Explore Open-Source Databases on Linux on POWER
Open-source software has transformed the IT landscape, now serving as a core component for production of next-generation cloud- and analytics-based growth solutions.
By Chuck Calio02/06/2017
Open-source software has transformed the IT landscape, now serving as a core component for clients to produce the next generation of cloud- and analytics-based growth solutions.
Organizations have explored open-source software for three main reasons:
- Cost. The less a company spends on proprietary software, the more it can dedicate to other facets of digital transformation efforts.
- Open-source databases can be either an alternative to proprietary, commercial databases or a base for brand-new applications that handle specific types of unstructured or semistructured big data
- Open-source databases allow companies to derive benefit not only from their own IT employees, or from the employees of proprietary software vendors, but also from the open-source community that supports and grows the core product and ecosystem around these databases
Users tend to group databases into two categories:
- Relational: those that are optimized to store and retrieve structured data
- NoSQL: those that are optimized to support unstructured data
Most data produced and analyzed today is unstructured, but no single database fits everyone’s needs. Users often take advantage of data stored in multiple databases in a variety of new applications. Enterprise-wide applications are being enhanced to support access to unstructured data and leveraging NoSQL databases.
For anyone interested in exploring open-source database solutions on Linux* on IBM POWER8* technology, MariaDB, EnterpriseDB, Redis, MongoDB and Neo4j are all good starting points. See Figure 1 below for more information (click to view larger).
Commercial Relational Databases
Commercial relational databases serve as the backbone of transactional systems, and continue to be enhanced with new functionality and new delivery models such as cloud computing and database as a service. When evaluating any new database, keep in mind that classic transaction atomicity, consistency, isolation and durability (ACID) properties remain a strength in the domain of relational databases.
Open-source relational databases include both MariaDB and PostgreSQL from EnterpriseDB. MariaDB, produced by the original developers of MySQL, is popular in modern, web-based architectures, and has benefited from the support of huge players in the IT industry, including IBM and Google. PostgreSQL is a powerful, open-source object-relational database system with over 15 years of active development and a proven architecture with a reputation for reliability, data integrity and correctness. PostgreSQL is fully ACID-compliant and offers native programming interfaces for C/C++, Java*, .Net, Perl, Python, Ruby, Tcl and ODBC languages.
End users want modern, multifunction applications that store, retrieve and discover value in new, unstructured or multistructured data. Agile developers and data scientists must quickly create new applications that leverage a variety of structured and unstructured data types—often across more than one database. Storing and managing vast amounts of unstructured data often relies on more specialized database technologies, generically called NoSQL databases. Also called nonrelational or post-relational databases, they provide a way to store and retrieve data without the tabular method used by relational databases. NoSQL databases often don’t require a fixed schema, and tend to be better suited for storing and retrieving big data. They’re adaptable to changes in data structures, and place the burden of data consistency and integrity on developers, rather than requiring the database itself to have a fixed schema. Many NoSQL database users run SQL queries, so the SQL language can be a bridge between the relational and NoSQL worlds.
Redis, MongoDB and Neo4j can be optimized on Linux on POWER*. Redis is a data structure store often used by application programmers as an in-memory cache. MongoDB is a leading document store. Neo4j is a graph database optimized for identifying and exploring connections between data. According to IBM, all three databases experience tremendous performance, scaling and price/performance advantages when running on Linux on POWER (ibm.co/1DP1XG0). Redis and Neo4j also exploit IBM POWER8 Coherent Accelerator Processor Interface flash technology for enhanced performance results.
How POWER8 Fits in the Picture
The POWER8 architecture was designed for modern big data platforms, delivering 4x more processor case, memory bandwidth and multithreading than commodity hardware platforms can provide, according to IBM internal measurements.
While some relational database management systems can be costly, the benefits and qualities of service inherited by commercial relational databases running on Power Systems* technology also apply to open-source relational and NoSQL databases on Linux on POWER. New advances in Power Systems scale-out processors, lower price points realized from Linux and open-source software, and vast amounts of innovation infused from the OpenPOWER Foundation members give open-source databases on Power Systems technology an industry-leading value proposition.
Take a Closer Look
Versions of these databases are available at no charge for download on Linux on POWER. Fully supported and enhanced enterprise-ready versions are available from commercial suppliers such as EnterpriseDB, MariaDB, MongoDB, Redis Labs or Neo4j. IBM partner companies—such as Avnet, through its RapidBuild offering—provide a completely integrated set of open-source database solutions on Linux on POWER.
Data takes organizations over the goal line, and POWER8 technology is optimized for handling and processing data.
Chuck Calio is an ISV and open-source solutions enablement architect, IBM Power Systems.
Post a Comment
Note: Comments are moderated and will not appear until approvedcomments powered by Disqus