Category

Blog

ODPi Takes Big Data Day LA

By | Blog

By Roman Shaposhnik

Earlier this month, I attended Big Data Day LA – a vibrant community gathering of data and technology enthusiasts in sunny Los Angeles. Located on the USC campus, the 5th annual event was organized by local Big Data user groups and volunteers!

Unlike many of the big data industry’s events, Big Data Day LA wasn’t a company-owned conference and registration was fully covered by data-driven sponsors, like Hortonworks, Disney Interactive, WANdisco and more – making the event free for anyone to attend. As such, it wasn’t surprising that it attracted such a big crowd, with more than 1,500 people in attendance!

During the conference, I presented “Big Data on The Rise: Views of Emerging Trends & Predictions from real life end-users” where I offered the audience an overview of key trends emerging in 2017 within the Hadoop and Big Data ecosystem. My session also covered data from the ODPi End User Advisory Board (TAB) and real end-user perspectives on how companies are using Big Data tools, challenges they face and where they are looking to focus investments. My talk was well received by those in attendance and quite a few people approached me following the session to discuss their new understanding of ODPi and how it relates to traditional vendors, the Apache Software Foundation, the Linux Foundation and the enterprise.

The remainder of the conference featured a great selection of talks, especially for data scientists and software developers and, unsurprisingly, the entertainment industry track was a huge hit – featuring talks from Netflix, Warner Brothers, Guitar Center and more.

And, of course, being a Silicon Valley guy I simply had to check out all the startup buzz at the conference as well. Not only did the conference feature an awesome Startup Showcase track, but there were also quite a few presentations pushing the envelope on state-of-the-art machine learning. Just to give you a quick taste, I suggest you go to http://novamente.ai/ and check out the truly SciFi projects these guys are tackling. Their presentation on how to apply AI to producing ever higher grossing movies (think scripts, casting, visual effects and more) had the audience on the edge of our seats.  

A few other sessions I particularly enjoyed include one from Bain & Company where they tried to put big data and machine learning in the context of organizational shifts required in any traditional enterprise, in order to realize full value from big data insight. On the flip side, even if you have your organization all lined up for digital transformation, you still have to be mindful of challenges on the technical side of machine learning. The notion of a hidden, growing technical debt in these complex, end-to-end machine learning pipelines is something that we all should keep in mind and Irina Kukuyeva’s presentation did a great job of highlighting some of the same important areas.

Overall, Big Data Day LA was a fun and dynamic event that harnessed the upstream community and showcased the importance data has on each level of business.

 

 

Making Production Hadoop Faster, Easier and More Productive: ODPi Member Conversation with Zettaset

By | Blog

In our fifth ODPi Member Conversation podcast, John Mertic spoke with Zettaset’s Director of Product Management, Sesh Ramaswami, and CMO, John Armstrong.

Their multi-perspective discourse covered off on the security challenges of deploying Hadoop in production, data governance and security expectations, along with how nonprofit organizations like OASIS and ODPi can help drive standards for the big data ecosystem.

Touching on big data security within production environments, Ramaswami dove right into enterprise expectations around data governance and data processing.

Referring to the role data governance plays in the Hadoop ecosystem, Armstrong noted that security wasn’t always a concern, as data was in isolation away from production.

However, the space has changed drastically over the years. Touching on this shift, Armstrong said, “When [enterprises] finally want to move into production, and want to take those data environments and put them into the mainstream… we need to have some real security in place before you plug into the rest of the network —  because Hadoop coexists in large enterprise environments in conjunction with relational database environments, maybe other NoSQL, so it’s not alone and it has to play well with others.”

To hear more of the group’s insight, tune in to the episode on our YouTube channel!

Subscribe to our YouTube channel and follow us on Twitter to catch upcoming episodes of the ODPi Member Conversation podcast series!

Making Production Hadoop Faster, Easier and More Productive: ODPi Member Conversation with IBM

By | Blog

In our fourth ODPi Member Conversation podcast, John Mertic met with IBM’s Worldwide Analytics Architect Leader for Data Lakes, Neil Stokes.

Their insight-rich conversation kicked off with a discussion around machine learning and cognitive vision, along with how an ODPi Compliant Hadoop environment relates to cognitive computing.

Stokes, who has spent more than 20 years working at IBM, has a unique perspective on how Hadoop and cognitive systems afford enterprises the ability to ingest and make sense of the rawest possible forms of data, and why systems are only as good as the corpus of data to which it has access.

Referring to these sets of broad, invaluable data, Stokes noted “Interoperability is key here… Groups like ODPi – who have the ability to facilitate that level of interoperability across vendors, across platforms, across different technologies – there is a tremendous value that groups like ODPi bring [to the ecosystem].”

To hear more of Stokes and John’s discussion, tune into the episode on our YouTube channel!

Subscribe to our YouTube channel and follow us on Twitter to catch upcoming episodes of the ODPi Member Conversation podcast series!

Twitter Poll Results: Is Apache Hadoop Running in Production?

By | Blog

Following the publication of our White Paper, 2017 Preview: The Year of Enterprise-wide Production Hadoop, we ran a series of Twitter polls to get a rough sense of where the market is on the following 4 questions:

  1. Does your company use Hadoop in production?
  2. What stage are most of your Hadoop deployments (lab, PoC, Pilot, Enterprise-wide production)
  3. When will you have Hadoop in Enterprise-wide produciton use?
  4. What challenges did you encounter while expanding Hadoop use?

We started with the basics, asking first:

The split between production and non-production use is in line with what we hear from our community.

As we discuss at length in the white paper, this concept of “production” Hadoop can be misleading. For instance, pilot deployments and enterprise-wide deployments are both considered “production,” but they are vastly different in terms of DataOps, as table 1 below illustrates.

Table 1: DataOps Considerations from Lab to Enterprise-wide Production

In the next poll, we learned that 72% of Hadoop deployments are stacked up in the pre-enterprise wide stages.

One of the other diagrams you’ll find in our white paper is the Enterprise Hadoop Deployment Continuum. In the version below, I have added the percentages from the Twitter poll in each stage.

Figure 1: Most Hadoop deployments are in pre and limited production.

With this established, we then asked the Twitterverse when they expect to be enterprise-wide with Hadoop? Reassuringly, the same 28% that told us they were enterprise-wide in poll #2 reiterated this in poll #3.

Less reassuring, however, is that only 9% of those that are presently pre-enterprise wide have concrete plans to get into enterprise-wide in the next 12 months, and even fewer have such plans in the next 24 months.

An eyebrow-raising 55% said that they’re not sure when they will reach enterprise-wide deployment.

And when asked about the challenges big data pros faced increasing their use of Hadoop, responses were very evenly distributed across the four big areas we hear from the ODPi community.

ODPi is here to remove risk and uncertainty from Hadoop and Big Data. We do this through comprehensive testing suites that improve predictability and through compliance programs to ensure interoperability. In other words, ODPi is here to smooth and illuminate the path to enterprise-wide production use of Hadoop for the 55% of respondents that don’t know when (if?) they will get there.

And the ODPi Special Interest Groups, or SIGs, were set up to address the widespread challenges that poll #4 surfaced.

Like all the technical work at ODPi, SIGs are wide open for all to participate in.

Join us and help drive toward solutions in these areas.

Making Production Hadoop Faster, Easier and More Productive: ODPi Member Conversation with zData

By | Blog

In our third ODPi Member Conversation podcast, John Mertic met with zData’s Senior Solution Architect, Gagan Brahmi.

Their in-depth conversation centered around the struggles commercial and enterprise corporations face with Apache Hadoop deployments, along with why standardization is an important driver in the Big Data community.

Brahmi has a unique perspective on Hadoop’s role in the Big Data industry and offered specific insight around the ways today’s enterprises can derive the maximum value out of the Hadoop cluster and their data.

Touching on how the framework’s capabilities have grown tremendously in the last few years, and urging enterprises to recognize how many tools throughout the ecosystem now complement one another, he noted “When we are dealing with [this] tool to make sure we derive the maximum value out of our existing systems, we have to keep an eye on all the tools available.”

To hear more of Gagan and John’s discussion, tune in to the episode on our YouTube channel!

Subscribe to our YouTube channel and follow us on Twitter to catch upcoming episodes of the ODPi Member Conversation podcast series!

Making Production Hadoop Faster, Easier and More Productive: ODPi Member Conversation with Pivotal

By | Blog

In our second ODPi Member Conversation podcast, John Mertic sat down with Pivotal Software’s Head of Data, Jacque Istok.

Their engaging discussion focused on the challenges enterprises face when trying to get value out of their data in Hadoop.

As a founding member of ODPi, and former technologist with a long history in data warehousing and data analytics, Jacque also weighed in on standards and how they enable application interoperability, portability, governance and security across the ecosystem.

Tying the importance of standards into his thoughts around ODPi, he noted “Our vision has always been to make it easier for customers and products/vendors/projects to be able to interact with an enterprise standard for Hadoop in an easy and common way.”

To hear more of Jacque and John’s insight, including why it was important to Pivotal that ODPi was a core organization with a common goal for the Hadoop platform itself, tune in to the episode on our YouTube channel!

Subscribe to our YouTube channel and follow us on Twitter to catch upcoming episodes of the ODPi Member Conversation podcast series!

ODPi 2.1: a “tick” for the future “tock”

By | Blog

By: Roman Shaposhnik, VP of Technology at ODPi

The release of ODPi 2.1 marks five-months worth of the ODPi technical community’s diligent work, though on the surface it may appear to be incremental change to last fall’s 2.0 release. While there aren’t any big, splashy additions to our specification this release is very noteworthy in its own way. Why? Because it follows in the great tradition of tick-tock releases and invests a lot of energy into the underlying infrastructure that is largely invisible to the consumer. This, of course, makes it a “tick” release and those are truly foundational to the success of the follow up “tocks” that get all the excitement. If you still don’t believe tick-tock pairs well with complex systems, ask any Sun microsystems SPARK engineer how well an alternative release model has worked out for them I believe they called it humpty-dumpty, but I digress, so back to ODPi 2.1.

One of the biggest underlying changes in ODPi 2.1 is that we have fully transitioned to leveraging Apache Bigtop for our reference implementation and validation testsuite needs. This required a lot of upstream backporting. Some of it was pretty straightforward, such as backporting all ODPi-developed tests into Bigtop, while some required us to engage with upstream communities and get their feedback on the best way to accomplish a similar goal. This was the story of our ODPi reference implementation stack for Apache Ambari. It started as a custom stack that was shipped as part of the ODPi reference implementation but, after receiving community feedback, it evolved into a standalone management pack that can now be developed and shipped independently of Ambari. This outcome benefits everybody because now any product based on Ambari can simply point at the management pack and deploy ODPi reference implementation.

ODPi 2.1 is our first release consisting of just the specifications. All of the software artifacts are also being released as part of Apache Software Foundation. Such renewed alignment with upstream community efforts allows us to be much more in tune with big data practitioners, regardless of whether they participate in ODPi directly or not. This is a win-win for both ODPi and upstream ASF communities. If Bigtop release 1.2.0 was any indication, ODPi’s focus on enterprise stability and readiness brings to light a lot of issues that would otherwise go unnoticed or would only be fixed in vendor-specific patch releases. ODPi’s Bigtop collaboration brings these issues up closer to the source, creating a feedback loop that results in much faster fixes.

On the flip side, Bigtop’s extensive platform coverage and a vibrant community of ASF developers means the ODPi specification will bring value far beyond what we believe are our core deployment targets. For example, we’ve never really considered IBM’s POWER as a supported ODPi platform, but since Bigtop runs on this hardware, we get it for free. Starting from ODPi 2.1, all of the engineering work will happen directly in the upstream ASF communities, and we expect this to make our development cycle extremely agile and asynchronous. Of course, we’ll continue releasing the specifications, which brings me to the last part of this release.

Most of our effort on the Operations spec was focused on standardizing Ambari 2.5 and taking care of upgrade and backward compatibility guarantees for future ODPi releases. On the Runtime side, we spent quite a bit of time future proofing it against Hive 2.0 (and looking at how known incompatibilities with Hive 1.2 can affect ISVs and end users). We also considered Spark 2.0 as the next component on which to standardize.

New Special Interest Groups Spark Exploratory Developments

Our Spark 2.0 work was interesting in its own right. Our take was that while Spark was still considered experimental and not at the level of maturity that is required of ODPi Core components, it was still highly important to enterprise readiness. We’re tackling this through a loose construct of Special Interest Groups (SIGs), rather than a highly-rigorous body of a Runtime PMC. Thus, Spark gave birth to our first SIG: Spark and Fast Data Analytics SIG.

With the increase in the popularity and usage of Hadoop and Spark, the notion of Spark replacing Hadoop is gaining traction. While this is possible in some use cases, Spark is already part of Hadoop and there are several components from the Hadoop stack on which Spark depends. Our Spark and Fast Data Analytics SIG, led by Pradeep Roy, advisory software engineer at IBM, is expected to publish guidelines for Spark deployment and recommend best practices on Spark and Hadoop use, along with providing guidelines for different deployment methods for Spark on YARN, Mesos or Spark standalone; comparisons of different SQL on Hadoop solutions; and more.

The formation of two new SIGs, Data Security and Governance SIG and BI and Data Science SIG, quickly followed.

Our Data Security and Governance SIG was formed to provide a place for industry experts to collaborate on a set of best practices aimed at solving the complexities of dealing with multi-tenant Big Data data lakes in a secure fashion and with considerations for control points demanded by enterprise regulatory environments and compliance policies. As the leader of this group, my fellow members and I plan to produce a series of whitepapers and validation test suites addressing both platform considerations and solutions practitioners may need to augment their platform practices. This SIG’s first deliverable will be a Security Guide Handbook, developed on GitHub by members from IBM, Hortonworks and Pivotal, that will bring much needed clarity to securing Hadoop-based data lakes infrastructure. We’ve also started working on codifying security-related deployment recommendations as part of the Apache Bigtop deployment capabilities, thus providing baseline functionality around security for the entire Hadoop ecosystem. Stay tuned for our outputs, coming soon!

For our BI & Data Science SIG, according to the group’s champion Cupid Chan, managing partner of 4C Decision, we have a two-fold goal. The first goal is to help bridge the gap between Relational Database Management Systems (RDBMS) and Hadoop so that BI tools can sit harmoniously on top of these systems, while also providing the same, or even more, business insights to the BI users who also use Hadoop in the backend. Another goal is to collaboratively explore ways for Data Science to better leverage the underlying Hadoop ecosystem. In order to attain an achievable result, the first deliverable for this SIG is to develop a “Data Science Notebook Guideline.” Stay tuned for the release of this group’s findings!

While these SIGs are still very young, they are pushing forward important exploratory work that, we hope, will form a basis for some of the future PMCs and specification updates within the broader scope of ODPi.

These SIGs also represent our lowest barrier of entry to date – so, if you feel like contributing to ODPi efforts but don’t know where to start, we encourage you to join an existing SIG or propose a new one.

By default SIGs are using odpi-technical mailing list for all on-line communications between the SIG members. This means that all you have to do to join a SIG is drop an email to the odpi-technical mailing list, introduce yourself and briefly describe why are you interested in the SIG activity. Include your GitHub ID in the introductory email so that a SIG Champion can add you to the GitHub group.

Contributing to the ODPi community is that easy!

Predictable Hybrid Hadoop Blog Series – Crossing the Chasm

By | Blog

In the previous blog in this series, we outlined some of the important ways that running Hadoop in production – especially in enterprise-wide production – differs from point solutions and PoCs.

As a leading downstream community of big data vendors, users and platform providers, ODPi is focused on tackling the security, governance, lifecycle management and application portability needed to run Hadoop at scale.

A classic way to think about technology maturity is Geoffrey Moore’s Chasm model. In our new white paper, we plot key Hadoop milestones against the technology adoption curve (see image below) and argue that the things the ODPi community are focused on are essential to continuing the adoption of this transformative technology.

An adaptation of Everett Roger’s famous S diffusion of innovations curve, the Chasm model argues that users on the left of the chasm are fundamentally different from those on the right. The chasm separates users by adoption trigger/motivation: on the left it’s all about competitive advantage at nearly any cost, on the right it’s about continuity of operations and keeping up with the Joneses.

As awesome as this model is, it has sometimes been co-opted. One way this happens is by applying it to a Product, when in fact it needs to apply to a Category. This is one reason why we are so bullish about our work at ODPi – we explicitly acknowledge that the only way Hadoop and associated Big Data solutions can cross the chasm to mainstream adoption is by working together to define category-wide – NOT vendor-specific – answers to questions like lifecycle management, security and governance, application portability – these are the things that address early and late majority users’ interest in stability and operational continuity.

When thinking about what it really means for a technology to be a platform, we like the way Sam Ghod puts it:

A platform abstracts away a messy problem so you can build on top of it. Platforms do this by delivering portability and extensibility.

With ODPi Releases 1.0 and 2.0 in place, we invited Application Vendors to self-certify that their applications work unmodified across multiple ODPi Runtime Compliance Hadoop Distros. As of this writing, twelve applications from leading vendors like SAS, IBM and DataTorrent have completed the self-certification.

We believe that savvy Enterprise CDOs, CIOs, CTOs and Chief Information Security Officers (CISOs) should carefully consider the platform independence that ODPi’s Interoperable Apps program delivers before making their Hadoop platform choices. If one of your preferred vendors isn’t listed either as an Interoperable App or as a Runtime Compliant Platform, let that vendor know that it matters to you.

In 2017, we’re heads down adding to our existing specifications and creating new workstreams through our Special Interest Groups. We invite you to get involved. If you are a twitter user, be sure to follow @odpiorg and participate in our ongoing polls.

Looking at the latest Gartner Magic Quadrant for Business Intelligence and Analytics Platforms

By | Blog

By John Mertic

I spent some time reviewing the latest Gartner Magic Quadrant for Business Intelligence and Analytics Platforms in preparation for my time at the Gartner Data and Analytics Summit last week. Overall, I’m really excited to see vendors overall scoring higher in ‘Ability to Execute’; Gartner toughly judges this so seeing the general shift upwards is great to see.

While the piece is clearly targeted towards buyers of these tools – I wanted to take a critical eye on the positioning of vendors in relation to their interoperability with Big Data and Hadoop tools. After all, it was a mere decade ago that all of data was covered by a single Gartner analyst. Enter the age of Big Data; with that variability, velocity, and volume has come a cornucopia of products, strategies, and opportunities for answering the data question.

In the same way, BI and Analytics has come from being purely the realm of “data at rest” to become cohesive with “data in motion”. It’s no surprise then to see two “pure play big data” BI vendors, Datameer and ZoomData, joining ClearStory which joined the MQ last year – cementing the enterprise production need of valuable data insights. And with a tip of the hat to the new breed of open source trailblazers such as Hortonworks, they heavily leverage Hadoop and Spark as not just another data source but instead a tool to better process data – letting them focus on their core competency of delivering business insights.

However, what really struck me was the positioning of data governance as a whole in this report – let’s dig into that more.

Data governance and discovery is being pushed farther out

If you’d compare the 2016 report to the 2017 report – you’d immediately notice this line from 2016…

By 2018, smart, governed, Hadoop-based, search-based and visual-based data discovery will converge in a single form of next-generation data discovery that will include self-service data preparation and natural-language generation.

…became…

By 2020, smart, governed, Hadoop/Spark-, search- and visual-based data discovery capabilities will converge into a single set of next-generation data discovery capabilities as components of modern BI and analytics platforms.

Two year delay in just a year is something of note – clearly there is a continual gap in converging the technologies. This aligns with what our members and end-users in our UAB mention as well – the lack of a unified standard here is hurting adoption and investment.

Governance no longer considered a critical capability for a BI vendor

This really stood out to me in light of the point above – is sounds like Gartner believes that governance will need to happen at the data source versus the access point. It’s a clear message that better data management needs to happen in the data lake – we can’t secure at the endpoints for true enterprise production deployment. This again supports the needs of driving standards in the data security and governance space.

I recently sat down with IBM Analytics’ WW Analytics Client Architect Neil Stokes on our ODPi Member Conversations podcast series and the discussion of data lakes was a very present one. To listen to this podcast, visit ODPi Youtube.
I’m reminded of the HL Mencken quote “For every complex problem there is an answer that is clear, simple, and wrong.” Data governance is hard, and not ever going to be something one vendor will solve in a vacuum. That’s why I’m really excited to see the output of both our BI and Data Science SIG and Data Security and Governance SIG in the coming month. Starting the conversation in the context of real world usage, looking at both the challenges and opportunities, is the key to building any successful product. Perhaps this work could be the catalyst for smarter investment and value adds as these platforms continue to grow and become more mature.

Predictable Hybrid Hadoop Blog Series – DataOps Considerations From Lab to Enterprise-wide Production

By | Blog

In last week’s blog, The Hadoop Deployment Continuum, we covered how “in production” actually refers to a very diverse set of deployment scenarios. Anything from a PoC, to point solution, departmental deployment to enterprise-wide production can and often is called “production” use.

This blog focuses on the step-change DataOps requirements that come when you take Hadoop into enterprise-wide production.

As enterprises plan to move Hadoop and Big Data into enterprise-wide production scale out, they face a number of challenges.

Table 1, taken from our recent White Paper, details how running Hadoop and Big Data at enterprise-wide production requires a significant re-think across multiple dimensions.

The good news is that these are the very same challenges that ODPi big data community has been working on for over a year. Through our ODPi Compliance and Interoperable Apps programs, enterprises get stacks that are validated across a number of platforms, providing needed support for  multi-vendor procurement policies. In the words of Gene Banman, CEO of ODPi member DriveScale: “Enterprises have varying big data needs that require flexible and interoperable platform components. Becoming a member of ODPi will allow us to better maximize data center efficiency for Hadoop with interoperability for enterprise-grade deployments.”

Our ongoing work to validate workloads across cloud environments promises to extend ODPi predictability even further.

From a lifecycle management perspective, our Application Installation and Management specification covers requirements and guarantees for custom service specifications and views. Importantly, this spec, like all ODPi specs, is developed in the open and guided by the ODPi Technical Steering Committee (TSC), which is pulled from the entire Big Data industry. ODPi benefits from the involvement of end users, Hadoop platform providers, solution providers, and ISVs.

Last but certainly not least, our Special Interest Groups (SIGs), are looking into the following areas that are key to predictable enterprise-wide operations:

If these things matter to you, we invite you to get involved with any of these SIGs and/or join our slack channel and work with us to co-create a predictable hybrid future for Hadoop.