IRI is an independent software vendor (ISV) specializing in commercial data processing, and paticularly, the high-performance transformation, migration, reporting, and protection of structured big data sources. The company was founded as Information Resources, Inc. of New York in May 1978 and got its start moving coroutine sort (CoSort) technology off the mainframe and onto personal and mini computers -- CP/M 80 and 86 in 1980 and MS-DOS in 1982. CoSort was then the first commercial Unix sort package, developed in C on an AT&T 3B2 in 1985.
IRI began to grow more rapidly in the early 1990s amid a sort platform downsizing trend that continues to this day. Business improved further from CoSort’s adoption in the world's largest Unix data warehouses, and from its unparalleled flexibility in Windows environments. Since the advent of commercial Unix, all major hardware manufacturers, including: HP, IBM, Intel, Siemens, Sun, and Unisys, have continued to seek CoSort benchmarks and cross-certifications while recommending CoSort to their customers. In addition, many leading DBMS, data warehousing, and vertical-industry ISVs -- as well as consultants in these spaces -- embed or recommend the use of CoSort technology within applications to improve performance in high data volume environments.
During a key expansion in 1995, IRI moved to Melbourne, on Florida's high technology "Space Coast," and changed its name to Innovative Routines International, Inc. The Company moved to larger space again in 1996, and announced the release of CoSort for Windows NT, Windows 95 and OS/2 Warp. In 1997, CoSort became the first Unix sort package to run across SMP CPUs and set industry performance standards. CoSort version 6.2 sorted a gigabyte in under a minute. Later that year, PC Week declared CoSort the fastest sort for Windows. In 1999, CoSort 7 was introduced, and featured the only single-pass join technology in the sort market, in addition to extensive drill-down aggregation and cross-calculation functionality. Multi-threaded for SMP servers, CoSort 7 also introduced a unique Java GUI to allow users to read and write SortCL specifications and then execute them locally or any networked Windows or Unix platform.
In succeeding years, CoSort saw increased adoption in database and data warehouse extract-transform-load (ETL) operations, vertical market application software, and e-commerce (clickstream analytics). Version 7.5 was an interim release in 2001 that enhanced internet and international data type support, made the original CoSort API thread-safe, and improved aggregation and reporting features in SortCL. In 2002, IRI also lead the sort world by porting CoSort to Intel's 64-bit Itanium platform and IBM eServer iSeries (AS/400) for Linux and OS/400 PASE. 2003 marked IRI's 25th year and a new "V8 Engine" in CoSort. Using just 6 Sun 12K CPUs, CoSort sorted 2.4GB in 39 seconds, and with just 4 CPUs on an IBM p690, CoSort sorted 1GB in 12 seconds, which in those days was very fast!. To serve its rapidly growing base of data warehouse architects, CoSort 8 introduced a new API call for the "sortcl_routine", support for click-stream data types and web log metadata, plus markup language formatting in SortCL for web reporting. Support was announced for Linux and HP-UX on Itanium, and for Linux on IBM zSeries mainframes.
In 2004, IRI enhanced both CoSort's transformation performance and its "point solutions" for its cross-platform users. The Fast Extract (FACT) product was introduced to rapidly unload Oracle tables and build metadata for instant ETL flows through CoSort's SortCL (transformation and reporting engine) and SQL*Loader. That year IRI also "YES!"-certified CoSort for SUSE Linux (SLES) 9 through Novell, and for use on IBM's 64-bit Power5 processor running AIX 5.3. In 2005, IRI released CoSort v8.2, and a robust test data generator and custom file synthesizer based on SortCL called RowGen. That year IRI made CoSort "Red Hat Ready" for Enterprise Linux, and released versions for Solaris 10 on x86, FreeBSD 5.3, and the then latest Unix, Linux and Windows platforms. In 2006, IRI updated its exclusive sort plug-ins for IBM's DataStage and Informatica's PowerCenter ETL suites, and partnered with Meta Integration Technology to automatically convert file layouts in ETL and BI tool repositories to SortCL and RowGen data definition files. IRI continued to expand its base of resellers and expert consultants around the world, and worked hard to enhance the CoSort, FACT, and RowGen tool sets.
2007 marked the start of the CoSort Version 9 era and a major expansion of functionality and services for solution architects, IT managers, compliance officers, and application developers. In addition to its already combined file processing and presentation (manipulation and reporting) functions (e.g. change data capture), CoSort's SortCL tool uniquely integrated field-level protections for sensitive files, plus safe test data generation and complex transformations -- all into that same job script and I/O pass. Also introduced as unique features were: built-in processing and cross-conversion support for multi-byte data and file types, including Shift-JIS, Big 5, Unicode, ISAM, LDIF, and XML; multi-file joins and dimensional lookups; and, integrated Perl Compatible Regular Expression (PCRE) logic. In 2008, RowGen v2 went into general release for the creation of massive, safe, and intelligent database table and flat-file test sets. IRI also introduced more spin-offs from SortCL, including a powerful data masking and encryption tool for content-aware data loss prevention (DLP) called FieldShield, and a file-format and data-type conversion tool called NextForm.
In 2009, IRI released improvements to FieldShield and NextForm, and began work on another significant upgrade in CoSort functionality and ergonomics. Released in 2011, CoSort v9.5 can source and target data in relational databases, convert between Unicode and native multi-byte Asian character sets, recognize endianness at the field level, and handle more date and mainframe numeric data types. Perhaps the most exciting aspect of CoSort 9.5, however, was its new Graphical User Interface (GUI). The "IRI Workbench" is an Eclipse plug-in with wizards for job creation, metadata discovery and conversion, and job tuning. This visual, Integrated Development Environment (IDE) for all IRI products features purpose-driven, end-to-end job wizards, dialogs, and graphical componentry linked to a syntax-aware editor for SortCL (and the metadata-compatible spin-off product) scripting. The IDE also features database views, version control, metadata management, reporting, and client/server execution capabilities designed for enterprise data management.
By 2012, CoSort 9.5.2 was shipping with more data warehousing and reporting features like pivoting, slowly-changing dimension and fuzzy logic lookups for discrete, trending, and windowed aggregate reporting. At the same time, the dramatic acceleration effects of SortCL's big data transformation capabilities alongside ETL and BI tools were reported. In 2013, IRI is releasing third-generation versions of FieldShield, RowGen, and FACT, and a new NextForm package should tackle database migrations afterwards.
With CoSort and its adjunct tools and connectors, IRI has delivered both standalone solutions and seamless accelerators for big data integration, business intelligence, and data security. Please subscribe to the IRI newsletter for more details on the “innovative routines” available today, and those you can expect next.