Jerry Bi, Senior Product Manager of Huawei Big Data Analytics Platform

Dr. JC Dorng, VP of Huawei Big Data Business Consulting

(tmforum 2016/17  Perspectives)

Big data is key for industry digital revolution

Big data has established itself as one of the key competitive differentiators in the emerging digital economy. Companies that have embraced big data and put insights into action are revolutionizing entire industries – just look at Uber, AirBnB and Facebook. But the benefits of big data aren’t limited to emerging companies. Many tried-andtrue enterprises such as GE, UPS and Bank of America are also using big data to stay ahead of the competition.

In recent years, telco operators increasingly have embraced big data and analytics to gain more insights into customer behavior and volatile market conditions, to make data-driven business decisions faster and more effectively than the competition, and to explore digital services, assets monetization and ecosystem development.

Challenges for telco big data

Many operators have invested in the Hadoop-based big data platform and formed the ‘DWH (data warehouse) + Hadoop’ coexisting big data platform. When collaborating with global carriers for big data deployment and operation, challenges are found common in carriers’ big data deployment and daily uses:

• Departments usually have their own systems and datasets that have created siloed fragmented environments. Some operators also have siloed departmental Hadoop clusters, leading to limited, department-only datasets and inefficient/low system utilization.
• Data complexity makes it difficult for operators to govern, integrate and process the volume and variety of data in a timely manner (velocity) in order to generate business value.
• Lack of skills or qualified data scientists to explore and discover customer insight or market opportunities with the increasing demands of data, modeling, algorithms, services or applications for digital operation.
• Data operation pains to support and integrate tools or functions from multiple vendors, which result in high development and maintenance costs.
• Digital services demand agile business operation and products/ services to meet customers’ demands and rich digital experience plus ecosystem development to build and offer innovative services.

Data-driven transformation

Telco operators are increasingly adopting ‘data-driven’ and ‘digital transformation’ as a company strategy. Big data plays a key role in implementing this and must be intelligent, agile, open and on-demand.

• Intelligent

  i  Unify and govern B/O/M domain data into a convergent data model to manage and integrate ambiguous, inaccurate, incomplete datasets.

  ii  Develop deep customer insight with automatic customer knowledge discovery, customer profiling, tagging and micro-customer segmentation.

  iii  Fast and efficient business analysis and exploration with keyword-based automatic data search and preparation, and self-analysis and modeling.

• Agile

    To empower agile business innovation and facilitate collaboration, big data must enable business and marketing teams to do business analytics or application development easily and quickly. In addition, big data must be open to external partners to allow easy integration with outside capabilities such as industry best practice techniques

• Open and On-demand

    As everything moves into the cloud and can be presented as a service, big data provides open analytics on-demand.

Huawei Fusioninsight-Universe Big Data Analytics Platform

Huawei’s telco-focused FusionInsight-Universe Big Data Platform provides four key features:

• Open and Integrated Big Data Analytics Platform
• Convergent Data Governance
• Agile Development and Self-Analytics/ Self-Service
• Big Data as a Service

Open and integrated big data analytics environment

The platform offers an open and integrated environment built on open source technologies and best business practices including:

•   Hadoop and Spark for batch and real time data processing;
•   Strom and Spark streaming engine for real-time event processing;
•   data analytics of auto-modeling business and prediction; and
•   visualization with the telco’s own and top industry tools

Convergent data governance

Trustable data governance is the basis of all discoveries and decisions. However, there is no data governance across DWH, massively parallel processing databases (MPP DB) and Hadoop. Huawei FusionInsight-Universe Big Data Platform provides a convergent data governance with following capabilities:

• unified metadata management to manage and define all the data and data lifecycle across DWH, MPP DB and Hadoop; and
• a convergent data model of business and operational support systems, and management based on TM Forum’s Information Framework (SID) to present dimensions of customer, market, partner, service, resource, event and enterprise.

Agile development with self-analytics and self-service

Huawei FusionInsight-Universe is designed to support different types of big data users such as business people, data scientists, developers or external partners.

• Business people and analysts can do self-service data preparation, automatic modeling and prediction, and visual-based data discovery without knowing statistics or technology.
• Data scientists are usually proficient in mathematics and algorithms, which can be combined with their business knowledge to create predictive analytical models. Huawei big data platform provide predictive analytics modeling that scientists need, without coding or knowing where data is stored that enable data scientists to focus on their work to improve efficiency.
• Huawei Big Data Platform provide the agile development environment that includes capabilities of data ingestion, data model design, data analytics processing and orchestration, etc., for developer or partner to build application. It also provide Agile Scrum and Project Management that partner and developer can do online development, testing, and rapid release of the services and applications to the product environment.

Big data as a service

Big data as a service is the next big thing when operators move into cloud. Huawei Big Data Platform can be deployed on private/public clouds to support multitenant deployment to offer big data platform as a service, data as a service and big data application as a service.

Big data platform as a service

Huawei Big Data Platform decouples and exposes layered capabilities as a service, which includes following:

• big data storage and computing;
• data ingestion and integration;
• analytics capabilities (statistical analysis, data mining and prediction); and
• visualization.

Data as a service

For business people, it would be convenient and efficient to get desired data, without learning or knowing specific big data technologies. Data as a service enable users to query data with two approaches. One is to use a data assets map, which lets the user query data asset metadata definitions and corresponding data after obtaining permission. Another is based on domain-specific knowledge data service. According to business domains and requirements, the big data platform provides the pre-defined data service – for example, when a customer calls in, the customer service representative is able to get customer profiles, statistics and recommendations to help make a decision.

Big data application as a service

The big data application normally is the end-to-end, standalone application, which includes data processing, analytics, visualization, the user interface panel, interactive operations, etc. Big data application as a service offers a software-as-a-service mode to support multiple tenants. The end users can directly use the big data application service without hardware preparation and installation.