Projekt HSM
Hierarchical Storage Management
Overview
Motivation
The fast growing amount of information by companies and private users creates new
challenges for storage management systems. Also the heterogeneity and multimediality of documents and their distribution across different (e.g. mobile) devices and servers create new problems. Furthermore, the requirements for storage technologies and data management are very different for each document. Those requirements also affect problems of data security, data protection, data revision and availability of information over a long period.
One solution is offered by systems for hierarchical storage management (HSM), which extend classic filesystems. They address the mentioned requirements and enable a virtualization of different storage media (Image 1). Furthermore, they provide an automated and transparent access to a database and enable an efficient and cost-optimized management of the storage hierarchy. Moreover, existing legal requirements are considered for the deposition of legal relevant information. HSM systems also enable a long-time archiving and ensure that documents were not forged or modified.
Traditional HSM systems are often server-centered and disregard the management of multimedia documents, which increasingly are created by mobile devices. Thereof, we focus on the following research goals:
Research goals
In this project we want to design and create a HSM system with the following innovative features.
- Multimediality
Multimedia documents set new requirements to a data management because of their real time properties, their different formats and their internal logic.
- Long time archiving
An investigation of mechanisms for a life cycle management of heterogeneous information (in a collaborative use case) is needed.
- Optimization of the energy usage (greenIT)
There is a gap between a fast access and a cost- and energy-efficient long time storage.
- Virtualization of the storage hierachy
Distributed client-server-systems are often run with distributed hierarchical storage systems (cache, hard disk, online and offline media). A future HSM system has to virtualize this storage types, classify them and use logical storage targets.
- Semantic
Classical HSM systems do not consider semantic aspects of information. The integrations of such metadata may essentially increase the efficiency of a document management.
- Client-Server-Architecture
Documents are increasingly created and used in a distributed way. So all distributed storages should be involved into a HSM system. Thus, one goal is to create a distributed HSM architecture, which also could be used by mobile devices. Especially mobile devices create new requirements, because they want to use the storage system transparently and may be offline temporarily.
- Statistics and user profiles
Information about the user behavior and statistics about data may help to predict dataflows to increase the efficiency.