Auto scaling

from Wikipedia, the free encyclopedia

Autoscaling (also auto scaling ) is a method in the cloud computing environment that ensures that the number of servers within a server farm is automatically scaled. This means that as the load increases, the work is done by a larger number of servers. If the load drops, redundant servers are automatically shut down.

The term was strongly coined by Amazon.com and is often used in connection with the Amazon Elastic Compute Cloud ( EC2 ) product.

Areas of application

Autoscaling is used in cloud computing environments in which the server load is shared between different servers and where there are strong fluctuations in the load on the servers. With these websites it often happens that few visitors visit the pages during the day and a large number of visitors in the evening. During the day the load is therefore low and too many servers are available. In the evening, however, the number of servers is insufficient and this leads to performance problems. With autoscaling, the number of servers is automatically reduced during the day when the number of visitors is low, in order to use as many servers as necessary in the evening when the number of visitors is high. The German Telekom offers this service, for example, as part of the Open TelekomCloud , a public IaaS product on.

advantages

By reducing the number of servers used when the server load is low, costs can be reduced. At the same time, performance problems are avoided by using a higher number of servers at peak times. Furthermore, DDoS attacks are technically almost impossible - but financial risks remain.

disadvantage

Setting up such an environment for the first time is usually more complicated. The use of auto scaling harbors immense financial risks, especially if the load can be traced back to DDos attacks. The use of load balancers minimizes the risk, but it is still possible for attackers to generate resources at the application level.

References

  1. ↑ Michael Armbrust et al .: Above the Clouds . A Berkeley View of Cloud Computing. February 2009 ( PDF - Technical Report).

Web links