HomeTechnologyComponents → OntoFS
 
 
OntoFS - Ontological File System
   
 

Introduction
OntoFS is a new file system.

Advantages
Ontologic File Systems in general have a very high level of

  • speed,
  • flexibility and
  • security.
Compared with the other relevant semantic store systems, which are all based on some kind of a database management system like triple stores, our file system-based approach of the OntoFS architecture isn't faster, but exorbitant faster. It just plays in a total different league and this even holds if our OntoFSs are compared to the few other solutions that are based on a pure file-based approach.

Another reason that makes Ontologic File Systems extremely interesting is, that they integrate file system and database management system funtionalities. In this way the additional

  • installation,
  • definition,
  • optimization and
  • technical service
of databases is simply said needless.

By integrating the ontology-oriented and ontologic paradigms as well, the foundation of

are parts of the Ontologic File System.

Furthermore, due to the fact that in hightech operating systems the file system is the central and extremly important part more features and benefits can be established. Prominent examples for these features are the

But to end here was for sure not enough for us, so that we applied the already existing features for the integration of our

File system
The Ontologic File System has the following basic features:

  • It was designed with high-security and high-speed in mind;
  • It's very fast, often the fastest in many tests, especially by its basic data structure and the highly efficient and intelligent integration with modern operating system architectures and their kernels;
  • It uses B+- or B*-trees as the basic data structures, like they are used by every modern database management system as well;
  • It uses B*-trees in conjunction with dancing trees for very space efficient packing and handling of small files;
  • It uses B+- or B*-trees for very fast handling of large files, too;
  • Its disk-layout is dynamically optimized through allocate-on- flush/delayed allocation;
  • Its atomicity supports transactions, so that file system oper- ations entirely occur (see also ACID principal);
  • Metadata is stored as sub-files, so a file is actually a folder and a file, and metadata can be attached to any other file;
  • It has an efficient journaling through wandering logs;
  • It has a flexible plug-in system;
  • It supports transparent compression, which can be customized and optimized as needed by developing special plug-ins;
  • It's mature and works since more than 4 years;
  • Simply said: It's a dream of a file system that has become true.

In this way, our file system has also defined implicitly a superset of an approach in the field of database management systems that is called Not Only SQL (NOSQL). Instead of splitting this approach into its main categories, which are labeled document-oriented, column-oriented, key-value, and graph databases/stores, the OntoFS unites directly from its start all of their foundational concepts with many other paradigms, like the row-oriented one, without compromising the consistency of the stored data.

Ontologic File System
By integrating semantic (world wide) web technologies, especially the Resource Description Framework (RDF), the Web Ontology Language (OWL), but also the object-oriented paradigm, the relational database paradigm (third normal form for referential integrity), and additional artificial intelligence-based methods, we have enhanced the paradigms of standard file and database management systems with the ability of

  • file management system defining,
  • file indexing,
  • file searching,
  • file system self-managing, and
  • further file management system enhancing
based on the semantics stored in the data itself, the metadata of a file or even of the whole file system.

Now, users and most fascinating the information processing system itself are able to

  • interact with the ontologic file systems by browsing graphs, or trees/hierarchies of files/directories,
  • query the file systems like information retrieval systems,
  • orthogonal persistently store data,
  • orthogonally process, as well as to
  • reflect.

Data Integrity
By using the plug-in system of the file system, data backup and re- covery functionality can be integrated easily and much more per- formant processed than by other ways. Specialized software already does the latter with file systems in general by bypassing the related functionality in the so called Virtual File System (VFS) software layer of Linux®. But with the basic concept and architecture of our file system OntoFS this is done in an ordered way, which supports third parties in their efforts to add and extend such operations.

Operating System Security
The foundational concept of the operating system Unix® and all re- lated operating systems, especially Linux®, is simply described by one short sentence: Everything in Unix® is a file. This means that not only data are handled as a file, but also all processes, inter-process communcation between processes, pipes between software applications, channels between computers and devices.
With its special concept that is based on the metadata of a file, the Security-Enhanced Linux® (SELinux) approach by the National Se- curity Agency/Central Security Service gives every user the option to add into a Linux® operating system mandatory access controls that can confine very flexibly and fine grained the actions of any process, including an administrator process. This means in the end to have control over the access to everything. "The security mech- anisms [type enforcement, role-based access control and (military) multi-level security] implemented in the system [as a Linux® securi- ty module] provide flexible support for a wide range of security pol- icies. They make it possible to configure the system to meet a wide range of security requirements. The flexibility of the system allows the policy to be modified and extended to customize the security policy as required for any given installation. [...] The Flask archi- tecture demonstrated in the Security-Enhanced Linux® reference implementation has been ported to several other operating systems, the Xen hypervisor/virtual machine monitor, and has been applied to applications such as the X Window System, []bus-system and []SQL database.", [National Security Agency, 2009].

One of the most important and fascinating feature of OntoFS is in this conjunction its full file metadata support, so that the approach of SELinux is directly and optimally supported. Indeed, we do not know a better and more elegant way of how to do it as with this combination.
The implications of the support of the file based security approach by SELinux with our OntoFS are as follows:

  • The combination of the unique features of OntoFS and SELinux realizes a flexibility to control the data access that is normally known from fully fledged database management systems, with which the access to every row of a database table can be controlled. This fact is an extremely nice and powerful example that our OntoFS file system has indeed database management system features, quantities and qualities, so that not only re- dundant functionality of software components and applications can disappear, but in this case database management systems become obsolete in general;
  • By its reflective and intelligent characteristic, the operating system supports the user with the very complex task to define and apply tens of thousands of policy rules and exceptions, and helps to realize and manage other security elements of a secure system such as system assurance or security audit- ings;
  • Furthermore, by the plug-in mechanism cryptographic algo- rithms can be added to the file system and integrated into security related operations;
  • A certification system can be supported by the security mech- anism of the file system as well;
  • Another option given with the OntoFS plug-in system is to create plug-ins for protecting the operating system against malicious processes that come from external sources;
  • Even better, the warranty of security is done how it has to be done, namely by the operating system and not an application like a web browser;
  • The concept of virtual machines is realized how it has to be realized, namely as an integrated part in a core operating sy- stem, that runs directly on top of the hardware, and not some software layers above in an application like for example a web browser. The integration of virtual machines into web browsers was only made due to missing security features in all major op- erating systems and the attempt to push proprietary internet programming language standards into the market. To bypass this deficit of unimplemented security features and to secure the establishing of such an own standard, a so called sandbox concept was created in the internet browsers with the inte- gration of related virtual machines, but not in the operating system where a sandbox concept has to be from the point of view of software technology and developers of operating syst- em architectures since decades.
    The elimination of virtual machines in web browsers does not imply that the whole concept of interpreted object-oriented programming languages is thrown away, but it keeps operating system tasks out of a programming language. An interpreter or virtual machine of a programming language hasn't the task to establish a sandbox, for example. It also simply defines every desktop application written in any programming language as a web enabled (mini) app(lication);
  • At the end, the always obsolete layer introduced by the con- cept of web browsers can be omitted again;

Finally, the results of this security approach and its implications are

  • shorter process paths,
  • less administrative overhead and
  • less possibilities to attack a computer system,
and a net gain by
  • more speed,
  • more security,
  • more flexibility,
  • more activity,
  • more productivity and
  • more fun.
If at all a now possible overhead is induced by the security part of the operating system, then it can be drastically reduced or even al- most completely eliminated in many cases by a smarter usage of all operating system features.

Ontologic Virtual Reality Environment
At the heart of the Ontologic File System is as the foundation geo- metry, so that the Ontologic Scene Graph, which as a super set is including semantic and ontology based scene graphs as well, of the Ontologic Collaborative Ontologic Virtual Environment (OntoCOVE), which is an essential part of our Ontoscope component, is seemless- ly integrated as well. This is done by simply interfering the file syst- em structure with the 3D scene graph structure through the ontolo- gy-oriented paradigm. At the end this means: If a user moves a file, then with its content the sensorial (visual, acustical, tactual and/or others) representation is as well simultaneously moved in 4 dimens- ions. But the true revolutionary part of this technology is, that in fact an object of the universe is moved in the universe, which opens the gate for undreamed-of future technologies (see the Terms of the 21st Century, especially Pure Rationality, Teleportation #1 Stellar , Teleportation #2 Interstellar and Warp Velocity and Hyperspace, but also Replicator #1 Molecular Assembler).

Sources
Actually we offer two different versions as the foundation for im- plementing an OntoFS, which are the:

  • real file system based on
    • our fork of the Reiser4 file system, or
    • the transformation of a virtual file system to a real file system without changing conceptually the existing Application Programming Interface (API) of the used libraries,
    and
  • virtual file system based on
    • the integration of a file system, a database engine, and a gluing virtual file system.

Real OntoFS
In former times the Real OntoFS implementation was based on the Reiser4 file system until its development seems to have finally ended in the Summer of 2011. Due to the low support of the file system by the developer community we had to remove this option, and are now focusing on two alternatives, which are

  • our announced fork with a totally overworked architecture and source code of the Reiser4 file system, as we have already designed in the December of 2009, and
  • the integration of a file system of the Linux® standard file system series, the extended file system series, with a data- base library, like SQLite and Berkley DB.
The replacment of the actual virtual version with a real version can be realized by substituting the software libraries Filesystem in USEr- space (FUSE) and Libsqlfs with either the: in the near future. This substitution will not change conceptually the existing Application Programming Interfaces (APIs) of the used Ext- ended FileSystems and database libraries, and the Libsqlfs userspace file system library, so that interested persons and entities can alrea- dy base their developments and existing software solutions on the actual APIs.

You can find the old content about the former connection to the Reiser4 file system with our OntoFS, especially of this paragraph, on the Reiser4 file system webpage.

Virtual OntoFS

Virtual OntoFS Installation
The installations of the needed packages are given on their web- pages.

OntoFS Manual
A manual for the OntoFS is coming soon.

   
 
© and/or ® 2006-2012
Christian Stroetmann GmbH
Disclaimer