PostgreSQL (“Postgres”) ranks as the 4th most popular database according to DB-Engines, and excels as a candidate for most OLTP apps and as a hub for data federation.  Its weakness in Massively Parallel Processing (MPP), however, often keeps it out of the running for advanced analytics and high-end OLTP use cases.

The MPP limitation was addressed earlier this month when TransLattice announced it would open source some of the Postgres-derived StormDB code, which TransLattice acquired in the autumn of 2013, creating a new Postgres branch called Postgres-XL (“XL” stands for eXtensible Lattice). Curiosity will compel plenty of the already well-established open source PostgreSQL community to give Postgres-XL a peek.  If they like what they see they may contribute, and eventually turn to TransLattice for support and service.  The world is certainly big enough for two commercialized Postgres choices fostered by EnterpriseDB and TransLattice.

Why would TransLattice take such a step?  First, it primarily acquired StormDB to enhance its position as a bare metal Database-as-a-Service provider.   With TED – the TransLattice Elastic Database which ranks as one of the very few truly geographically distributed databases – already available via Amazon Web Services and appliance options, StormDB provides a private cloud companion database for TED. In addition to StormDB giving TransLattice market coverage, it also brings a related but fresh set of engineering and design to the TransLattice IP fold.  Furthermore, giving developers access to some of StormDB’s underlying code could potentially foster a new community, and eventually customers, for TransLattice.  Who might care?

The Postgres-XL Value Proposition

Enterprise-class developers, data architects, and database administrator might care about Postgres-XL, and here are 5 reasons why.  The 4 four reasons apply whether you work with open source PostgreSQL or Postgres-XL, and the fifth applies to Postgres-XL only:

  • Enterprise-class:  First, because since it is based on PostgreSQL, Postgres-XL already includes serious enterprise-class characteristics that typically you would find in the most well-established commercial brands like Oracle, IBM DB2, and SAP Sybase.  The larger Postgres community and EnterpriseDB have seen to it that in areas like data protection, disaster recovery, and security Postgres can stand shoulder-to-shoulder with the big thoroughbred commercial databases.
  • Dev, Data Access, and APIs:  Second, Postgres supports a wide range of development tools, data access methods, and important APIs like REST.
  • SQL:  Third, Postgres supports ANSI-SQL, meaning IT’s investment in SQL expertise retains its value, and thus it is relatively easy to port applications to or build new apps for Postgres.
  • Dependable and Fairly Priced:  Fourth, despite all the marketing hype around the inaccurately named “NoSQL database” community, Postgres is a proven and trusted database solution, and has enterprise momentum.  Many of the NoSQL databases are playing catch up to what Postgres has delivered in production for years.  If you are looking for less expensive alternative commercial options in the enterprise-class database field, with a zero to low cost appetizer, Postgres and MarkLogic are where you should look first.  Consider, for example, that the 4th largest ERP vendor, Infor, recently certified on Postgres Plus, the commercial Postgres version from EnterpriseDB.
  • Big Data, Fancy Data Types, Fancy Infrastructure:  Fifth, for those interested in Big Data, but fearful of snake oil and the high cost of services required to make Hadoop work, and not enamored with learning new query languages, MPP Postgres, aka Postgres-XL, offers an enticing alternative.  Postgres-XL is also scale-out ACID RDBMS, acts as a key value store, supports multi-tenancy/as-a-service, and supports abstract data types and specialty data types like geospatial.  Of particular interest to those interested in scale:  you can choose to either partition individual tables across nodes, or replicate them. Typically it makes sense to replicate fairly static tables, while partitioning large write-heavy tables if the workload is OLTP or large fact tables for analytics.

EMA View:  The Market Impact of Postgres-XL is TBD

I have written previously that PostgreSQL is the quiet giant of enterprise databases (see the bottom of this page).  PostgreSQL, however, had been to some degree missing out on some of the market interest in database-as-a-service, big data, and scale out.  Postgres-XL aims to remedy that situation.  Given the high degree of professionalism and commitment of the Postgres community, if Postgres-XL tests nicely, and the XL open source variant doesn’t stray far from the parent PostgreSQL open source, XL should give the entire community a boost.

The one hurdle is lack of hype:  Hadoop and NoSQL databases have been making noise for several years now, and the open question is will the quiet giant remain too quiet?  Postgres, whether the PostgreSQL original or the Postgres-XL MPP version, needs to make much more noise, and more loudly position itself as a Hadoop and NoSQL alternative that meets enterprise-grade requirements with ease.  If XL also means eXtremely Loud, Postgres-XL will emerge as the dark horse of the big data database sweepstakes.