IT infrastructure software "Made in Germany" for self-hosters and providers
OSL Storage Cluster
Virtual Storage and Clustering all in one
Modern virtual storage and cluster engine.
Global pools, global administration, application HA ...
OSL Unified Virtualisation Environment
Hyper-Converged Infrastructure
Fully-virtualized VM infrastructure, radically simplified.
V3: Virtual Storage + Virtual Network + Virtual Server.
OSL Remote Storage I/O
Data Center I/O via Network
Virtual, cluster-enabled storage via Ethernet/IB.
Advanced features (performance, multipath).
Our principles
-
Specialized design
The OSL UVE was developed specifically to be used in a self-hosted data center with the purpose of storage virtualisation and clustering in mind. Therefore our solution has a hardware and software architecture design which is logical, easy to understand, robust and efficient. A highly portable storage virtualisation and cluster engine developed by OSL was customized specifically for the UVE to be used with an abstract (and therefore in principle exchangeable) hypervisor. This way separate layers of the complete solution work together perfectly, but at the same time without having disturbing dependencies, ensuring good performance and high availability.
-
Simplicity instead of complexity
This is one of the main principles for OSL: from design to the solution (hardware, software, processes, user interfaces) and all the way to the internal implementation in the process of development we focus on the simplicity and the reliability. Otherwise a small company like OSL would not be able to dare to compete with much larger enterprises and their complex solutions.
-
Standardization and openness
In the fast-developing IT world there is always a danger to follow a hype and to find oneself in a technological dead-end street or to be caught in a cost trap. The way for the commercial and professional users to deal with this challenge is probably in finding a balance between utilizing the chances of the new developments, on the one hand, without endangering the basis of the continuous and reliable business processes, on the other hand. Besides the many pseudo-standards, that basically only exist(ed) to secure the market positions, we can also observe technologies and standards staying stable in the long term, like e.g. ANSI-C, Ethernet and TCP/IP, SCSI or NVMe. OSL relies exclusively on a minimal set of well-established standards when developing own technologies. Thanks to this we are able to implement IT architectures with long-term stability that stay reliable even when replacing individual components. That means an increadibly free selection of the hardware and the system software for the end user, which offers certain protection from uncontrollably exploding costs. A vendor lock-in is avoided or at least reduced.
-
Avoiding disadvantageous dependencies
At first sight, the decision-makers might appear to have only three options to choose from:
- Standalone data center operation with commercial products: One falls into the hands of the market leaders and surrenders technologically and commercially to their business models.
- Cloud: One goes to the cloud, surrenders technologically and commercially to business models of the cloud providers ("XY as a Service") and forfeits the sovereignty over data and processes and also the related know-how.
- Standalone data center operation with Open Source: One relies consequently on Open Source. Unfortunately, however, many open-source projects have a focus different from operating a data center. And: they are usually not a product. The high amount of required engineering efforts makes you very much dependent on external consultants. Or: Maybe your IT experts always wanted to put a guru topic XY into their CVs, and after having completed an expensive and hopefully successful implementation you suddenly do not have any of your experts anymore. All this does not have much to do with your applications and processes.
OSL offers a fourth option - to prepare a working VM infrastructure in cooperation with partners, and to keep it running. While your system house that knows your business requirements implements together with OSL the optimal infrastructure according to the guidelines from your IT experts, your team has enough time to take care of the specific concerns and processes of your organization.
-
Shorter path to desired result
Of course also the OSL UVE cannot eliminate all the risks posed by complex hardware and software stacks (doesn't matter, whether coming from a market leader or based to a large extent on open-source projects). However: No carrier will get an idea to buy the vehicles as a construction kit and to let its own drivers or car mechanics to assemble them, not even if they know how to do it. In the same way, it makes sense to let the OSL partners to prepare your software-defined data center / your VM infrastructure, so that you can operate it professionally with your own team afterwards. You can stay focused on your business and as free as possible from the details and risks of complex systems.
-
Engaged support
We also had to have this experience: if one really has a problem, it can take weeks or months until the submitted support request finally reaches a point where a really proficient ingeneer or developer examines the issue. And of course one would wish it would go differently.
In the small OSL team everyone is more or less engineer, developer and supporter. Issue reports appear immediately on the radar and it is a matter of honour to not let the user "sweat it out". And: With the respective maintenance contract you talk from the very beginning to a system engineer that understands your problem and can help you right away or can at least directly initiate the necessary action.
OSL has a support model suitable for your company: from the maintenance 7x24 with 2 hours reaction time, to the pure software subscription.
-
Business ethics
OSL follows a cooperative business model with the focus on the mutual benefit, adheres to the strengthening of the autonomy and to acting responsibly, respectfully, in accordance with fundamental humanistic values. For OSL all the partners and clients are „key accounts“. We listen to you and for us the importance of your IT is obvious. Strategically, the autonomy, resilience and sustainability are more important than efficiency at any cost. In the end, what matters are the people and the quality of life. We have the chance to create something special together and for each other. The money is here only a means for organization and exchange. We are looking forward for inspiring collaboration and shared experiences.
-
Made in Germany
Our solutions, especially our software, are developed and supported exclusively by a well-coordinated, professional OSL team of permanent employees in Germany. Everywhere in this world there are people with good ideas and creative energy. We want to be a part of this community, but as a German company we are focused on the needs of the local and European users, we speak their language, in the good tradition of the German engineering we are questioning the dominant concepts of the IT metropolises, and we try - if necessary - to develop alternative, simple but still perfect solutions that would put smiles on the faces of our users.
Technology
-
Cluster-enabled cross-platform storage virtualisation
The internal engine of the OSL SC and the OSL UVE is a technology seamlessly integrating storage virtualisation and clustering out of the box. This way it is possible to use the infrastructure that is maybe already present in the data center (Solaris server, data-center RAID systems etc.) „on the side“ also as Virtualisation Server, which is a core of a VM infrastructure e.g. for x86-based applications (or also SPARC-/Solaris-based applications). OSL is probably the only provider on the market that can combine different CPU architectures, Solaris, Linux and as a guest system also Windows in one cluster. But of course it is possible to build one's VM infrastructure completely from the scratch, completely Linux-based. Together with the partners OSL can create a suitable service and support package depending on the user requirements.
The availability of the portable storage virtualisation enables e.g. hypervisor-independent live data migration, synchronous data mirroring, creating data copies, clones and a lot more. Separate RAID systems are not necessarily required in the OSL UVE. But if they are present, they can also be integrated easily. These storage virtualisation capabilities are missing in many other solutions or are available only as optional features (at additional costs).
Virtual Machines usually run on the block I/O interfaces, that might be mapped to files in a filesystem. The UVE works directly with (better performing) block I/O interfaces without making a detour around a filesystem. Special functionality, like e.g. expansion of the volumes, is provided by the storage virtualisation.
-
Coexistence of the old and the new systems
OSL UVE relies on a minimal set of well-established standards. This way it is equally possible to integrate both older systems that you might already have in your data center (e.g. those with 1G-Ethernet), and super fast, modern servers with 200G-Ethernet or Infiniband. The same opennes applies also for storage systems, where (almost) anything is possible, from internal NVMe-SSDs up to the RAID systems connected over FC or iSCSI. Also for OSL Storage Cluster you have a wide selection of storage systems, SAN and LAN technologies and servers.
-
Wisely designed, convergent network
A wisely designed network concept combines two (or more) independent networks in such a way, that the inteface visible in the guest works in a high-available / redundant way without any special precautions. The need for bonds, teaming, LACP etc. disappears. Switches and network cards of different manufacturers and generations can be used side by side. With the Virtual Network Objects developed by OSL the interfaces of the respective hypervisor OS are managed in a platform-independent, easy-to-over manner. This is done automatically by the UVE engine, the administrator usually does not have much or anything to do with it. At the same time, the software takes care of a possible separation of networks of different clients, so that different worlds separated from each other can run in one cluster.
And last but not least: the practice has shown, that generic network traffic and block I/O can run over the same media without any problems. What is immediately obvious with 100/200G-Ethernet, in many cases works even with 1G-networks quite decently. And: there is no need to use special switches or network cards. A manageless switch saves the need to manage it – the relevant network configuration is done by the software on the participating virtualisation servers or clients.
-
Block-I/O via RSIO
RSIO (Remote Storage I/O) is a protocol developed by OSL for data-center block I/O over network interfaces (Ethernet, IB). RSIO does not only enable an efficient, parallel and high-available data transfer over available network paths. From the very start it was designed to be used with modern networks and clusters, and therefore it can communicate with the virtualisation engine in advanced way. Thanks to this, the UVCs (Virtualisation Clients/Compute Nodes) can use alle the relevant functions of the storage virtualisation, without the storage virtualisation running on the UVC. The storage virtualisation comes as a service from the UVS to the UVC client and therefore to the VMs. At the same time, the required network bandwidth, e.g. for data mirroring, is radically reduced compared to other solutions on the market. RSIO can also be used in OSL Storage Cluster, allowing the hosts without FC ports to "co-use" the storage infrastracture availbale in the backend. RSIO servers can be used also with internal disks, so that the RAID systems can be skipped completely.
-
Hyper-convergence in a client-server model
The OSL UVE as a hyper-convergent system is consequently software-centered in the aspects of design and daily use. This means that an administrator in the everyday operation of the data center does not have to be busy with the hardware configurations (like storage systems, switches etc.) anymore. Different from the known HCI solutions from the overseas, that mostly have a brick layout, the OSL UVE has a client-server design. This approach brings many advantages - shorter I/O paths, faster IPC in the server instead of expensive inter-node communication with the brick design, more simple and robust flows and a lot more. The theoretically unlimited scalability of the brick layouts is for the most part invalidated by the practical obstacles. In addition, it is probably irrelevant for almost all users in Germany / Europe. In the end, the OSL UVE scales better in practice and offers considerable advantages.
-
Organization Entities, Roles, Monitoring, WebGUI and an intuitive CLI
The OSL UVE offers for an administrator a choice between a WebGUI that can be easily used intuitively, and an easy-to-learn CLI, that can be a good alternative especially for a professional use, when there are hundreds or thousands of objects to administer. The access over the WebGUI can be restricted and customized using pre-defined and custom Organization Entities, permissions and roles. An integrated monitoring enables neatly visualized real-time analyses and high-resolution representations of the stress situations from the past.
-
One-click console, proxy and desktop agents
With a special proxy and desktop agent developed by OSL it is possible – depending on granted permissions – to access the console of a VM over (No)VNC or other protocols, also during or after live migration of a VM. These consoles can also be used in common sessions with multiple users, e.g. for support purposes, with the possibility to restrict possible interactions if desired.
-
Hypervisor abstraction
A VM is a VM, is a VM. It has RAM, CPU, storage disks, network interfaces (and sometimes also some additional special devices). OSL UVE works with an abstract model of a VM, so that it is possible for a user to change the VM hypervisor (with certain restrictions).
Highlights
Self-hosted IT infrastructure can be vastly superior in terms of performance, technological flexibility and costs. Here some of the many OSL software features:
Global Storage Virtualisation
enables a fully flexible usage of all the storage resources, without having to enforce a static assignment to specific hosts.
Use the hardware of your choice!
We develop our software for it to be manufacturer-independent, so that you can keep your flexibility.
Backups and Archiving
are possible at the runtime. Minimize your downtime with the OSL software.
High Availability
of your data can be garanteed by using an intelligent solution with priority-assignment without additional hardware.
Disaster Recovery
using synchronous and asynchronous mirroring brings impressive performance even with limited hardware resources.
Performance Monitoring
The performance of the nodes and the virtual machines in the cluster can be observed using our real-time monitoring and the saved performance history.
OSL Network Virtualisation
enables an economical and manufacturer-independent multi-site interconnection of your data center.
WebGUI
Uncomplicated management and administraton of the cluster, the VMs and the users is possible by using our graphical user interface.
Live Migration
Migrate running applications like Zones and Virtual Machines easily and without the need to stop the machine.
Personal Support
is available to all our clients. You can use our Shared Terminal for a secure remote support.