Open Source Helps China Unicom to Lead Services Expansion

Big Data

China Unicom wants to improve its services to 320 million subscribers by means of using open source software / Photo by: Bjoertvedt via Wikimedia Commons


Chinese telco China Unicom is leading the way to expand services and improve telecommunications performance for its more than 320 million subscribers with the help of open source software.

The fourth largest telecommunications provider in the world has gone under for a wide rebuild of its internal software stack, founded largely around open source, TechRepublic reports. It says China Unicom's growing business in 4G and emerging 5G networks was already placing a hefty load on the legacy network data processing, which was based on IBM midrange computers, Oracle database, and EMC storage systems.

The company's telco cloud seeks to integrate the cloud computing model into telecommunications infrastructure through the development of software that can operate on commercial off-the-shelf (COST) hardware, said China Unicom's big data engineer Zhang Ce.

According to TechRepublic, a micro-service framework lessens the needed time in developing and deploying applications made up of independent, autonomous, and modular pieces of code. It also says that instead of building software as something monolithic, the application can be devised as distributed applications based on software components that can function across the cloud.

"The architecture of our incumbent computing platform was too complicated and didn't let us effectively use resources," Ce said, adding that open source was China Unicom's path forward.

The big data engineer further said the key to the telco's project success was the emergence of new frameworks on big data with the likes of Apache Kafka, Spark, Hive, and Alluxio. These developments allowed China Unicom to re-imagine its software stack to support batch, as well as stream processing business requirements through the use of relatively similar open source-based models developed by Kubernetes.

The two workloads are operated in Spark clusters and both store data in Alluxio for quick retrieval and access—be it in HDFS, object store, or even disk.

Alluxio serves as a virtual pool for ingesting and accessing data from disparate storage devices throughout the network. It has a unified namespace feature that allows the mounting of HDFS paths of varying clusters, providing users the performance benefit of data locality.

Ce said moving data, even large jobs, is nearly eight times faster when done under the new framework. He added that the new open source model can support seven different lines of business services at China Unicom.