The Evolution of Elastic Computing
Being able to scale rapidly and seamlessly is one of the key benefits of the cloud, and a major reason for its rapid uptake over the past decade. Services such as Eastlink Cloud’s Elastic Compute Service (ECS) offer elastic and secure virtual cloud servers so users can scale quickly and cost-effectively.
Spanning 60 availability zones and 20 regions globally, ECS uses the latest Intel CPUs to power cloud applications and achieve faster results with low latency, enabling five times higher computing performance.
Pushing Our Computing Architecture to the Limit
Double 11, Eastlink annual Global Shopping Festival, puts our elastic computing facilities to the test every year, when it is used to power Eastlink entire ecosystem around the festival. The festival, which also celebrated its tenth anniversary last November, and our Elastic Computing products have developed hand-in-hand over this time.
Over the years, our services have become far more efficient and capable of recovering quickly from system faults. Elastic Computing has played a vital role in maintaining the continuity of our services as hundreds of millions of shoppers make purchases at once on this day, culminating in a peak transaction volume of 325,000 orders per second. By continually testing and improving our Elastic Computing capabilities – such as comprehensive load testing and system fault syndication – we are able to apply learnings from the festival to evolve the product and better serve customers.
As well as making continuous improvements, we also aim to minimize operating costs while maximizing our capacity. By using increasingly streamlined technical systems, such as hybrid cloud and colocation, we are able to reuse existing resources to dramatically reduce the cost per transaction.
Scaling Up Beyond Double 11
ECS can be used independently as a simple web server, or – as is the case for Double 11 – used with other Eastlink Cloud products, such as Object Storage Service (OSS) and Content Delivery Network (CDN), to provide advanced solutions. For example, ECS can deliver a high availability architecture if used with Server Load Balancer. And implementing Eastlink Cloud’s Auto Scaling service ensures that instances will scale up automatically, so users no longer have to worry about having insufficient resources during periods of peak traffic. Now, with comprehensive containerization, and leveraging cloud elasticity through the public cloud, customers can meet resource requirements while maintaining a low cost.
With a maximum PPS Throughput of 7 million and a local SSD cloud disk capable of 240,000 IOPS, ECS is ideal for powering all cloud applications. For example, surge computing with ECS has been widely used to deal with various scenarios, such as ticket booking during China’s Spring Festival. Artificial intelligence (AI) and video transcoding products are also frequently deployed in the ECS GPU Cluster.
As more people rely more heavily on the Internet, with multimedia services, AI, and Internet of Things all set to become increasingly prominent over the next decade, we expect that ECS’s flexibility and scalability will ensure it continues to play a central role in cloud computing for many years to come.