Ancelus in-memory database achieves certified one billion transactions per minute performance mark, efficiently handles “Big Data” challenges

Share Article

The new “Big Data” generation Ancelus database: capable of one billion transactions per minute; increased database processing speed of 5-7 orders of magnitude; real-time updating capacity; and database size – “Big Data” or smaller – is irrelevant to performance. “This Ancelus database system possesses the true transformational capacity to dramatically change how software applications are developed and implemented,” said John Layden of the Ancelus design team. “Ancelus was specifically designed to overcome current industry-wide limitations of scalability and ‘Big Data’ growth.”

The power of Ancelus

Ancelus, an in-memory advanced high-speed database, has achieved certification of its capacity to exceed one-billion transactions per minute, officially dwarfing the computing capacity of its nearest rival. According to John Layden, CEO of Time Compression Strategies (TCS), the certification was achieved using randomized data in the challenging three-table-join benchmark, with a billion-row join table.

The three-table-join represents one of the most common design challenges for application developers. This function is a special problem for traditional databases, including typical in-memory systems.

The independent certification confirms critical industry transformational elements: that Ancelus achieves processing speed improvements of 5-7 orders of magnitude; that Ancelus database systems can efficiently address “Big Data” challenges on 64-bit readily available commercial systems; that Ancelus performance metrics are essentially independent of data sizes or challenges; and that it is possible for “Big Data” databases to be updated “live” and to strategically manipulate or decide on table structure “on the fly.”

“We picked one of the toughest challenges, and the Ancelus database is now independently certified as being extremely fast with numerous other strategic performance benefits,” said Layden. “It addresses virtually all of the complex reliability and performance concerns about ‘Big Data’ in today’s business environment, and all on commercially available platforms.”

Running on 64- bit hardware platforms available to any business, the Ancelus database specifically outperforms in the areas of speed, reliability, real-time revisions, scalability and agility. “Revolutionary is often an over-used word, but in this case the Ancelus database possesses the capacity to actually change how application software is designed and developed. Previously intractable problems are now within easy reach,” Layden said.

Through unique and patented conceptual methods that took almost a decade to produce, the Ancelus database uses a direct representation of stored data and all relationships rather than the tradition process of being “flattened” into two-dimensional tables that store only data. “The Ancelus database was specifically designed to overcome current industry-wide limitations of scalability and ‘Big Data’ growth,” explained Layden. “Ancelus eliminates inefficient chained-together databases, exotic hardware, or ‘big iron’ installations.”

The certified benchmark tests, conducted on a 64-bit six-core microprocessor readily available to any business, produced throughput transactions per minute (TPM) of 1.3 billion (at 100 million rows), 1.2 billion (at 200 million rows) and 1.1 billion (at one billion rows).

The 2012 certification tests represent the first published benchmark for Ancelus since U.S. patents were awarded for this new concept in database design, according to Layden.
TCS contracted with the Purdue Enterprise Company, LLC in West Lafayette, Indiana to secure the independent certification. Vijay Pai, PhD, of the Purdue University Electrical and Computer Engineering Department, served as technical consultant and advisor in the test on benchmark software design and measurement methods. Dr. Pai independently audited the test software, monitored the execution of the benchmarks, audited the stored and retrieved results, and confirmed the recording of results.

Becoming known as the “database that never sleeps,” Ancelus deployments are expected to save operational costs through continuous high processing speed on smaller servers. All maintenance, admin and schema revision functions execute while the database is live. Data can be “pipelined” into and out of the database without downtime. Re-indexing occurs on every transaction and is included in the benchmark measures.

Ancelus deployment costs are manageable from the perspective that the database can be installed and operated either as a stand-alone high-speed system or as a “front-end” in tandem with existing relational database installations, according to Layden. Replicated in-memory versions and the streamed disk store assure security and integrity of data, even during power or hardware failures.

Ancelus also enhances developer productivity, a critical success factor in application development. Ancelus eliminates multiple layers of complexity with existing systems and makes the database essentially transparent to either human users or applications, regardless of size or complexity.

“Scalability and normalization are the most common failure modes in large relational database systems,” said Layden. “An Ancelus deployment significantly reduces development cost and banishes both scalability and normalization failures.”

In addition to high-speed performance, the patented Ancelus architecture delivers several competitively unique features:

  •     Fast: Ancelus get/put performance is a constant, regardless of the size of the database

o    Latency of 200 nanoseconds, single thread throughput of 300 million transactions per minute (TPM)

  •     Non-stop: All database schema revisions, admin, backup and data purging are performed in a live database

o    Less than eight microseconds interrupted access to add a column to a one billion row table
o    Pipelining data for uninterrupted operation
o    Eliminates planned and unplanned downtime; the practical limit for the “Big Data” world of massive data sets
o    Extreme reliability through simplified hardware and application software
o    Dramatic reduction in disk hardware eliminates the primary failure mode
o    Reduced layers of software eliminate complexity failures

  •     Big: Scalable to Exabyte size without loss of performance

o    Massively parallel access to the database is accomplished via multiple sockets; a unique locking strategy allows this creative option and eliminates the need for racks of disk drives
o    Cloning sockets is free; cloning disk drives isn’t

  •     Lean: 80% database compression is typical with Ancelus

o    No duplication of data, no empty cells; variable width fields
o    True zero “sparcity”

  •     Agile: Query paths are stored in the data directory with live edit

o    Planned downtime becomes a thing of the past
o    Fast prototyping and direct query paths eliminate SQL ambiguity

The Ancelus design team first developed the foundational 64- bit technology between 2003 and 2007. It was first offered for commercial use following four years of beta-testing in a manufacturing environment.

ABOUT ANCELUS – The Ancelus design team of first developed and delivered real-time systems in 1983. In 1986 the design team produced its first in-memory database, operating in a 16-bit environment. A second generation 32-bit version was patented in 1993, and the Ancelus database patents were secured in 2010 and 2011. The new Ancelus database is either currently in use or being deployed by a variety of companies and manufacturing operations. Running on 64- bit hardware platforms available to any business, the Ancelus database specifically outperforms in the areas of speed, reliability, real-time revisions, scalability and agility. It is available through Time Compression Strategies. For more information, please visit: http://www.ancelus.com or http://www.tcsdb.com.

Share article on social media or email:

View article via:

Pdf Print

Contact Author

John Layden
Ancelus
(317) 709-2161
Email >

Michael Snyder
Visit website