IPs/Building Blocks

MindTree proactively invests in building intellectual property such as technology building blocks and frameworks. We believe this allows us to serve you better, helping you bring new products to market faster as well as reduce costs.

iSCSI Target Protocol Stack

MindTree has developed an independent high performance iSCSI target stack. Integrated with a Transport Engine (ToE), it has support for multiple connections, error recovery level 1 and 2, connection, and session re-instatement. The stack is customizable for different transport and SCSI interfaces, and portable with different versions of Linux (initiator and target), Windows (initiator), and Solaris (initiator). MindTree can provide source code licenses, and has already licensed the stack to multiple customers and integrated it with their products.

Key features of the iSCSI stack include:

• Zero copy in the I/O Path
• No context switch in data processing
• Thread pooling
• Dedicated threads for session establishment (no load on I/O threads)

Storage Management Framework

MindTree’s StorageMind is an independent framework for SAN-based device management. Capable of communicating with SAN devices using the emerging Common Information Model (CIM) as well as other non-standard management interfaces, the framework is implemented in Java using JDK1.3, runs on Linux 7.2, and uses MYSQL as the database. The Web functionality and interface allows the framework to manage devices from anywhere.

The SAN management framework-

• Discovers all SAN devices that respond to SNMP V1 requests and CIM
• Facilitates fault management of SAN devices
• Helps users configure their SAN devices
• Facilitates addition and removal of SAN devices

Global Name Space

With the growth of storage systems and data, managing disparate file systems is a real challenge for IT managers. MindTree’s Dynamic File Virtualizer (DFV) is a Distributed File System (DFS) solution that addresses enterprise data management by providing efficient file management services. DFV provides a common platform based on the concept of Global Name Space, an abstraction layer over the underlying file systems, for data administration across the network. Based on a search paradigm, DFV is dynamic, and different from traditional NAS and other distributed file systems. DFV virtualizes the storage and the actual location of the data behind it. A user can access all files by using a predefined query or defining new ones. Predefined queries are available as:

• Predefined share points via the CIFS (Common Internet File System) interface, or
• Predefined mount points via the NFS interface, or
• Plain “query-folders” or “Views” via the HTTP interface also commonly called as Web-Services

The “View” concept (or query-folders) makes it easy to find and organize your files. Viewing content in saved query-folders on the computer network adds flexibility. Query-folders also foster easy migration from one computer to another, while providing advanced navigation and a concise picture of the data.

Key features of the DFV include:

• Support for online expansion, contraction
• Query-based searchable logical views of data
• Type-centric search (including contents)
• Platform dependent for FS, all others are Java implementation
• Extensibility for encryption and translation