Our InfiniDB Enterprise column database has been built to handle the demanding workloads generated in business intelligence, data warehousing and big data analytics environments. With a linearly scalable column-oriented architecture at its core, InfiniDB enables organizations to leverage the best database the industry has to offer for keeping up with the pace of business.
InfiniDB Enterprise has been built to handle the demanding workloads generated in business intelligence, data warehousing and big data analytics environments.
- Column-oriented architecture: The InfiniDB columnar database is architected in a column-oriented manner vs. a legacy row-based format. This allows InfiniDB to service queries that need to read anywhere from a few hundred GB to many TB’s of information faster than row-oriented databases.
MPP distributed database operations
- MPP distributed database operations: InfiniDB is multi-threaded and can make use of today’s modern hardware that is multi-CPU/core based. More CPU’s and/or cores allow InfiniDB to increase performance of big data analytics with no application modifications being necessary. Further, deploying InfiniDB across multiple commodity servers as an MPP deployment delivers linearly scalable performance capable of adjusting to meet your user, data and performance demands.
High speed data loader
- High speed data loader: InfiniDB supports traditional insert/update/delete operations, but to load massive amounts of big data for business intelligence, a high-speed load utility is made available. InfiniDB’s loader allows TB’s of data to be loaded much more quickly into an InfiniDB column-oriented database than other traditional methods.
- Automatic partitioning: InfiniDB transparently uses vertical partitioning to store data and only read the needed columns to satisfy a query. InfiniDB also uses a form of logical horizontal range partitioning that does not require special storage placements or design. Using both vertical and logical horizontal range partitioning allows InfiniDB to reduce I/O in both directions (column and row). For many organizations, some data value lessens over time. Both vertical and horizontal partitioning are automatically handled by the InfiniDB columnar database and require no user intervention. Automatic partition drop within InfiniDB enables ultra fast removal of data partition boundaries and associated data no longer required.
Data compression with real-time decompression
- Data compression with real-time decompression: InfiniDB uses column-level compression to physically compress data on disk, as well as real-time decompression functionality to uncompress data. Columnar compression offers greater compressibility than row-based capabilities because of the similarity of data values within the column file. Real-time decompression means that data can be decompressed while it is being read from disk, and without the loss of read performance. InfiniDB combines the spacing savings derived from physical compression and elimination of indexing and materialized views. The benefit is increased query response when reading from disk, smaller disk footprint, and improved disk bandwidth and throughput.
User-Defined Function (UDF) for in-database analytics
- User-Defined Function (UDF) for in-database analytics: InfiniDB’s UDF capability enables data owners the flexibility to write business logic and analytic functions specific to their business needs. UDF’s provide a mechanism for extending the functionality within the database server (in-database) by embedding logic that can be evaluated in SQL statements. With InfiniDB’s distributed MPP database architecture, UDFs are fully parallelized and scalable, as is InfiniDB’s out-of-the-box SQL. The resulting benefit for in-database calculations is improved latency and high-throughput for deep operational and predictive analytics.
- High concurrency: More and more users are demanding access to the corporate data asset. Within InfiniDB, scalability is assured to meet the growing user workloads for big data analytics and business intelligence.
- Transactional support: ACID-compliant transactional (insert, update, delete) support is provided in the InfiniDB columnar database. Transactions can easily be committed or rolled back, and deadlock detection support is also provided to handle conflict resolution.
- Crash recovery: InfiniDB provides for full crash recovery capabilities. In the event of a system crash, InfiniDB automatically maintains data integrity and upon system restart, InfiniDB supports roll forward and back operations to return the database to a consistent state.
- MVCC design: InfiniDB supports multi-version concurrency control (MVCC) or “snapshot read” so query operations are never blocked on a table; a query will always see the big data as it existed at the time the query was issued.
No need for indexing
- No need for indexing: Because of InfiniDB’s transparent use of both vertical and logical horizontal partitioning, there is no need for indexing. In essence, the data in a columnar database is the index. In addition, InfiniDB also automatically maintains a small, important structure called the Extent Map, which is used to reduce I/O. The Extent Map also removes the need for any manual data partitioning.
- Low maintenance: In addition to removing the need to index tables, InfiniDB also doesn’t require other objects such as materialized views and summary tables to achieve fast performance. This helps remove the need for typical database administration maintenance functions as well as cuts down on the complexity of the overall system ensuring the capability of rapid implementations for InfiniDB.
- Performance diagnostics: To help tune performance, InfiniDB supplies monitoring and diagnostic utilities that help users monitor their database and troubleshoot poorly running SQL.
MySQL front end
- MySQL front end: InfiniDB utilizes MySQL for its front end. This allows anyone familiar with MySQL to become immediately productive with InfiniDB. For those not acquainted with MySQL, the learning curve is minimal as MySQL supports almost all ANSI standard SQL operations. Moreover, there are many freely supplied GUI tools as well from other vendors that may be used to develop against and administer an InfiniDB columnar database.