In March 2023 our new software generation entered a pilot phase. Even though until now it took us longer than we hoped: the long development time was worth it, and we are happy that we were able to create a small miracle. With the Virtual Storage Domains we have arrived at a point where we wanted to be since long ago. A highly modern, portable and scalable storage virtualisation technology running in the user space, having an enormous potential for the future. Most important for the users are the functionality, stability and an easy to use interface, but also the ability to run on the chosen OS and with the chosen applications. To be able to satisfy these expectations, the new version 4.8 combines the user interfaces and solutions already known to our users (Application Management, HA, Backup, DR etc.) with the new Core Engine. And with the new engine the OSL Storage Cluster 4.8 including all storage connectors (FC, SCSI, iSCSI, SAS, RSIO ...) is available for Solaris und Linux at the same time.
The cross-platform capabilities of the version 4.8 make it possible to combine all hosts and storage systems in one cluster. Possible host platforms are Solaris 10/11 (on demand also earlier versions starting with Solaris 8) for x86 and SPARC, and all modern Linux systems. Without additional in-band appliances, with software only, it is possible to keep track of things even in larger environments, making the administration tasks much easier. The image on the left shows a conventional data center stack and the places (marked with a gear) where there is a need to perform administrative actions, e.g. for creating a new storage object and assigning it to an application. The image on the right illustrates that the OSL Storage Cluster becomes a "Single Point of Administration", from storage to application description. It is worth noting that any configuration data are independent from the hardware and exist in a detached from the hardware form. Therefore it is easy to repalce e.g. a RAID system or a server, and this task does not require a large administrative effort.
Another specialty is that you are really building one cluster for all participating systems / platforms. Thanks to the Extended Data Management (XDM), additional features like data mirroring, incremental synchronisation and Live Data Mobility (migration of the on-disk data of the running application) are also possible. Besides, the global view of all storage objects enables an easy realisation of the clusters with the n:n high availability. Compared to the widely-spread 1:1 clusters with other HA solutions, the resource costs are much lower.
Using RSIO it is possible to keep the infrastructure even more simple and further lower the costs. The existing RAID systems can continue to be used. RSIO servers in the Storage Cluster mode will then use the storage resources of the RAID systems in the backend and offer the virtualized storage over the RSIO protocol.
As RSIO clients Solaris or Linux servers get access to this already virtualized storage. In the daily administration there are practically no differences to the operation of a Storage Cluster node with a FC backend. The creation of the virtual Application Volumes, the data mirroring, backup and DR solutions are managed exactly the same way on the surface. Compared to other I/O protocols, the RSIO provides an already virtualized storage, and only administrative metadata are being transported from/to the client for mirroring, data mobility etc., noteably reducing the network load.
This concept can be scaled further by abandoning the usage of the RAID systems altogether and keeping enough storage resources e.g. in the RSIO servers themselves. This makes sense especially with fast (SAS/SATA SSD) or NVMe storage on the PCI bus.
Especially with ultra-fast internal storage (SSD, NVMe) and with a fast network connection very convincing throughputs and latencies are possible. A fast network for I/O can be used at the same also for a generic network connection ("converged networking").
A modular software structure enables simple handling and reasonable costs suitable for a specific project. The base package already contains fundamental virtualisation functions, I/O multipathing, all storage connectors, the cluster engine and an intuitive WebGUI. Solaris user have a choice between the VV module for the Virtual Storage that they already know and the new VSD technology. Unless higher availability is required, there is actually everything you need here.
For the management of the applications, resource-based self-management of the cluster and for the high availability the system can be extended with the Application Control Option (ACO). A tight connection between the application and the storage management stands out, providing a well-integrated solution.
With the Extended Data Management (XDM) there is a possibility to explore further functions like permanent mirorring, backup to disk or data mobility for running applications.
A special feature is that an increadibly simple and integrative approach of the OSL Storage Clusters can make a long-term impact by radically simplifying your data center processes, making you more agile and able to profit from the capabilities of the modern hardware available to reasonable prices. The software provides you with important management and availability features.
Especially for the customers with a Solaris background a single cluster for Solaris and Linux might notably reduce the complexity of potentially required migration.
Find more information in the OSL Storage Cluster 4.8 data sheet.
We are still at the beginning of our road with the new software generation. Especially the availability for Linux will bring a lot of change, because there are many things that can be done now, like a high number of applications, file services, backup solutions, integration with the Microsoft world and access to the latest hardware. In the second step, our HCI solution, the OSL Unfied Virtualisation Environment (UVE), will also profit from this.
At our house conference, "OSL Aktuell" on the 14th of June 2023, the abilities of the new software generation 4.8 will be highlighted in more detail, and we hope to be able to pass our exitement on to you. Please come and visit us!