Exclusive Interview: Stephen O’Grady, RedMonk

May 18, 2011


The open source movement is expanding, and it is increasingly difficult for commercial software vendors to ignore. Some large firms have embraced open source. If you license, IBM OmniFind with Content Analytics, you get open source plus proprietary software. Oracle has opted for a different path, electing to acquire high profile open source solutions such as MySQL and buying companies with a heritage of open source. Sun Microsystems is now part of Oracle, and Oracle became an organization of influence with regard to Java. Google is open source, or at least Google asserts that it is open source. Other firms have built engineering and consulting services around open source. A good example is Lucid Imagination, a firm that provides one click downloads of Lucene/Solr and value-add software and consulting for open source search. The company also operates a successful conference series and has developed specialized systems and methods to handle scaling, big data, and other common search challenges.


I wanted to get a different view of the open source movement in general and probe about the more narrow business applications of open source technology. Fortunately I was able to talk with Stephen O’Grady, the co-founder and Principal Analyst of RedMonk, a boutique industry analyst firm focused on developers. Founded in 2002, RedMonk provides strategic advisory services to some of the most successful technology firms in the world. Stephen’s focus is on infrastructure software such as programming languages, operating systems and databases, with a special focus on open source and big data. Before setting up RedMonk, Stephen worked as an analyst at Illuminata. Prior to joining Illuminata, Stephen served in various senior capacities with large systems integration firms like Keane and consultancies like Blue Hammock. Regularly cited in publications such as the New York Times, NPR, the Boston Globe, and the Wall Street Journal, and a popular speaker and moderator on the conference circuit, Stephen’s advice and opinion is well respected throughout the industry.

The full text of my interview with him on May 16, 2011 appears below.

The Interview

Thanks for making time to speak with me.

No problem.

Let me ask a basic question. What’s a RedMonk?

That’s my favorite question. We are a different type of consultancy. We like to say we are “not your parents’ industry analyst firm.” We set up RedMonk in 2002.

Right. You take a similar view of industry analysts and mid tier consulting firms that I do as I recall.

Yes, pretty similar. We suggest that the industry analysis business has become a  “protection racket…  undoubtedly a profitable business arrangement, but ultimately neither sustainable nor ethical. In fact, we make our content open and accessible in most cases. We work under yearly retained subscriptions with clients.

Over the last nine years we have been able to serve big household names to a large number of startups. We deliver consulting hours, press services, and a variety of other value adds.

Quite a few firms say that. What’s your key difference?

We are practical.

First, RedMonk is focused on developers, whom we consider to be the new “kingmakers” in technology. If you think about it, most of the adoption we’ve seen in the last ten years has been bottom up.

We’re “practitioner-focused” rather than “buyer-focused”. RedMonk is focused on developers, whom we consider to be the new “kingmakers” in technology. If you think about it, most of the adoption we’ve seen in the last ten years has been bottom up. Our core thesis is that technology adoption is increasingly a bottom up proposition, as demonstrated by Linux, Apache, MySQL, PHP, Firefox, or Eclipse. Each is successful because these solutions have been built from the ground floor, often in grassroots fashion.

Third, we are squarely in the big data space. The database market was considered saturated, but it exploded with new tools and projects. A majority of these are open source, and thus developer friendly. We are right in the epicenter of that shift.

Do you do commissioned research?

No, we don’t do commissioned research of any kind. We just don’t see it as high value, even if the research is valid.

How has the commercial landscape of search specifically, and data infrastructure generally, been impacted – for better or for worse – by open source?

As with every other market with credible open source alternatives, the commercial landscape of search has unquestionably been impacted. Contrary to some of the more aggressive or doom crying assertions, open source does not preclude success for closed source products. It does, however, force vendors of proprietary solutions to compete more effectively. We talk about open source being like a personal trainer for commercial vendors in that respect; they can’t get lazy or complacent with open source alternatives readily available.

Isn’t there an impact on pricing?

Great point.

Besides pushing commercial vendors to improve their technology, open source generally makes pricing more competitive, and search is no exception here. Closed source alternatives remain successful, but even if an organization does not want to use open source, search customers would be foolish not to use the proverbial Amdahl mug as leverage in negotiations.

When the software is available for free, what are customers paying for?

Revenue models around open source businesses vary, but the most common is service and support. The software, in other words, is free, and what customers pay for is help with installation and integration, or the ability to pick up the phone when something breaks.

A customer may also be paying for updates, whereby vendors backport fixes or patches to older software versions. Broadly then, the majority of commercial open source users are paying for peace of mind. Customers want the same assurances they get from traditional commercial software vendors. Customers want to know that there will be there someone to help when bugs inevitably appear: open source vendors provide that level of support and assurance.

What’s the payoff to the open source user?

That’s my second favorite question.

The advantages to this model from the customer perspective are multiple, but perhaps the most important is what Simon Phipps once observed: users can pay at the point of value, rather than acquisition. Just a few years ago, if you had a project to complete, you’d invite vendors in to do a bake off. They would try to prove to you in an hour or two demo that their software could do the job well enough for you to pay to get it.

This is like an end run, right?

In general, but we believe open source software inverts the typical commercial software process. You download the software for free, employ it as you see fit and determine whether it works or not. If it does, you can engage a commercial vendor for support. If it doesn’t, you’re not out the cost of a license. This shift has been transformative in how vendors interact with their customers, whether they’re selling open source software or not.

The general complexion of software infrastructure appears to be changing. Relational databases, once the only choice, are becoming rather one of many. Where does search fit in, and how do customers determine which pieces fit which needs?

The data infrastructure space is indeed exploding. In the space of eighteen months we’ve gone from relational databases are the solution to every data problem to, seemingly, a different persistence mechanism per workload.

As for how customers put the pieces together, the important thing is to work backwards from need. For example, customers that have search needs should, unsurprisingly, look at search tools like Solr. But the versatility of search makes it useful in a variety of other contexts; AT&T for example uses it for Web page composition.

What’s driving the adoption of search? Is it simply a function of data growth, as the headlines seem to imply, or is there more going on?

Certainly data growth is a major factor. Every year there’s a new chart asserting things like we’re going to produce more information in the next year than in all of recorded history, but the important part is that it’s true. We are all–every one of us–generating massive amounts of information. How do you extract, then, the proverbial needle from the haystack? Search is one of the most effective mechanisms for this.

Just as important, however, has been the recognition amongst even conservative IT shops that the database does not need to be the solution to every problem. Search, like a variety of other non-relational tools, is far more of a first class citizen today than it was just a few short years ago.

What is the most important impact effective search can have on an organization?

That’s a very tough question. I would say that one of the most important impacts search can have is that a good answer to one question will generate the next question. Whether it’s a customer searching your Web site for the latest Android handset or your internal analyst looking for last quarter’s sales figures, it’s crucial to get the right answer quickly if you ever want them to ask a second.

If your search fails they don’t ask a second question, you’ll either have lost a potential customer or your analyst is making decisions without last quarter’s sales figures. Neither is a good outcome.

Looking at the market ahead, what trends do you see impacting the market in the next year or two? What should customers be aware of with respect to their data infrastructure?

There are a great many trends that will affect search, but two of the most interesting from my view will be the increasing contextual intelligence of search and the accelerating integration of search into other applications. Far from being just a dumb search engine, Solr increasingly has an awareness of what specifically it is searching, and in some cases, how to leverage and manipulate that content whether it’s JSON or numeric fields. This broadens the role that search can play, because it’s no longer strictly about retrieval.

And integration?

Okay, as for integration, data centers are increasingly heterogeneous, with databases deployed alongside MapReduce implementations, key-value stores and document databases.

Search fills an important role, which is why we’re increasingly seeing it not simply pointed at a repository to index, but leveraged in conjunction with tools like Hadoop.

What kind of threat does Oracle’s lawsuit over Google plus Java pose to open source?How does it compare to the SCO controversy with Linux some years back?

In my view, Oracle’s ongoing litigation of Google over Java related intellectual property has profound implications for both participants, but also for the open source community as a whole.

The real concern is that the litigation, particularly if it is successful, could have chilling effects on Java usage and adoption. As far as SCO is concerned, this is somewhat different in that it targets a reimplementation of the platform in Android rather than the Java platform itself. SCO was threatening Linux rather than a less adopted derivative.

While users of both Java and MySQL should be aware of the litigation, however, realistically the implications for them, if any are, are very long term. No one is going to abandon Java based open source projects, for example, based on the outcome of Oracle’s suit.

It seems like everyone who is anyone in the software world has an open source strategy, even through to Microsoft’s embrace of php. Should information technology executives and decision makers, who were once suspicious of open source, be suspicious of software vendors without a solid open source strategy?

With the possible exception of packaged applications, open source is a fact of life in most infrastructure software markets. Adoption is accelerating, the volume of options is growing, and – frequently – the commercial open source products are lower cost. So it is no surprise that vendors might feel threatened by open source.

But even if they choose not to sell open source software, as many do not, those without a solid open source interoperability and partnership story will be disadvantaged in a marketplace that sees open source playing crucial roles at every layer of the data center. Like it or not, that is the context in which commercial vendors are competing. Put more simply, if you’re building for a market of all closed source products, that’s not that large a market. In such cases, then, I would certainly have some hard questions for vendors who lack an open source strategy.

Where can a reader get more information about RedMonk?

Please, visit our Web site at www.redmonk.com.

ArnoldIT Comment

RedMonk’s approach to professional services is refreshing and a harbinger of change in the consulting sector. But more importantly, the information in this interview makes clear that open source solutions and open source search technology are part of the disruption that is shaking the foundation of traditional computing. Vendors without an open source strategy are likely to face both customer and price pressure. Open source is no longer a marginalized option. Companies from Twitter to Cisco Systems to Skype, now a unit of Microsoft, rely on open source technology. RedMonk is the voice of this new wave of technical opportunity.

Stephen E Arnold, May 18, 2011


One Response to “Exclusive Interview: Stephen O’Grady, RedMonk”

  1. Lucid Imagination » Using open source isn’t free? Could be search for big data is the win on May 19th, 2011 11:07 am

    […] Arnold, noted Enterprise Search blogger behind the Beyond Search blog, spoke to Stephen O’Grady, the co-founder and Principal Analyst of RedMonk, who had an […]

  • Archives

  • Recent Posts

  • Meta