STACKIT as redundant infrastructure Disaster Recovery in the European Cloud

Grafik: Mann steht vor einem visuellen Bücherregal mit Datenbanken

Home / Our services / Infrastructure Projects / Digital Sovereignty / Disaster Recovery in the European Cloud: STACKIT as redundant infrastructure

inovex’s Managed Services Team (iMS) has developed a STACKIT-based disaster recovery strategy to ensure the availability of the company’s internal IT services in emergency situations. The use of the European cloud alternative not only increases the systems’ resilience, it also strengthens inovex’s technological sovereignty.

Background: The inovex Cloud Services (iCS)

inovex’s internal IT infrastructure forms the backbone for its employees’ daily work. For this purpose, an internal team operates the inovex Cloud Services (iCS) – a private cloud based on OpenStack which runs on dedicated hardware in a Frankfurt data centre. Between sixty and seventy internal services are hosted here, and the setup gives inovex the maximum possible degree of independence. As with every on-premises or private cloud solution, however, there is one major risk: if the infrastructure fails, whether physically or logically, the services will shut down. To ensure its availability, the data was previously mirrored in other data centres as part of a multistep backup concept, but the setup lacked the computing capacity for rapid restoration. Until recently, the measures specified in inovex’s Restore Time Objectives (RTO) had been sufficient. In order to considerably exceed these targets, however, both the concept and its technological basis needed to be radically revised.

The challenge: Adding geo-redundancy to increase reliability

The iCS and iMS Teams’ rationale was clear: For the unlikely event that the iCS were unavailable or experiencing technical difficulties, a reliable backup and restore (disaster recovery) plan was required – one which did not involve inovex investing in their own high-capacity, continuously available standby infrastructure.

The search was launched for a cloud provider which fulfilled the following criteria:

Storage

Sufficient, flexibly available computing and storage capacity to handle the entire inovex workload in case of a disaster – in addition to regular backups and georedundant services.

Compatibility

As the iCS is based on OpenStack, it was highly desirable for the target system also to be OpenStack-based – or at least built on a similar, compatible concept. This would simplify migrations and restorations from backups.

Data sovereignty

As a German company, inovex sets great store by data protection and GDPR compliance.

Managed services

In order to minimize the internal team’s administrative workload, the provider needed to offer Managed Kubernetes (K8s).

Deciding in favour of STACKIT

The company decided to go with its technology partner STACKIT. As well as being an existing partner, STACKIT impressed the teams primarily with its technological similarity (OpenStack basis), its location in Germany/Europe, and the services it provides.

During the evaluation phase, inovex tested a variety of STACKIT’s services, including Block Storage, Object Storage, and the company’s managed Kubernetes offering. The rapid development of STACKIT’s portfolio made a particularly positive impression. Although, for example, the CLI (command-line interface) was still partly in the beta stage, it was developed into a mature tool in less than a year. STACKIT’s support also proved rapid and expert.

The solution: Integrating the STACKIT Kubernetes Engine (SKE)

The iMS team created a complete Kubernetes infrastructure based on the STACKIT Kubernetes Engine (SKE). The aim was to create an environment with the same features as the iCS’s internal clusters.

One significant benefit of the SKE was the considerably lower effort involved compared to self-managing a Kubernetes cluster.

The result: Initial successes in live operations

Although the project has not yet been completed, the solution has already proven itself in practice.

Proof of concept

Backups of the internal systems have been restored immediately and without a hitch in the STACKIT environment.

Emergency testing

During a temporary interruption in the iCS’s performance, the GitLab Runners (essential for CI/CD pipelines) were quickly rerouted to the STACKIT infrastructure. This emergency solution worked seamlessly and confirmed that the disaster recovery strategy was ready for operation.

VPN integration

A future VPN solution is already running in a semi-productive environment on the new platform and integrates it seamlessly into the company network.

Future plans

Going forward, STACKIT will be a fixed component of the inovex infrastructure, ready to launch services promptly and independently if the iCS develops problems. Services that are particularly critical will run redundantly on both infrastructures to ensure an even greater level of availability and independence. The integration also enables inovex to flexibly access STACKIT’s “as a service” features without having to develop these themselves.