Actian Unleashes SQL Users on Hadoop with Easy Access, Faster Performance

Actian’s approach to big data aims to break down barriers that often frustrate companies trying to wring more value and performance from Hadoop.  Actian Analytics Platform - Hadoop SQL Edition throws the doors wide open for millions of SQL users, offering easier access, faster performance and more. IDN speaks with Actian CTO Mike Hoskins

Tags: Actian, analytics, BI, big data, data, Hadoop, Hive, query, SQL, structured, unstructured, YARN,

Mike Hoskins

"Today’s [Hadoop] performance is still completely unacceptable to millions of SQL programmers out there. We want to unleash SQL users on Hadoop."

Actian’s latest update to its Actian Analytics Platform looks to overcome many of the barriers that often frustrate companies trying to wring more value and performance from Hadoop.


Big Data in Motion Summit
Manage Expanding Data Volumes for Analytics & Operations
Jan 29, 2015

Actian Analytic Platform – Hadoop SQL Edition throws the doors wide open for millions of SQL users, offering easier access, faster performance and more, by bringing together a set of rich technologies to support native SQL in Hadoop, Actian CTO Mike Hoskins told IDN.


The gravitational force of Hadoop is increasing. Companies are dealing with data in scale and complexity they have never seen before, and they are going to Hadoop-type infrastructure. Old architecture is not going to be used anymore,” Hoskins said. “So, we want to unleash SQL users on Hadoop with easier to use and high-performance design time and runtime [features].


Hoskins listed several key improvements Acian focused on with this release, including:

  • Fixes for slow query speeds (or even incomplete query functionality) in Hive or Impala;
  • Complete ability to update or delete data; and
  • Bringing more reliability to how ODBC/J DBC drivers connect with third-party BI or data visualization tools.

Actian Analytics Platform - Hadoop SQL Edition also eliminates the need for specialized skills, such as the need to learn MapReduce, Hoskins added.


In the big picture, Hoskins is most proud of Actian’s efforts to deliver what he called “SQL completeness” to Hadoop. This is where all of SQL is fully and natively supported – without compromise or trade-offs – within a HDFS (Hadoop Distributed File System) cluster, he said.


“Quite frankly, we saw the ‘State of SQL’ was under-performing, and so we took those challenges on,” Hoskins said. “Now, we believe we’re opening the doors wide for millions of SQL developers and users.”


Actian’s Jim Hare, senior director for product marketing, described the under-the-covers Actian architecture this way:

SQL users want ‘industrialized’ SQL, which means they want mature enterprise-ready solutions that are fast, secure, reliable and can scale for all of their users and applications. Plus, they want to be able to use the full SQL language without constraints. And, they want to run SQL on the freshest Hadoop data, which means they don’t want to have to move data out of Hadoop into another database each time they want to run a query.

We refer to [Actian’s approach] as ‘SQL in Hadoop.’ It’s mature SQL and fully integrated into Hadoop. Actian has taken the world’s fastest TPC-benchmarked analytic database (Actian Vector) and extended it to run inside Hadoop natively via YARN. Industrialized SQL in Hadoop is married with an analytics and data science workbench enabling users to discover nuggets of valuable data hidden in Hadoop and make it accessible to all their SQL users. In essence, Actian has transformed Hadoop from a low cost data storage lake into a high performance analytics platform.


By bringing “industrialized SQL” technologies together in a Hadoop-friendly way, you also get super-fast performance, Hoskins noted.

By optimizing the X100 vector processing engine (patent pending), it now runs across every node of the massively scalable Hadoop cluster architecture, Hoskins said. “This delivers an unimaginable performance increase, especially for users used to using Hive or Imapala,” Hoskins told us. Benchmarks note up to a 30-times speed increase. Actian’s high-performance data flow engine also speeds data blending, discovery, and predictive modeling with an embedded high performance data flow engine.


Other barrier-breaking features of Actian Analytics Platform – Hadoop SQL Edition are:


Data science workbench: Allows users to visually blend and analyze Hadoop data – without coding.


Libraries of analytics: Functions to make the entire analytic process more consumable and easier to manage.


Parallelized pipelines: These will help users take the best advantage of the Hadoop cluster capacity data flow framework.


Pre-built visual data analytics operators: Users can visually build end-to-end analytics workflows or data bindings in minutes with more than 1500 pre-built “drag and drop” data analytics operators. Users can spend less time preparing and enriching data.


Security & Compliance: Delivers industrial-strength DBMS security, encompassing users, roles, and authentication with ACID-compliant reliability


YARN-Ready: Support for end-to-end native analytic processing for data blending, enrichment, analytic computation and operational BI.


We asked Hoskins what it has taken the Hadoop community to come up with solutions like these that could broaden the reach of Hadoop and deliver even more business impact. His response talks not just to the need to boost Hadoop technology, but to widen the circle of technologies Hadoop should leverage.


“In most cases, we have not been asking the right people. Until now, enterprise SQL users have not even been in the [Hadoop] game. It’s been a bit closed to the Hadoop programmers. And while they are just thrilled that they’ve made Hive faster in the last three months, the truth is that even with that jump the level of today’s performance is still completely unacceptable to millions of SQL programmers out there,” he told IDN.


“With our Hadoop SQL Edition, we are moving Hadoop from being a low-value data lake, to a high-value high-performance analytics platform. It will be transformative for big data,” he added.


At least one noted Hadoop analyst firm sees the point.


A note from a Forrester Research “Quick Take” report reads in part: “Hadoop has largely been the domain of developers and data scientists with programming skills. [B]ut high performing, compliant-SQL-on-Hadoop makes it interesting and accessible to a large community of business intelligence professionals and other SQL-jockeys in enterprises.”


Actian Analytics Platform – Hadoop SQL Edition is shipping now.