Scientific research, engineering design, and knowledge discovery are increasingly data-centric, and the effectiveness of an organization depends on its ability to collect, store, process, mine, and visualize data. Whether performing statistical analysis, driving design through numerical optimization, or exploring data, scientists, engineers and analysts require ready, efficient access to data servers that enable a variety of clients to access and manage the data repository. Midas Platform, an open-source data storage system, was created to support the emergence of data-centric computing.
Unlike conventional database systems, the Midas Platform is scalable, capable of handling massive amounts of data, and easily integrates with a variety of open-source computing tools including Kitware's industry-leading open-source systems the Visualization Toolkit (VTK), the Insight Toolkit (ITK), and ParaView. In this model, large data repositories are assembled with Midas Platform and auxiliary satellite processes are used to access, visualize, analyze, annotate, process, and manipulate the data. Midas Platform can be easily customized for integration into an organization's workflow, supporting data upload and download, batch processing, visualization, and interaction. As a web-based application, the Midas Platform uses a scalable client-server architecture to support multiple users and large data, with deployment as simple as using any standard web browser.
One of the most significant benefits of the Midas Platform is its ease of integration into existing products and processes. It allows for remote collaborators and terabytes of data to be seamlessly integrated into the workflow of individual research projects and large publication efforts. The key to the Midas Platform's ease of integration is the diversity of data, processing, and access formats that it supports. The following are some of the many benefits that Midas Platform provides:
||Midas Platform provides batch and interactive upload methods and supports arbitrary file types with arbitrary metadata fields. It can manage raw and processed data as well as companion reports and summary statistics.|
||Midas Platform understands scientific data, and extracts and indexes metadata in more than ten different files types including XML, Microsoft Word, and PDF documents. It also automatically indexes header information from more than twenty different image types, including MetaImage and DICOM header information.|
||Through the robust metadata, the Midas Platform enables powerful, flexible search capabilities. The extensible nature of the Midas Platform also enables the use of plug-ins to further extract additional metadata. The system supports the OAI data-harvesting standard, so that Google and other search engines readily discover public data. Private data is readily searched using a local search engine.
||Midas Platform provides automated processing and a scalable framework for server-side processing that enables high-throughput computing. Local, server-side, and distributed processing can be automatically initiated on upload of data to a Midas Server or initiated using thin or thick clients.|
||Midas Platform provides an open, self-documenting API that enables the construction of applications that interact with a Midas Server without the restraint of a browser. For many tasks involving the transmission of large data, the browser is limiting and cumbersome. Users can access data with Midas Platform without downloading the data, or downloading the data over the web or directly into legacy software using Pydas for programmatic access. Pydas is a Python client for Midas Server that provides a powerful suite of functions to integrate the Midas Platform into scripts.|
||Midas Platform readily integrates with a variety of visualization and analysis clients including ParaView, 3D Slicer, and VolView. This integration capability enables the rapid building of data-intensive applications that efficiently interface into custom workflows.|