Storage Virtualisation

OSL Storage Cluster (SC) is a host-based software that integrates virtualisation, clustering and storage networking. Without additional hardware it can simplify server and storage management, and also data center processes. It can also improve resource utilization and system performance. Without any hardware restrictions the OSL SC provides the resources of a RAID system (disk storage) to a group of compute nodes (cluster). Disk resources are assigned to a global storage pool. The cluster nodes can use the resources from this stroage pool depending on their storage requirements, without having to waste the storage.


Nowadays even professitional IT users cannot afford to use only a small part of the expensive storage resources. But there is also no reason to do so, because the modern storage virtualisation technologies not only improve the utilization of the full storage capacity, but they also improve the availability and offer additional investment incentives due to an accelerated Return on Investment. OSL implements such functionality with an integrated Volume Manager as a host-based software solution. This Volume Manager uses a virtualisation layer to allow dimensioning of the volumes independently from the hardware base underneath. At the same time, there is a potential to achieve performance advandages, e.g. with volume striping.

Traditional Storage Management
In traditional environments the volumes (LUNs) of the RAID systems are assigned exclusively to single hosts, using separate controllers and zoning, or LUN masking, or disk groups of a given Volume Manager. Free storage available on a LUN can therefore be used only on one dedicated host. For this reason, if a new host is connected, investing in new storage disks becomes necessary, despite still having free storage. The storage administrator might experience difficulties with having a good overview in the complex environments, because there is no central management framework with a uniform view onto all devices. The clearly-structured and easy to overview IT infrastructures are particularly valuable for the safety of operation, especially in critical situations after hardware failures.

Storage Management with OSL Storage Cluster
The global virtualisation of the OSL Storage Cluster enables a completely flexible utilization of all resources. A new system integrates seamlessly into the cluster and can use the free resources if necessary. The available capacities and bandwidths can be utilized in the most optimal way possible. Thanks to this, there is no need to make expensive investments into superfluous hardware. Very comfortable for the administrator is the uniform overview of all devices available through the global storage pools. This way, the work of the administrator is supported substantially by the software.

 Traditional Storage Management

Storage Management with OSL SC

SC 1

OSL Storage Cluster in a classical data center with FC SAN

The existing data center infrastructures can continue to be used. OSL Storage Cluster makes it possible to do the daily administration tasks comfortably from a single host. This includes e.g. provisioning of new environments / systems, data mirroring, backup or high availability. Multiple RAID systeme can be connected. A tight integration of all the components enables simple administration routines.
OSL Storage Cluster can be used on Solaris and Linux systems, also mixed with different versions / CPU architectures.


SC 2

OSL Storage Cluster in a DR-enabled configuration

Protection against disaster situations can be achieved with a cluster organized over multiple fire compartments or even different geographic locations. Also here OSL Storage Cluster offers suitable mechanisms and procedures, like data mirroring over different locations, replication of configuration data and procedures for a fast (restore-free) recovery, also after the loss of a whole location.


SC 3

OSL Storage Cluster with storage connected over Ethernet (OSL RSIO)

A distinguishing feature of the OSL Storage Cluster is the ability to transport the block storage virtualized in the cluster engine over network interfaces (Ethernet, Infiniband). This enables e.g. a setup illustrated here. Two cluster nodes are connected directly to the RAID systems, and at the same time they are working as RSIO servers. Remaining nodes access the virtualized block devices via Ethernet, and can also use all the functions of the global storage virtualisation. With 2x 1G-Ethernet it is possible to get about 230 MiB/s throughput per client, with 10G- and up to 400G-Ethernet even more, respectively. In general this enables a cost-efficient storage connection for a large number of machines.


SC 4

OSL Storage Cluster - Cluster nodes with internal storage as RSIO Server

Using internal storage in the RSIO Servers allows to avoid using external RAID systems completely. Considerable cost reduction and at the same time enormous performance gains can be achieved thanks to a much shorter data path, e.g. to the modern NVMe storage. And of course this setup can be redundant as well.

Supported hardware

Servers

OSL Storage Cluster is available for Solaris and/or Linux OS. Thanks to the modular design of the software and a clean integration into separate driver levels of the OS there are no additional platform requirements. Therefore the Storage Cluster can be used with a wide range of hardware. This is also valid for connected RAID systems. The choice depends on the existing data center environment and on the customer-specific decision criteria. Please follow the OSL recommendations when choosing the operating system (see software descriptions / data sheets). The available options for the server hardware are resulting from the chosen operating systems. In particular, the OS choice determines which network, SAN or storage controllers can be used. If you are not sure, please do not hesitate to contact your system partner or OSL.

Operating Systems

  • Linux (in principle, starting with kernel 4), currenlty recommended SLES 15.5. Other versions / distributions (e.g. Tumbleweed) upon request.
  • Solaris 10 / 11, since Version 7 upon request

Please also consider the release notes of the manufacturers for the related systems.

Administration of the virtual data center

The OSL Storage Cluster can be scaled to a large number of the clustered nodes and zones. The cluster stack is completely integrated and is delivered with a standard installation. In order to deal with all the tasks of your virtual data center, you can use the command-line interface or the web-based graphical user interface.

Command line
The complete management and operating of the virtual data center can be done on the command line. For this, the documentation in form of manual pages is provided. With the help of the command line you can administer your whole cluster from any cluster node. You do not need a dedicated manager node for this.

WebGUI - web-based graphical user interface
The graphical user interface of the OSL SC is integrated with the cluster and provides a comfortable way to control your virtualized data center. The web-based interface can be used with a regular browser (recommended Firefox). The WebGUI helps you to operate your cluster and to observe the state of every single node or zone. The WebGUI can be also used to manage the storage resources and the cluster applications. Furthermore, an extensive performance monitoring of the cluster nodes is available as a live log and also as a history log. This way a system administrator can be always informed of how much stress is on the systems.

SC-WebGUI

Hochverfügbarkeitscluster

Was den OSL Storage Cluster aus technologischer Sicht einzigartig werden lässt, ist die Verknüpfung von Speichervirtualisierung" target="_blank">Speichervirtualisierung und Clustering. Durch Kombination dieser beiden Technologien gelingt es der hostbasierten Softwarelösung von OSL, Applikationen auf dem Host zu integrieren. Nur die Betrachtungsweise vom Host aus ermöglicht schließlich eine ganzheitliche Sicht auf Anwendung, Betriebssystem, Hardwarebasis sowie ergänzende Soft- und Hardwaretechnologien. Wahre Applikationsorientierung kann demzufolge im Grunde nur per hostbasierter Software erreicht werden.


Cluster sind allgemein als sehr geeignete Lösungen für Hochverfügbarkeit und Ausfallsicherheit anerkannt. Anders als viele Wettbewerbslösungen benötigt der OSL Storage Cluster keine auf Standardnetzwerktechnik aufsetzende dedizierte und redundant ausgelegte Cluster-Interconnect-Verbindung. Mithilfe von im ausfallsicheren RAID-System abgelegten Informationen (Cluster-Datenbank) kommunizieren die Knoten über die Datenkanäle (z.B. FC-SAN) miteinander. Die dabei ausgetauschten Daten belegen  einen kaum messbaren Anteil der zur Verfügung stehenden Bandbreite.


Erreicht wird damit nicht zuletzt eine deutliche Verringerung der durch den Administrator zu verwaltenden Komplexität. Mit klarer Befehlssyntax, zentraler Administration von jedem beliebigen Clusterknoten aus und durchweg einheitlichen Gerätenamen besticht der OSL Storage Cluster mit weiterer Bediensicherheit.

Hochverfügbarkeitscluster

Was den OSL Storage Cluster aus technologischer Sicht einzigartig werden lässt, ist die Verknüpfung von Speichervirtualisierung" target="_blank">Speichervirtualisierung und Clustering. Durch Kombination dieser beiden Technologien gelingt es der hostbasierten Softwarelösung von OSL, Applikationen auf dem Host zu integrieren. Nur die Betrachtungsweise vom Host aus ermöglicht schließlich eine ganzheitliche Sicht auf Anwendung, Betriebssystem, Hardwarebasis sowie ergänzende Soft- und Hardwaretechnologien. Wahre Applikationsorientierung kann demzufolge im Grunde nur per hostbasierter Software erreicht werden.


Cluster sind allgemein als sehr geeignete Lösungen für Hochverfügbarkeit und Ausfallsicherheit anerkannt. Anders als viele Wettbewerbslösungen benötigt der OSL Storage Cluster keine auf Standardnetzwerktechnik aufsetzende dedizierte und redundant ausgelegte Cluster-Interconnect-Verbindung. Mithilfe von im ausfallsicheren RAID-System abgelegten Informationen (Cluster-Datenbank) kommunizieren die Knoten über die Datenkanäle (z.B. FC-SAN) miteinander. Die dabei ausgetauschten Daten belegen  einen kaum messbaren Anteil der zur Verfügung stehenden Bandbreite.


Erreicht wird damit nicht zuletzt eine deutliche Verringerung der durch den Administrator zu verwaltenden Komplexität. Mit klarer Befehlssyntax, zentraler Administration von jedem beliebigen Clusterknoten aus und durchweg einheitlichen Gerätenamen besticht der OSL Storage Cluster mit weiterer Bediensicherheit.

Hochverfügbarkeitscluster

Was den OSL Storage Cluster aus technologischer Sicht einzigartig werden lässt, ist die Verknüpfung von Speichervirtualisierung" target="_blank">Speichervirtualisierung und Clustering. Durch Kombination dieser beiden Technologien gelingt es der hostbasierten Softwarelösung von OSL, Applikationen auf dem Host zu integrieren. Nur die Betrachtungsweise vom Host aus ermöglicht schließlich eine ganzheitliche Sicht auf Anwendung, Betriebssystem, Hardwarebasis sowie ergänzende Soft- und Hardwaretechnologien. Wahre Applikationsorientierung kann demzufolge im Grunde nur per hostbasierter Software erreicht werden.


Cluster sind allgemein als sehr geeignete Lösungen für Hochverfügbarkeit und Ausfallsicherheit anerkannt. Anders als viele Wettbewerbslösungen benötigt der OSL Storage Cluster keine auf Standardnetzwerktechnik aufsetzende dedizierte und redundant ausgelegte Cluster-Interconnect-Verbindung. Mithilfe von im ausfallsicheren RAID-System abgelegten Informationen (Cluster-Datenbank) kommunizieren die Knoten über die Datenkanäle (z.B. FC-SAN) miteinander. Die dabei ausgetauschten Daten belegen  einen kaum messbaren Anteil der zur Verfügung stehenden Bandbreite.


Erreicht wird damit nicht zuletzt eine deutliche Verringerung der durch den Administrator zu verwaltenden Komplexität. Mit klarer Befehlssyntax, zentraler Administration von jedem beliebigen Clusterknoten aus und durchweg einheitlichen Gerätenamen besticht der OSL Storage Cluster mit weiterer Bediensicherheit.

Hochverfügbarkeitscluster

Was den OSL Storage Cluster aus technologischer Sicht einzigartig werden lässt, ist die Verknüpfung von Speichervirtualisierung" target="_blank">Speichervirtualisierung und Clustering. Durch Kombination dieser beiden Technologien gelingt es der hostbasierten Softwarelösung von OSL, Applikationen auf dem Host zu integrieren. Nur die Betrachtungsweise vom Host aus ermöglicht schließlich eine ganzheitliche Sicht auf Anwendung, Betriebssystem, Hardwarebasis sowie ergänzende Soft- und Hardwaretechnologien. Wahre Applikationsorientierung kann demzufolge im Grunde nur per hostbasierter Software erreicht werden.


Cluster sind allgemein als sehr geeignete Lösungen für Hochverfügbarkeit und Ausfallsicherheit anerkannt. Anders als viele Wettbewerbslösungen benötigt der OSL Storage Cluster keine auf Standardnetzwerktechnik aufsetzende dedizierte und redundant ausgelegte Cluster-Interconnect-Verbindung. Mithilfe von im ausfallsicheren RAID-System abgelegten Informationen (Cluster-Datenbank) kommunizieren die Knoten über die Datenkanäle (z.B. FC-SAN) miteinander. Die dabei ausgetauschten Daten belegen  einen kaum messbaren Anteil der zur Verfügung stehenden Bandbreite.


Erreicht wird damit nicht zuletzt eine deutliche Verringerung der durch den Administrator zu verwaltenden Komplexität. Mit klarer Befehlssyntax, zentraler Administration von jedem beliebigen Clusterknoten aus und durchweg einheitlichen Gerätenamen besticht der OSL Storage Cluster mit weiterer Bediensicherheit.

High availability cluster

A unique technological feature of the OSL Storage Cluster is the combination of storage virtualisation and clustering. Thanks to the combination of these two technologies it is possible to integration applications on the host. Only the host view enables eventually a comprehensive view on the application, operating system, hardware base and additional software and hardware technologies. Therefore a real application orientation can be only achieved with a host-based software.
Clusters are generally recognized as suitable solutions for high availability and protection agains failures. Different from many competitors, the OSL Storage Cluster does not need a dedicated cluster interconnect with redundant setup, based on the standard networking technology. With the help of the cluster database saved in the failure-protected RAID system, the cluster nodes communicate over the data channels (e.g. FC-SAN). The exchanged data use a very limited amount of bandwidth.

The OSL Storage Cluster has an efficient protection from hardware failures and takes care of migrating the applications to another piece of hardware in case of a disaster. Another suitable host takes over the application in such a case. Thus offline times are avoided. This approach requires, however, that an abstract access to the hardware resources is possible. Processes and applications must be described and adapted if necessary. This also leads to a radical reduction of the system administration complexity. Furthermore, the OSL Storage Cluster offers a comfortable management interface, with a clear command-line syntax, central administration from any cluster node, and uniform device names.

Active/active, active/passive or none of both?

Conventional cluster solutions are known to be either active/active or active/passive. The active/active solutions require at least two application servers to be online, and in case of the failure of one server, the second server takes over the duties of the first. The active/passive solutions require one standby-device to be online, which can take over the job of a primary system in case of its failure. Although such architectures are also supported, the OSL cluster concept is not limited to these two traditional approaches. The OSL Storage Cluster can also represent much more intelligent configurations. With our Application Control Option (ACO) the cluster software controls your cluster applications automatically via resource-based application self-management and assigns them to the hosts that are available according to their resource requirements. In cases of hardware failures the affected applications are reordered and assigned to the remaining hosts according to their priorities:

The OSL user assigns priorities to the applications. Based on the performance benchmarks, the application descriptions and priorities, the applications get assigned to different nodes in the cluster.  This can happen automatically depending on the capacity of the available hardware.

If a cluster node experiences a failure, the cluster launches the process of the redistribution of the applications over the remaining cluster nodes. Applications with higher priorities can even drive away the applications with lower priorities, so that the applications with higher priorities can stay online. This way the user can assign higher priorities to the production applications, and lower priorities to the test, development or QA applications, so that the important applications stay available even in case of disaster.

If the failed hardware later becomes online again, the stopped applications can be restarted. Another migration of the running applications is not necessarily required then.

Disaster Recovery

Disaster Recovery is possible if there is a completely redundant setup of all data center components used. To achieve this goal, especially large enterprises with the 24/7 availability requirement, rely on a second data center that is located in another place, so that the loss of the primary data center (e.g. in case of a power outage) can be overcome with the secondary data center hosting the most important applications. One of the OSL's goals is solving disaster recovery problems in an simple and efficient manner.

All large storage manufacturers offer also DR solutions in different forms. In most cases one large production RAID system is assumed, that is at least in parts permanently mirrored onto another location (emergency data center). If a production RAID system fails, the emergency data center should take over. In principle this is enought to achieve a large part of the desired protection against failures. Unfortunately this approach makes the operation of the while data center environment very complex, especially because of the required assignment Application ↔ LUN of the RAID system, that is not always easy to establish. The same challenges are known from the backup and restore concepts, which oftentimes do not allow a fast re-run in a serious case unfortunately.
Particularly critical use case are the database systems, where up until the moment of failure not only the database itself should be available, but also the log archive, online logs and control files (e.g. Oracle databases), so that a restore and re-run is possible without transaction losses. Thanks to the permanent mirroring there is also an additional problem of the instantaneous replication of the database errors onto the remote system. This way, also the mirrored data can become unusable at once.


A comprehensive solution concept for large databases/SAP systems with OSL Storage Cluster
The OSL Storage Cluster offers especially for large databases/SAP systems a comprehensive solution that combines Disaster Recovery, Backup-to-Disk, Backup-to-Tape and Instant Recovery. A balanced combination of the synchronous and asynchronous mirrors ensures a maximal protection from the typical error and failure situations. Our solution has the following distinguishing features:

  • "fast restore and re-run for databases" - especially for large databases oftentimes a problem
  • impressive performance and
  • low hardware requirements and therefore noticeable cost-reduction.

A big advantage compared to other solutions available on the market is the application orientation. An administrator defines Storage Cluster applications with all required resources (virtual interfaces, file systems, start/stop and break routines etc.) and gets with each application an easy-to-manage object that can be started and administered on any cluster node.


Application mirrors as Instant Recovery for Disaster Recovery purposes
Using a simple command, you can create mirror applications (XDM) for an existing application. Thanks to the OSL universe concept the production and the backup applications have the same name, but different universe extensions. This way it is easy to keep the overview even if there are many applications. Using a meaningful separation of the universes it is possible to set up desired backup scenarios, including permanently mirrored applications and regular Backups on Disk at defined times. The backup mirrors can be always synchronized with the production, and also disconnected atomically and therefore (time-)consistently. This way you can always have a startable image of the whole application on the disk, which can be used in disaster case. When using databases (e.g. Oracle), a backup mode is activated additionally, so that an online backup is available. In combination with the permanent mirrors (redologs) you have the possibility to roll the backup to any desired state (also automatically), without getting the tape involved. In a word: Instant Recovery.

The backup concept can also be extended with a tape connection solution (e.g. Networker). In this case the backup is first created on the disk, and then it will be written on tape by the central backup server. This approach has the following advantages:

  • The tape backup does put any additional load onto production, since only a backup application gets written on tape.
  • Only one backup server per cluster is used, and this way the licences for the backup clients are saved.
  • The tape data stream runs through dedicated SAN connections with notably increased performance. Therefore if the production data center fails, the production can continue running in the emergency data center with the help of the mirror applications.

Disaster Recovery

Application integration with Virtual Runtime Environments
The OSL Storage Cluster provides system resources to your applications via uniform, virtualized interfaces.  This concept is called Virtual Runtime Environments because of the customization of the environment parameters of the applications. Compated to the widely used virtual machines, these application environments are easy to administer because of the clear separation of the processes and services. Since the hardware is not separated by an abstract layer of a hypervisor, the applications in the Virtual Runtime Environments work wih the performance of a native execution.


Application as the pivotal point
Applications are the greatest good in the data center. As long as the applications can run without data loss, the hardware failure can usually be tolerated. The Virtual Runtime Environments consequently put the main focus of the data center operation on the applications. Only real application orientation helps to adequately cover further challenges like backup, archiving, system copy and disaster recovery or storage management. The OSL concept includes optimal host-based management of the processes by an intelligent software from the application's point of view.

Framework for virtual machines

The cluster framework offers the full comfort of the Application Control Option also for virtual machines under Solaris or Linux. Thanks to a custom abstraction layer, Solaris Zones, LDOMs and Linux-based hypervisor technologies are integrated using uniform interfaces and processes. On Linux you can choose between KVM, VirtualBox or XEN as a virtualisation techology for a virtual machine. The cluster framework takes care of the main functionality like provisioning, resource management, (live) migration, high availability, cluster-wide backup and disaster recovery.

Backup

The OSL Storage Cluster can create ready-to-run backups - while the application is running (online). For a long-term archiving these backups can be saved on tapes - without consequences for your productive applications.

Data must never get lost. This is one of the main goals of the IT infrastructures. But a safe keeping of the information is only one of the aspects of the data management. In order ot avoid the standstill times of the business critical applications, certain data sources must be constantly available for the productive access. These requirements can be met only by convincingly combining different technological and organizational approaches. The host-based software solution OSL Storage Cluster provides necessary instruments for this purpose.


From backup to archiving with an integrated and high available complete solution
The availability of the applications should be in the center of any backup concept. Running applications always require the access to the related data basis. If hardware fails, this pre-condition is endangered. This shorcoming cannot be eliminated by using redundant systems and mirror configurations alone. Because if we rely only on them, then in case of a disaster the application still would not be able to access the required data. Therefore further software intelligence is needed. OSL fulfills these requirements with an application-oriented solution, where the application's point of view is in the foreground.

Production data of business-critical applications are saved on the high-grade RAID systems (primary storage). With the help of the application resource descriptions and application-oriented mirroring mechanisms the OSL Storage Cluster software creates application images on the affordable secondary storage (e.g. SATA-RAID). These images get synchronized periodically and fast.  If the production storage fails, application images are immediately ready for the productive access - without any data loss thanks to the permanent mirroring of the log files. This way the restore step that usually precedes the restart is not necessary anymore. Production data are simply re-synchronized at runtime onto the repaired primary storage. If a node fails, the cluster engine ensures that the application can be taken over by another host.

If a backup node is connected, the writing of application images can be done comfortably, followed by the high-performnace disk-backup onto the tape (band streaming). The usage of the reasonably-priced SATA drives as a secondary storage is actually recommendable, since it is well-suitable for the sequential data flow, that is almost exclusively used here.

Backup