In computer technology, the term elasticity refers to the ability of a system to adapt to different load conditions by automatically making resources available when required and releasing them again after the load peak has subsided. An example of elastic load balancing is a television advertisement which says: “Call now” or “Order online now at www.example.com.”
As soon as such ads appear on the TV screens, the demand for resources in the ordering department increases dramatically: Instead of one call center employee, twenty are now needed to handle the calls coming in. Instead of one web server, an entire server farm now has to process online orders – these load peaks are cushioned by elastic load balancing in the cloud.
Elasticity, also known as scalability of resources or elastic load balancing, is one of the basic functions in the concept of cloud computing. The system architecture of the Elastic Cloud ensures that companies have instant access to virtually unlimited resources when they need them. The computing capacities of the Elastic Cloud Servers adapt elastically to changing conditions.
Elastic load balancing technology avoids an undersupply or oversupply independently and without any user intervention. In the event of a shortage, the IT systems used would be slow and the end customers annoyed. In the event of oversupply, on the other hand, systems would be fast and end customers would be happy, but companies would have to pay for resources that they don’t need. In the Elastic Cloud, however, customers only pay for the Elastic Cloud Server resources if they actually use them.
René Büst, a former analyst from Crisp Research, viewed elastic load balancing as one of the top five cloud trends of 2017: “Rule-based infrastructures automatically and independently ensure that the necessary resources are increased and reduced to suit the current requirements of workloads and applications.”
Three technologies work in harmony when it comes to elasticity in the Open Telekom Cloud: Elastic Cloud Server, Auto Scaling and the Elastic Load Balance service. Elastic Cloud Servers aren’t physical servers, but virtual machines in nine different “flavors”, each optimized for a specific purpose.
The high-performance servers used for the Open Telekom Cloud address high-performance scenarios, for example for simulation calculations. The GPU-optimized servers of the Elastic Cloud are suitable for applications with a lot of graphics and videos. The Disk-intensive flavor is aimed at large data volumes, for example for big data analyses, while the Massive Memory flavor is suitable for in-memory databases such as SAP HANA. Companies can change the characteristics of an Elastic Cloud Server at any time, either manually or based on rules using the Auto Scaling service.
Open Telekom Cloud’s Auto Scaling service scales Elastic Cloud Server resources up and down according to predefined rules. These rules determine, for example, the desired response times of a service or application in the Elastic Cloud. If necessary, Telekom customers can change these rules and adapt the Elastic Cloud to suit their needs. Both the rules and the load states of the IT systems can be monitored at any time for precisely tailored management of Elastic Cloud resources. With Elastic Cloud storage, storage capacity can be added or revoked as required.
Finally, with the Elastic Load Balance network service, the Open Telekom Cloud automatically distributes data traffic over multiple Elastic Cloud Servers. Thanks to the immediate, elastic response to peak loads, the response times of the Elastic Cloud remain fast and the error rate of the systems is reduced. The Elastic Cloud thus adapts to both higher and lower loads. The Elastic Load Balance service supports up to 100,000 connections simultaneously.