In general, storage resources are allocated to each application separately. As it is almost impossible to define storage requirements of an application exactly in advance administrators are usually allocating too much storage to be prepared for the worst case. But so plenty of capacity is wasted in all other case.

With the intelligent storage management of OSL Storage Cluster on behalf of global storage pools there is no excess or wasted capacity remaining. OSL's device driver allows to create virtual volumes from different disks or LUNs. Such a virtual volume may well consist of physically separate disks from different storage systems. But disks can not only be put together as an assembly. It is also possible to cut any size of volume from a disk or to put a volume together from parts of different disks (such as waste capacities).

With this concept OSL is going a step further than the well-known technology of thin provisioning. No only virtually existing storage has to be pretended. The administrator can increase virtual volumes easily on demand. Thanks to the waste-free allocation lavishness of expensive RAID resources is part of the past. The software is also able to increase performance by striping disks.

Clusters are generally recognized as very suitable solutions for high availability and reliability. But they are also known, in part because of old-fashioned software concepts of some well-known solutions, as "elitist and complex". This is a main obstacle for small and medium business customers to use modern cluster technology. OSL does away with this prejudice. Unlike many competetive solutions OSL Storage Cluster does not require a redundant cluster interconnect. On behalf of centralized stored information (cluster database in RAID system) the cluster nodes are communicating over storage network (eg FC-SAN). The transferred cluster-own data needs even with older hardware a barely measurable bandwidth of the storage connection.

So not at least a significant reduction of complexity is achieved. By the way, OSL Storage Cluster avoids with that concept a cluster typical problem: With conventional cluster software it is never clear if there is a real failure of the host or only a problem of the heartbeat. With a clear command syntax, centralized administration from any cluster node and consistent uniform device names OSL Storage Cluster stands out with ease of use rather than unnecessary complexity.

Conventional cluster solutions usually rely on standby machines, which are jumping in if a productive server quits the service. Depending on the operating condition of the secondary machine this is called active / active or active / passive cluster. To obtain high availability, a very high price has to be paid: One has to steadily retain current hardware, which would not be done regularly.

OSL follows also at this point a different path. With OSL Storage Cluster, you can assign a priority to your applications. Thus, for example, productive business applications are prioritized higher, test or QA systems even lower. If a productive server fails, the application moves to the server running the lowest priority application. The less important application is shut down and restarts as soon as sufficient server capacity is available again.

This intelligent application migration is supported by a resource-based self-management. The cluster software automatically identifies by benchmark, whether the computing power of a server is sufficient for an application or not. Optionally OSL Storage Cluster is aware to assign all applications to cluster nodes according their cpu and memory needs automatically (resource based self management / adaptive computing).

Please find more information about high availability under Solutions.

Thanks to enormously increased capacities of individual servers, using the virtual environments for applications in data centers has become very common nowadays. This way, data center operators can improve the load distribution on their systems through resource sharing, while still keeping applications isolated from each other.

But especially in business critical fields where a lot of administrative work is done for tasks like backup, restore processes and disaster prevention, VMs have considerable weaknesses.

OSL Storage Cluster and OSL Unified Virtualisation Environment provide a compelling alternative to classical virtualisation technologies and help to keep your VM environments under control. A full integration with custom backup procedures is possible in almost every case.

OSL Storage Cluster consistently implements a central administrative concept being very easy to understand. The whole high availability environment can be managed from any cluster node. Devices and paths are named the same regardless from which host the information is requested. The nomenclature for virtual volumes can be freely defined. Where competing solutions come up with complicated but maybe arty browser interfaces OSL Storage Cluster convinces with a clear command-line syntax. This provides clarity and helps to keep a clear head in complex environments and critical situations.

OSL customers usually know the value of consistently user-friendly designed tools and their editing ability with standard Unix tools. Those who are preferring the comfort of a curses-based configuration tools can of course make use of its features.

Modern business management software is a great support for sales, administration and management. Unfortunately, it is good enough to block any I/O in the SAN through huge database queries. To handle such situations more smoothly, or better yet, prevent them completely, OSL Storage Cluster 3.1 has an integrated bandwidth control. For individual volumes or complete applications this allows both strict and flexible load-sensitive bandwidth limitations. As a consequence, all systems always remain manageable. Productive applications won't get disabled any longer completely.

Further, an important use case of this function exists beside I/O regulation. Bandwidth control can restrict resource requirements of synchronization operations in mirroring environments effectively. So even at high traffics trouble-free operation of productive applications can be secured.