Yesterday, ParAccel, provider of high performance analytic database, announced its partnership with NetApp, storage and data management vendor, to offer an integrated solution. This integration of ParAccel Analytic Database (PADB) with NetApp Unified Storage Architecture will offer highly boosted performance of Analytic Database at a much affordable cost. This integration comes as ParAccel joins the NetApp Alliance Partner Program.
State Of The Enterprise Database Market
In the enterprise database market dominated by the big three, Oracle, Microsoft and IBM, we are seeing many startups trying to disrupt their dominance using some innovative technologies in analytics, use of new kinds of architecture based on standard x86 systems, etc.. Even though the big three vendors have a stronghold in the marketplace, businesses are slowly getting frustrated with their high licensing fees. Especially, in this weak economic scenario, organizations are rethinking their belief on the value of big vendor database at high licensing costs. They are slowly realizing how they are made to pay these hefty licensing fees by vendor lock-in. They are not only looking for low cost offerings but also for more open systems without the lock-in.
In fact, the frustration with the big vendors was highlighted in the new Information Week research survey of 750 business technology professionals.
Meantime, our survey of more than 750 business technology professionals shows discontent with steep licensing and upgrade fees. Fifty-two percent of survey respondents characterize licensing costs for their primary databases as either somewhat overpriced or outright highway robbery, and they’re not targeting only Oracle. That resentment is nothing new, but it’s being exacerbated by workloads and data volumes that are multiplying at a staggering rate, sending costs even higher.
There is clearly an opportunity here and smaller vendors like ParAccel, Aster Data, Netezza, Teradata, etc. are attacking different segments of this market like Data Warehousing, Transaction Processing, etc..
ParAccel High Performance Analytic Database
ParAccel offers a leading edge RDBMS without the overheads associated with the traditional analytics database management systems. They use industry standard hardware (x86 systems) to offer high performance analytic database. Traditionally, in order to do high performance analytic queries, enterprises used datawarehousing tuning techniques like specialized physical schemas, aggregation tables, indexes, materialized views, etc. on traditional RDBMS like Oracle, SQL Server, MySQL, etc.. This caused significant overheads, including data redundancy, which caused heavy performance degradation.
With PADB, these are not needed and hence the elimination of unnecessary overheads. ParAccel relies on a more normalized schema than these specialized ones used by traditional RDBMS systems and, even, some of the newer analytic systems. This resulted in an extensible system with much higher efficiency. This flexibility allowed the database administrators to focus on providing newer applications than spending their time tweaking the configuration for performance gains, which usually resulted in incremental performance increases.
In order to offer high performance, PADB was architected to use
- Shared Nothing Massively Parallel Processing where the processing is distributed uniformly across several self contained nodes (standard x86 machines with its own CPU, Memory, I/O, Interconnect and which can be optionally attached to their own SAN storage devices) which benefitted from the raw power of a parallel configuration fabric that connects these nodes. Their use of ParAccel Interconnect Protocol (PIP) instead of TCP/IP over standard Gigabit Ethernet offered the high performance available in their system while also offering linear scalability
- Unlike in the case of transaction processing where row-wise databases were highly appropriate, ParAccel used columnar storage for their Analytic Database. Columnar storage is very much aligned with analytic queries because they are typically concerned with only a fraction of the columns defined in a table. More importantly, their columnar storage implementation is completely transparent to the applications and data could be retrieved via SQL as in the case of traditional RDBMS
- High-end query optimizer which took advantage of PADB architecture to handle complex queries and it is MPP, Columnar and compression aware so that the queries are handled seamlessly as in the case of row-wise databases
- PADB is SQL92 compliant and supports the most important analytic extensions in the SQL99 and SQL:2003 standards including analytic functions, window functions and correlated sub-queries, along with some features specific to PADB
- Usually, the usefulness of SAN in MPP environment is limited by their finite channel throughput. But ParAccel’s patent pending technology helps overcome these disadvantages allowing the nodes to take advantage of speed and scalability of Direct Attached Storage devices
In benchmark tests, PADB running on VMware’s vSphere achieved a 7.7x better performance over the previous performance record holder using 37% fewer servers. This combo also had a 8.7x speed over the previous record holder. Even though these numbers can be reproduced under ideal conditions, it is evident that PADB offers a much higher performance for a much lower cost.
Today’s announcement meant that customers can use PADB on NetApp’s Unified Storage Architecture getting much higher performance on their analytic database with low infrastructure costs. Along with the high performance available through NetApp SAN or NAS, customers can also take advantage of NetApp’s advanced data management solutions like advanced backup, recovery, restore, etc..
Clearly, the market is ready for newer technologies from smaller vendors. Companies like ParAccel are well positioned to take advantage of these needs. It is vital for these smaller companies to forge right kind of partnership to take on the likes of Oracle and IBM. ParAccel’s partnership with NetApp can be seen in this light. However, the big vendors are not taking it easy. They are not only trying to price aggressively but also acquire newer technologies to match these smaller vendors. Interesting days are ahead in the enterprise database market.
Related articles by Zemanta
- ParAccel and NetApp Extend Benefits of the World’s Fastest Analytic Database to Storage Area Network Customers (eon.businesswire.com)
- EMC Buys Greenplum – Big Data Realignment Continues (enterpriseirregulars.com)
- Ceridian Stored Value Solutions Taps ParAccel to Deliver Rapid Insight into Consumer Preferences (eon.businesswire.com)
- IBM Acquires Netezza – ADBMS Consolidation Heats Up (enterpriseirregulars.com)
- Some thoughts on the announcement that IBM is buying Netezza (dbms2.com)
- Research: State Of Enterprise Databases (informationweek.com)