We are living in the age of polyglot persistence, which really just means that it makes sense to store data using the technology that best matches the way the data will be used by applications. The age of trying to force everything into a relational DBMS is over, and we now have NoSQL, NewSQL, in-memory, and Hadoop-based offerings that are being used to store data. But you really should also be looking at the algorithmic approach offered by Ancelus Database.
Ancelus has achieved incredible performance benchmark results that are orders of magnitude faster than comparable relational database management system (RDBMS) results. The results show constant read/write latency under 100 nanoseconds regardless of size (up to 16 exabytes). Ancelus can run a billion-row search in 2.4 microseconds; and it adds only about 500 nanoseconds for each 10x increase in database size. Basically, queries that run for hours or days on typical database implementations can run in microseconds on Ancelus.
What powers this capability under the hood? The Ancelus algorithmic database uses a patented approach to data storage that replaces predefined storage structures with an algorithm that determines the physical storage location of data. This algorithmic approach decides where to put data based on efficiency of access to it. This means that physical data storage and retrieval are independent of any predefined data layout.
Ancelus couples this storage approach with in-memory data access techniques which are also based on patented memory management methods. By combining algorithmic data storage and in-memory data management, Ancelus has achieved constant performance even for very large databases and complex applications.
There is no duplication of data in an Ancelus database. Each data element is stored only once and data is linked using pointers to its data dictionary. A consequence of this design is that referential integrity is not optional—it is designed right into the database … and it does not impede performance.
Nested and recursive lists are used by Ancelus to directly implement an entity-relationship diagram (ERD). The ERD is loaded directly to Ancelus without requiring it to be transformed into another format such as DDL for tables. Of course, you can design from tables if you so choose. And there is no need for denormalization with Ancelus as the data can be accessed efficiently and effectively at full speed regardless of how the “tables” look.
With this architecture, users can implement multiple logical schemas over the data at the same time, to access the data as it is in relational tables. Even though it is not. The actual physical schema is not exposed to users.
Ancelus is ACID-compliant and works efficiently for both transactions and analytics, supporting HTAP (Hybrid Transaction Analytical Processing) and translytical environments. Ancelus is not a SQL database system; data is typically accessed by a native API. However, Ancelus does offer TQL (Threaded Query Language), which converts Ancelus schema structures into SQL-readable formats for ease of integration. You can also use JDBC and PHP to access Ancelus.
But what about administration? It is possible to implement a true, nonstop operational environment with Ancelus. It requires no downtime to make schema changes, re-index, or compress, refresh, and archive data. And remember, there is no need to create separate, denormalized instances for analytics, so design time is freed up, and there is no ETL because the data is all in one place.
From a development perspective, coding is simpler because programmers do not need to worry about deadlocks. Ancelus uses a multi-lock function that prevents the primary causes of deadlocks, blocks most race conditions, and clears stale locks caused by abnormal exits.
And, the size of your databases will not expand out of control due to internal storage issues. Ancelus does not duplicate data nor does it implement sparse matrices, so dataset size can be up to 20x smaller than for traditional systems.
Ancelus can be run on Linux, UNIX, and Windows platforms.
Finally, a word about the company. Although you may not have heard of Ancelus before, the company that developed it, Time Compression Strategies Corporation (TCSC), has been in business since the early 1980s, so there is a history of innovation and development behind Ancelus.
The Bottom Line
The bottom line is that you may be able to reduce cost, speed up development, and improve operations using Ancelus. Indeed, some applications that may have been perceived as impractical or too costly may become achievable with Ancelus.