an initial load or on a continuous basis
- Bidirectionally integrate with a variety of data stores to read and write to them in real-time, on an ongoing basis
- Provide a variety of interfaces (both SQL and non-SQL based) and microservices to manipulate data in the platform in a secure fashion
- Execute complex compute tasks
- Do all of the above at massive speeds and scale
GridGain’s Digital Integration Hub
An in-memory computing platform lets you load data into your RAM from various sources via a set of simple APIs. The platform can then run a variety of compute operations on that data in the same memory space.
GridGain Digital Integration Hub Reference Architecture
The GridGain DIH architecture built on top of the GridGain In-Memory Computing Platform takes this basic in-memory computing concept further. The DIH architecture provides the following additional enterprise-grade abilities that mission-critical applications need.
High availability, to make sure that critical customer and operational 5G data is always and easily available across the telco network.
Advanced security and role-based access controls, to make sure that only authorized personnel can access information, to protect customer privacy, and to secure both data at rest and in motion.
ACID & ANSI SQL support, to provide critical data management functionality with the benefit of high availability and scale. With the support for ANSI SQL, GridGain allows companies to work with their existing, legacy relational databases with minimal rip-replace of their current applications.
Hybrid analytical and transactional processing (HTAP or HOAP), to prevent large data movement across the network for analysis. Think classic ETL processes between your OLTP & OLAP. This feature enables analytical processing on the same data layer that supports transactional processing.
Support for training and execution of ML/AI algorithms, thereby enabling the execution of complex AI models and facilitating real-time decisioning for deeper customer insights and generating new revenue opportunities.
Furthermore, with its Transactional Persistence feature and ability to horizontally distribute data across commodity hardware, GridGain supports data access at various SLAs and unlimited scale
GridGain Supports Globally Distributed Telco Deployments
Finally, GridGain supports deployment across multiple geographically located data centers or cloud platforms with protection against network segmentation. This offers a level of resiliency, availability and scale that is well suited for the kind of data management use-cases that telcos need.
We have successfully helped Telco customers meet their specific needs by deploying slight variations of this telco reference architecture. If you are dealing with similar use-cases and data management problems, please reach out to us. We will be happy to discuss how GridGain can help you in taking your organization one step closer to becoming a data driven enterprise.
- Blog Post: In-Memory Computing and the 5G Ecosystem
- IMC Summit Talk: How Apache Ignite Powers Real-Time Subscriber Offers for a Leading Telco (presented by Teradata)
- Digital Integration Hub Overview