Telecommunications Operational Data Store Avada Support 2019-06-11T14:35:13+00:00
Solutions delivering competitive advantage and differentiation through faster business insight
1 Tier Telecom Company with Several Hundred Million Users
China Telecom was established in 2002, the main business covers wireline, mobile telecommunication, Internet access service and integrated information service. China Telecom is No.1 in wireline service and the 3rd largest mobile communication service provider in China. Its fixed-line and broadband market share in China is more than 50%, and has the world’s largest fixed telephone network and data communications network, covering the country’s urban and rural areas, accessible around the world.
High Latency and Low Efficiency on Current Hadoop Platform, Workload Occupies Too Much System Resource, Tasks Cannot Even Run Concurrently. When Customer’s Analytical Business Increases Dramatically, the Response Time Cannot Meet SLA.
China Telecom takes Hadoop as core technology, HDFS as storage, using Hive for business analytics and accurate queries, e.g. log analysis, wide table analysis, customer data association analysis, fast phone query, reporting etc. Hive uses MapReduce which is high-latency and low efficiency for processing data. SLA cannot be met when analytical business increases. If no specific configuration on resource control management, the CPU utilization will be 100%, thus other tasks cannot be performed simultaneously.
Search for the Perfect Match
China Telecom is looking for new technology, it can leverage current Hadoop platform (current cluster has 500 nodes) and overcome the drawback of high CPU utilization and long response time. Meanwhile it can meet different business demand including accurate query, reporting and various complex analytical business.
EsgynDB Makes the Match
EsgynDB, the industry’s first all-in-one SQL database for Big Data, can meet various demands from China. Telecom As converged big data platform, it can choose different storage engine for different workload type. ORC format has better performance for complicated analytical business, Hbase can serve better for accurate query like telephone number query instead of ORC. Above 2 distinct workloads can run on unified platform. As the mature 4th generation SQL engine within EsgynDB, by using a couple of patented technologies queries can achieve the best DoP and execution plan, so that resource management could be optimized. Due to MPP system architecture, its linear scalability can meet the system expansion requirement easily when business increases in future. To top it all, Esgyn’s team of database experts and the consulting and professional services they provided removed risks from the solution implementation.
Multiple table join performance:12X
Wide table analytics performance: 2.5X
CPU utilization: down 5X
EsgynDB Delivered Value
Dramatically reduce query response time
Down CUP utilization rate, more queries can be executed simultaneously
EsgynDB Hbase support better performance for accurate query
EsgynDB ORC support better performance for analytical workload and data aggregation operation.
Access original data via external Hive table
As EsgynDB is ideal for ODS workload on EDW (this has been proven by HP), China Telecom is considering moving hundred million operational user data to EsgynDB and running ODS reporting workload together with current ones on the same platform. This will simplify system architecture and save IT cost.
“EsgynDB can handles different business demand efficiently, it achieves better performance of accurate query and reporting, as result customer’s satisfaction is improved. Benefit by ORC column store, analytical workload also has good performance.”