Deutsche Telekom embraces OpenStack with Open Telekom Cloud

At the OpenStack forum in Germany Deutsche Telekom's (DT) business customer division announced further enhancements in Open Telekom Cloud, its public cloud offering, launched in March 2016 at Cebit.

Open Telekom Cloud is based on OpenStack and is DT’s own first public cloud offering running on OpenStack, the OpenSource platform. DT chose OpenStack in order to keep further development and integration as simple and "open" as possible.

From mid-July 2016, a container service will be available via Docker, for transferring applications across Clouds. For this purpose, complete clusters of containers can be managed and displaced. Furthermore, a relational database, with extensive features for managing the database, backup and restore as well as a comprehensive monitoring and tools for performance optimisation, will be provided.

Docker is an open-source implementation of lightweight Linux containers. Linux containers have been part of the kernel since 2008, and enable virtualization through an environment that has its own process and network space, instead of creating a fully-fledged virtual machine.

Virtual machines are the cornerstone of cloud computing, where the hardware and software are virtualized away, and we can focus on running as many guest-operating systems as the infrastructure will support. This allows the data centre to trick one server into thinking it's 10, 100 or more virtual machines, all repeating across a telco’s entire server farm.

The sheer scale of the average data centre means that the gains in speed, resilience and resource utilization are huge. At the scale of a Google or an Amazon Web Services, it's what makes modern web services possible.

On the telco side the threat is from web companies. Telcos can't compete at the same speed because their proprietary networks are slow, so companies such as DT are looking at companies like Facebook and Google and realising that they have to innovate in the services layer. The push towards full virtualization through SDN or NFV is due to the need to be more agile.

They want to try new services and ‘fail fast if they don't work – they cannot afford the huge lead times to introduce new services.

If they do move towards more software-oriented architecture, they can fail a bit more easily and spread risk. Rather than having to make huge network investments they can be more agile, since software is easier to write and change than routers and switches.