Cloud MSW: Similarity Searchable DBMS
This brochure describes a similarity searchable DBMS - Cloud MSW. The DBMS can be used by people who develop big data applications and require distributed data storage, multiuser access and quick data similarity search.
ACTIVE DATABASE ARCHITECTURE FOR XML DOCUMENTS
Our presentation for the CAINE-2008 international conference (Andrey Logvinov)
Single-attribute Distributed Metrized Small World Data Structure
Our paper, which was sent to IEEE International Conference on Intelligent Computing and Intelligent Systems 2009. In this paper we discuss the basic variant of the structure which can store data units with a single searchable attribute in the form of a unique string and any amount of non-searchable data. One important application of this form of the structure are content-addressable systems (CAS) which map a content- derived identifier to the content element. Simulation results demonstrate that the proposed structure has logarithmical search complexity, avoids access bottlenecks and can be dynamically scaled to large numbers of data units.
Content Addressable Storage (in Russian)
This presentation overviews the structure of content-addressable storages (CAS) and their general characteristics. The paper also reviews the possibility of use of Metrized Small World (MSW) data structure for creation of CAS systems.
Analysis of service network architectures
A study of the different types of service network topologies conducted by Vladimir Krylov. The author created a mathematical model of networks and studied their properties in terms of network node availability, system availability, network scalability and the costs associated with sustaining the current level of performance. The results of this work are especially interesting if the conclusions are used in application to communication networks.
Metrized Small World Properties Based Data Structure
This paper gives an overview of an innovative data structure named Metrized Small World (MSW). The data stored in the structure is represented as a set of structured storage units on which a semi-metric can be defined which characterizes the relative relevance of each unit. Then a complex graph can be constructed whose vertices are the storage units and the edges are selected in such a way that the graph has the small world properties and is in accordance with the introduced metric. Due to the special properties of the constructed graph, the search is accomplished on average in the number of steps logarithmic of the storage size. The paper covers the main algorithms on the data structure as well as the properties of MSW data structure which were obtained experimentally on the prototype of the structure. See also the presentation for this paper.
Metrized Small World Based Data Structure
This paper describes in detail the Metrized Small World (MSW) data structure. The main principle is to link all data units whereby every data unit has to have the local only link list to provide the associative search mechanism by roaming in compliance to the links to all data units which are relevant to the input search mask. The known Small World stochastic graphs space equipped by semi-metric model is used to create the links for every new data unit. The XML data unit semi-metrics are developed and investigated to use this approach to create the distributed index. Testing of the search features has shown that the theoretical assumptions are correct and we can use the MSW approach to build the petabyte and exabyte storages.
Metrized Small World Properties Based Data Structure (presentation)
The presentation for Metrized Small World Properties Based Data Structure paper presented at the SEDE 2008 conference in Los Angeles.
Active XML Database Architecture
This paper gives an overview of the new database architecture to build very large, scalable, loosely structured distributed data storage. The main idea is that every data unit is active, capable of communication with other data units and database clients. Combined with the special overlay structure incrementally formed by data units (Metrized Small World Graph) this provides for effective distribution of data units among database servers and unbounded scalability of the resulting storage ensuring logarithmic search and append complexity. The paper covers the structure of the database, its properties and essential algorithms used in its operation.
RFID Based Document Storage System (presentation)
Conceptual view of a paper document storage facility which uses RFID tags to link paper documents to their digital descriptions stored in the MSW data storage and to determine their physical location within the facility.
EPC Object Code Mapping Service software architecture Web approach (presentation)
The concept of a system which stores Web-accessible product unit descriptions for EPC-tagged products. The description is dynamically updated as the product unit moves through its lifecycle stages.
Global scale inventory system concept based on RFID and Web technologies (presentation) (in Russian)
Global scale inventory system concept where each inventory item receives a RFID tag with a unique code and a globally accessible Web representation. This presentation is in Russian.
Active Database Architecture for XML documents
This paper introduces an innovative database architecture intended to build very large, scalable, loosely structured distributed data storage. The main idea is that every data unit is active, capable of communication with other data units and database clients. Combined with the special overlay structure incrementally formed by data units (Metrized Small World Graph) this provides for effective distribution of data units among database servers and unbounded scalability of the resulting storage ensuring logarithmic search and append complexity. The paper covers the structure of the database, its properties and essential algorithms used in its operation.
How to use Skoal
A set of instructions to the demo version of Skoal web-service, a similarity searchable database of chemical compounds.