Long-Term Storage (LZS)
Currently, a highly scalable object storage is under construction at the RRZ under the project name “Langzeitspeicher” (LZS), which, as a supplement to the file service and the central storage infrastructure (ZSI), will cover the constantly growing demand for storage space for data from science and research.
The LZS is suitable for, among others, the following purposes:
- Backend for file services
- Backend for archiving solutions
- Backend for research data management or CMS
- Backend for the delivery of multimedia files
- Storage of measurement data or results from simulation calculations
- Storage of larger collections of photos, films or digitized books
- On-premise cloud storage for scientific collaborations as a secure and affordable alternative to Amazon Simple Storage Service (S3)
The target group of the LZS are systems that speak S3 natively and have to store larger amounts of data at a low rate of data change. With UHHDisk, the RRZ offers a suitable platform for storing files, e.g. for home drives or project directories.
Due to the system architecture and configuration, the LZS is not suitable for use in structured data (backend for databases) or virtualization environments. For this the RRZ offers a suitable service with the central storage infrastructure (ZSI).
If you have any questions about the LZS, please contact the Serviceline of he RRZ.
Access and Protocols
The primary protocol for accessing the object stores is Amazon AWS S3. The manufacturer of the software used by the LZS guarantees full compatibility with the S3 standard. This allows access via Amazon AWS S3 tools and APIs.
Amazon AWS S3 is currently supported by many cloud-connected repository systems and measurement facilities. Clients are available for the current operating systems, such as the AWS CLI or Mountainduck.
In addition, there are libraries for all common programming languages and development environments.
The use of the LZS via special gateways is also possible via the UHHDisk and thus via the SMB/CIFS protocol.
Availability and Data Security
The long-term storage is designed for the storage of data with high availability requirements.
This applies to the fail-safety against defects and malfunctions as well as the safety against data loss due to technical errors and errors during use (accidental deletion, modification of data). Availability and data security are ensured by the following features:
- all data remains in UHH data centres (On-Premise Cloud)
- support of the LZS by qualified personnel as well as through support contracts with the participating system manufacturers
- construction of all components in access-protected areas with high-availability infrastructure (air conditioning, power supply and network)
- redundant storage of all objects in the primary system at three UHH locations using Erasure Coding
- clusters of load balancers at all locations allow distributed access with high availability
- full availability in case of failure or maintenance of any components
- the system remains online even in the event of a complete site failure
- asynchronous replication of data to a second object memory at another UHH location with storage on LTO tapes (off-site and offline backup of data)
- Buckets versioning. Deleted data remains in the primary object store and in replication as a result of versioning until rule-based deletion as previous versions
- SSL encryption during transport
- optional: WORM function for buckets for archivingoptional: encryption on the client
- optional: encryption on the client