Teradata continued its spending spree by acquiring the Mountain View, CA-based Hadoop consulting firm Think Big Analytics on Wednesday. The acquisition of Think Big Analytics will supplement Teradata’s own consulting practice. Think Big Data Analytics, which has roughly 100 employees, specializes in agile SDLC methodologies for Hadoop consulting engagements that typically last more than a month but less than a quarter of a year. According to Teradata Vice President of Product and Services Marketing Chris Twogood, Teradata has “now worked on enough projects that it’s been able to build reusable assets” as reported in PCWorld. Think Big Analytics will retain its branding and its management team will remain at the company’s Mountain View office. Teradata’s acquisition of Think Big Analytics comes roughly two months after its purchase of Revelytix and Hadapt. Revelytix provides a management framework for metadata on Hadoop whereas Hadapt’s technology empowers SQL developers to manipulate and analyze Hadoop-based data. Teradata’s third Big Data acquisition in less than two months comes at a moment when the Big Data space is exploding with a proliferation of vendors that differentially tackle the problem of data discovery, exploration, analysis and visualization with respect to Hadoop-based data. The question now is whether the industry will experience early market consolidation as evinced by startups snapped up by larger vendors or whether the innovation that startups provide will be able to survive a land grab in the Big Data space initiated by larger, well capitalized companies seeking to complement their Big Data portfolio with newly minted Big Data products and technologies. Terms of Teradata’s acquisition of Think Big Analytics were not disclosed.
Trifacta recently announced a deeper integration of its Data Transformation platform with Tableau, the leader in data visualization and business intelligence, as a key feature of the release of the Trifacta Data Transformation Platform 1.5. The Trifacta Data Transformation Platform 1.5 allows customers to export Trifacta data to a Tableau Data Extract format or register it with Hadoop’s HCatalog to facilitate the integration of Hadoop-based data from Trifacta into Tableau. Trifacta’s Chief Strategy Officer Joe Hellerstein remarked on the significance of the deeper integration with Tableau as follows:
Tableau creates huge opportunities for effectively analyzing data, but working with big data poses specific challenges. The most significant barriers come from structuring, distilling and automating the transfer of data from Hadoop. Our integration removes these barriers in a way that complements self-service data analysis. Now, Trifacta and Tableau users can move directly from big data in Hadoop to powerful, interactive visualizations.
Trifacta’s ability to output data to Tableau Data Extract format means that its customers can more seamlessly integrate Trifacta data with Tableau and reap the benefits of its renowned data visualization capabilities. The Trifacta Data Transformation platform specializes in enhancing analyst productivity in relation to Big Data sets by delivering a machine learning-based user interface that allows analysts to explore, transform, cleanse, visualize and manipulate massive data sets. Moreover, Trifacta’s predictive interaction technology iteratively learns from analyst behavior and offers users guided suggestions about productive paths for data discovery and exploration. The announcement of Trifacta’s deepened integration with Tableau means that Trifacta data which has experienced a process of transformation now encounters a streamlined segue to the Tableau platform. Meanwhile, the deepened partnership between the two vendors positions Tableau to consolidate its market positioning as the de facto business intelligence platform for Hadoop-based data.
Cloud Computing Today recently had the privilege of speaking with Dave McCrory, CTO of Basho Technologies, about the NoSQL space and Basho’s competitive differentiation within the NoSQL landscape. McCrory elaborated on Basho’s Riak “open source, distributed database” by noting its high availability, scalability and ability to handle any type of data as follows:
Cloud Computing Today: How do you envision the NoSQL space? What are your high level impressions of the competitive landscape amongst NoSQL vendors?
Dave McCrory (Basho Technologies): The NoSQL industry has many players for various use cases, but overall it is still young, especially from the enterprise point of view. I’ve been involved in big data for quite some time, and as data continues to grow, the NoSQL industry will grow with it. As the early adopters begin to move to the early majority – we are positioned in that space for crossing that chasm. Looking at how people want to build applications and data we will see, as an industry, in the next few years nearly half of enterprises will embrace NoSQL technologies to deal with the problems that traditional databases cannot deal with. Other NoSQL providers like MongoDB have an amazing presence in the market as it has made it easy for developers to give it a try. From my understanding from the market view, at the same time, it is limited in the actual applications that can be used. With so many companies offering NoSQL solutions for specific use cases and the high demand for data management, I can only see the industry continuing to expand and thrive.
Cloud Computing Today: Where do you see Basho within the larger NoSQL space at present?
Dave McCrory (Basho Technologies): We’re looking to provide the strongest key value solution and object store we can – that’s our priority right now. Although we at Basho are still a fairly young company, I think our technology speaks for itself. Since starting at Basho in the spring, I’ve been able to work with the outstanding Basho engineers and I’m amazed by what they have accomplished. Riak and Riak CS use simplified administrative features and a key/value system which enable anyone with command line experience to build a cluster in less than 15 minutes. I believe that Riak’s simplicity and usability are what separates it from other companies in the NoSQL space.
Some of that usability is our differentiation expressed in terms of high availability, fault tolerance and the ability to scale well beyond many of our competitors.
Cloud Computing Today: What are the key differentiators of Riak? What does Basho have planned for Riak in subsequent releases in the near future?
Dave McCrory (Basho Technologies): Riak’s key differentiators are its ability to offer high availability, massive scale and a variety of data types. Since Riak stores data as binary it is able to handle any type of data, unlike other solutions. Its top features include operational ease at large scales, always-on availability, and the ability to add and remove nodes easily and quickly as needed.
We are unique in that we have built object storage on our foundation and offer both key value and object store from the same platform. We have a thriving community, but our go to market in very focused on the enterprise. That has resulted in almost 200 enterprise customers including a third of the Fortune 50.
We have a lot planned for Basho and Riak in the coming months. We recently launched Riak CS 1.5 which offers additional Amazon S3 compatibility, performance improvement in garbage collection processes, and new, simplified administrative features. We are releasing Riak 2.0 in the fall which will provide enhanced search capability, expanded data types and more customer control over consistency, and we are hosting the annual RICON conference in Las Vegas in October, so you’ll be hearing a lot from Basho the rest of the year!
Google recently announced development of Mesa, a data warehousing platform designed to collect data for its internet advertising business. Mesa delivers a distributed data warehouse that can manage petabytes of data while delivering high availability, scalability and fault tolerance. Mesa is designed to update millions of rows per second, process billions of queries and retrieve trillions of rows per day to support Google’s gargantuan data needs for its flagship search and advertising business. Google elaborated on the company’s business need for a new data warehousing platform by commenting on its evolving data management needs as follows:
Google runs an extensive advertising platform across multiple channels that serves billions of advertisements (or ads) every day to users all over the globe. Detailed information associated with each served ad, such as the targeting criteria, number of impressions and clicks, etc. are recorded and processed in real time…Advertisers gain fine-grained insights into their advertising campaign performance by interacting with a sophisticated front-end service that issues online and on-demand queries to the underlying data store…The scale and business critical nature of this data result in unique technical and operational challenges for processing, storing and querying.
Google’s advertising platform depends upon real-time data that records updates about advertising impressions and clicks in the larger context of analytics about current and potential advertising campaigns. As such, the data model requires the ability to accommodate atomic updates to advertising components that cascade throughout an entire data repository, consistency and correctness of data across datacenters and over time, the ability to support continuous updates, low latency query performance, scalability as illustrated by the ability to support petabytes of data and data transformation functionality that accommodates changes to data schemas. Mesa utilizes Google products as follows:
Mesa leverages common Google infrastructure and services, such as Colossus, BigTable and MapReduce. To achieve storage scalability and availability, data is horizontally partitioned and replicated. Updates may be applied at granularity of a single table or across many tables. To achieve consistent and repeatable updates, the underlying data is multi-versioned. To achieve update scalability, data updates are batched, assigned a new version number and periodically incorporated into Mesa. To achieve update consistency across multiple data centers, Mesa uses a distributed synchronization protocol based on Paxos.
While Mesa takes advantage of technologies from Colossus, BigTable, MapReduce and Paxos, it delivers a degree of “atomicity” and consistency lacked by its counterparts. In addition, Mesa features “a novel version management system that batches updates to achieve acceptable latencies and high throughput for updates.” All told, Mesa constitutes a disruptive innovation in the Big Data space that extends the attributes of atomicity, consistency, high throughput, low latency and scalability on the scale of trillions of rows toward the end of a “petascale data warehouse.” While speculation proliferates about the possibilities for Google to append Mesa to its Google Compute Engine offering or otherwise open-source it, the key point worth noting is that Mesa represents a qualitative shift with respect to the ability of a Big Data platform to process petabytes of data that experiences real-time flux. Whereas the cloud space is accustomed to seeing Amazon Web Services usher in breathtaking innovation after innovation, time and time again, Mesa conversely underscores Google’s continuing leadership in the Big Data space. Expect to hear more details about Mesa at the Conference on Very Large Data Bases next month in Hangzhou, China.
Pivotal and Hortonworks will collaborate to accelerate development of Apache Ambari, the open source framework for provisioning, managing and monitoring Hadoop clusters. Pivotal will dedicate engineers toward advancing the “installation, configuration and management capabilities” of Apache Ambari as part of the larger project of contributing to software that promotes adoption of Apache Hadoop. In a blog post, Pivotal’s Jamie Buckley elaborated on the value of Apache Ambari to the Hadoop ecosystem as follows:
Apache Hadoop projects are central to our efforts to drive the most value for the enterprise. An open source, extensible and vendor neutral application to manage services in a standardized way benefits the entire ecosystem. It increases customer agility and reduces operational costs and can ultimately help drive Hadoop adoption.
Here, Buckley remarks on the way in which Ambari enhances the process of deploying and managing Hadoop by reducing costs and increasing the flexibility of customer choices regarding the operationalization of Hadoop. Meanwhile, Shaun Connolly, VP Strategy at Hortonworks, commented on the significance of Pivotal’s contribution to the Apache Ambari project as follows:
Pivotal has a strong record of contribution to open source and has proven their commitment with projects such as Cloud Foundry, Spring, Redis and more. Collaborating with Hortonworks and others in the Apache Hadoop ecosystem to further invest in Apache Ambari as the standard management tool for Hadoop will be quite powerful. Pivotal’s track record in open source overall and the breadth of skills they bring will go a long way towards helping enterprises be successful, faster, with Hadoop.
Connolly highlights Pivotal’s historical commitment to open source projects such as Cloud Foundry and its track record of success helping enterprises effectively utilize Apache Hadoop. Hortonworks stands to gain from Pivotal’s extraordinary engineering talent and reputation for swiftly releasing production-grade code for Big Data management and analytics applications. Meanwhile, Pivotal benefits from enriching an open source project that both vendors refer to in the context of a “standard” management tool for the Apache Hadoop ecosystem. The real winner, however, is Hortonworks, who now can claims the backing of Pivotal for the open source project Ambari incubated by some of its engineers, but also reaps the benefits of dedicated engineering staff from Pivotal that will almost certainly accelerate the rate of development of Ambari. The only qualification, here, is that Pivotal’s collaboration with Hortonworks is likely to ensure the optimization of Ambari for both the Pivotal HD and Hortonworks distribution, with the ancillary consequence that Ambari may be less suited for other Hadoop distributions such as Cloudera and MapR. Regardless, the collaboration between Hortonworks and Pivotal promises to serve as a huge coup for the Big Data industry at large both with respect to expediting development of Apache Ambari, and constituting a model for collaboration between competitors in the Big Data space that will ultimately enhance Hadoop adoption and effective utilization.
Databricks, the company founded by the team that developed Apache Spark, recently announced the finalization of $33M in Series B funding in a round led by New Enterprise Associates with existing participation from Andreessen Horowitz. The company also revealed plans for commercializing Apache Spark by means of the newly launched Databricks Cloud that simplifies the data pipeline for data storage, ETL processing and thereupon running analytics and data visualizations on cloud-based Big Data. Powered by Apache Spark, the Databricks Cloud leverages Spark’s array of capabilities for operating on Big Data such as its ability to operate on streaming data, perform graph processing, offer SQL on Hadoop as well as its machine learning functionality. The platform aims to deliver a streamlined data pipeline for ingesting, analyzing and visualizing Hadoop-based data in a way that dispels the need to utilize a combination of heterogeneous technologies. Databricks will initially offer the Databricks Cloud on Amazon Web Services but plans to expand its availability to other clouds in subsequent months.