Top Global Telco Answers Call for Speed at Scale

技术

平台即服务 (PaaS) - 数据管理平台

适用行业

电信

适用功能

商业运营

用例

供应链可见性(SCV)
需求计划与预测

服务

系统集成

挑战

The company, one of the largest telecom companies in the world, wanted a global platform that could scale new services for hundreds of millions of users and countless partners. The first challenge was a new pay-as-you-go global Wi-Fi service that was outstripping capacity. The company had built the service assuming it could sign up 5 million users in five years; it signed up that many in just six months. The runaway success meant the company needed 40 times that capacity—enough for 200 million users—to handle future demand, but scaling its existing infrastructure was far too costly. The company previously stored transactional and billing data in Oracle Real Application Clusters (RAC), but Oracle wanted $5 million for the kind of expansion the company needed. Worse, Oracle RAC hit a wall at around 150 business transactions per second—way below the company’s requirement of 2,000 per second.

关于客户

The customer is a $60 billion global telecom provider. As one of the largest telecom companies in the world, the company wanted a global platform that could scale new services for hundreds of millions of users and countless partners. The company had built a new pay-as-you-go global Wi-Fi service assuming it could sign up 5 million users in five years; however, it signed up that many in just six months. The runaway success of the service meant the company needed 40 times that capacity—enough for 200 million users—to handle future demand.

解决方案

The company rejected Oracle® Coherence after discovering it wouldn’t scale beyond approximately 500GB and decided against Memcached due to poor scaling, frequent crashes and lack of enterprise-grade features such as high availability and fast restart. The company eventually chose Terracotta BigMemory Max for its ability to deliver extremely low, predictable latency at scale. BigMemory’s highly available striped architecture, fast restartable in-memory data stores and reasonable cost were important factors in the company’s decision. The Terracotta BigMemory Max deployment easily achieved the company’s goal, demonstrating it could handle 200 million users—a 40x boost in capacity—while delivering more than 2,000 business transactions per second, a 13x improvement over Oracle RAC.

运营影响

The company’s engineers were initially skeptical about BigMemory because it seemed like yet another layer to manage, but easy deployment and operation have impressed them.
What they love most about BigMemory is that even when their database of record is down, applications continue to run at peak performance.
After achieving success in the Wi-Fi service project, the company is looking to use BigMemory to speed access to all of its billing and account systems for customers and B2B partners worldwide.

数量效益

Scaled capacity 40x, from 5 million to 200 million users
Boosted transaction throughput 13x versus in-memory Oracle® RAC—from 150 transactions per second to more than 2,000
Achieved performance and capacity goals while spending only 10 percent of the $5 million that Oracle® would have charged