OSBC 2008 The Hidden Session

There's a blank space in the agenda for this time slot. A panel has been brought together to discuss if open source companies can survive the M&A. The panel includes Harold Goldberg, CEO, Zend Technologies, Ben Sabrin, VP, Sales & Business Development, Appcelerator, Zack Urlocker, EVP of Products, MySQL, and Rex Wang, VP, Product Marketing, Oracle.

The community, the employees, the product - everything could change.

Zend and MySQL are strong partners, as over 50% of development on top of MySQL is in PHP. Sun acquiring MySQL initially was upsetting, but more from the standpoint that the world needs strong, independent open source companies. The two companies are dedicated to supporting the overlapping communities through the transition and to maintain the direction of MySQL. Sun and MySQL, whether still being in the honeymoon period or indicative of long term cultures, are having open, supportive conversations.

Red Hat and JBoss had no such conversations, and the cultural shift was very negative, though the results may be positive.

Harold brought out that there are three types of M&A: acquire a business, consolidate markets, purchase a piece of technology. Open Source doesn't make this different.

Sleepycat was acquired by Oracle two years ago, and it was announced at OSBC 2006, and thus has been the longest in the transition. Sleepycat personnel were integrated with comparable groups within Oracle, e.g. Sleepycat engineers joined database engineering, marketing joined marketing, sales formed with sales folk from other recent Oracle acquisitions, a new sales group focused on embedded databases and small devices. Though this integration and dispersal happened over many months, and some processes, such as the release cycle are still done the Sleepycat way and may never transition to the Oracle way.

There are gradations of M&A transition & integration policies, ranging from complete absorption into the new culture, to maintaining the acquired company intact while the acquiring company learns the new business. Most acquisitions fail. It's too early to tell how these acquisitions will run, but a lot depends on the M&A experience and competency of the acquiring company.

OSBC 2008 Server Out Network In

I haven't been doing any "live blogging" this OSBC. Yesterday was very hectic. I did some interviews, attended some sessions, and had many great conversations. Today was less hectic, and I even got to pull out the MicroTrak to record a podcast with Brian Reale of ProcessMaker.

I actually got to this panel session early enough to grab a piece of table and set-up my laptop. This panel is "Open Source: Out of the Server, Into the Network" moderated by Larry Augustin, Managing Director, Augustin Ventures and comprised of Artur Bergman, Director of Engineering, Wikia, Fabrizio Capobianco, CEO, Funambol, Kelly Herrell, CEO, Vyatta, and Mark Spencer, Founder and CTO, Digium.

Much of the conversation in the first part of the panel centered around Cisco as a monopoly, pricing of "enterprise class" at three or four orders of magnitude over similar items from the the local electronics store, and adapting Linux for network use. What I heard was the commoditization of the network. The same discussion I was having with the CEO of a communications VAR 10 years ago. At that time, the industry was going through a sea change where the old way was to have system engineering as a free, value-added service as part of the sales process for large, multi-million dollar PBX, EDGE Router, telecomms & networking deals. The new way hadn't been decided yet, but the margins were being forced down to under 15% - commoditization. Today this panel was talking about COTS hardware running specialized open source software to replace network gear and PBXs - commoditization.

But where's the value? Fabrizio hit the nail on the head: community, the power of community and using open source to build a community. Why is there value in community? I think we can sum this up with customization, localization, and common needs.

Larry summarized it as a vertical market coming apart horizontally.

Great session all the way around.

Adaptive Planning at OSBC2008

I met with Bill Soward, President & CEO, and Greg Schneider, Vice President Marketing of Adaptive Planning. Adaptive Planning provides packaged solutions for business performance management (BPM), but I think that the best description of Adaptive Planning is that they provide focused BI tools, both on-demand as SaaS and on-premise as open source software, which is licensed under GPLv2.

Since its founding in 2003, Adaptive Planning has tripled its customer base year over year, and has grown to over 220 customers, with 90% of these being customers of the on-demand SaaS. However, approximately half of the largest customers use the open source on-premise version. Most of these customers are companies with 100 to 2500 employees. In toto, they've taken US$29 million in VC funding.

The open source version is available from SourceForge and SugarForge, and has been downloaded over 65,000 times in 85 countries. They have data that suggests that a significant percentage of these downloads have been by finance folk, not IT. These fits well with our experience that end-users are starving for simple BI tools that simply aren't being provided by IT in many cases. I told the "lake of data" story, and I'll repeat it again. Several years ago, we were giving our benefits of BI pitch to a VP of Marketing at a business unit of a Fortune 100 company. She stopped us and brought in one of her marketeers. This very non-technical fellow had recreated data warehousing using over 100 linked MS Access database applications that he called the "Lake of Data". It was amazing. It worked, and it filled their needs. This story pre-dates any significant open source options for BI, but it is exactly this type of person who might download Adaptive Planning Express.

Their community model is a bit different and I believe that it is very worthwhile. They build their community around partners with specific domain expertise. This expertise may be within a vertical niche such as health or the public sector, or specific geography. One example of this is the Adaptive Planning partner in Japan worked with their customer AKTIO to implement localized version of Adaptive Planning's open source budgeting, planning and reporting tools for 800 seats, on-premise at AKTIO. There are currently 65 partners working on extending Adaptive Planning in different ways.

Adaptive Planning is also looking to partner with other open source companies, particularly in the area of data integration. When Adaptive Planning added reporting tools, they originally looked at using one of the open source tools. At the time however, the existing open source reporting tools were very much for developers, and not for end users. Adaptive Planning is very much about ease of use for the end users, so they decided to build their own dashboarding and reporting set. However, with the maturation of open source offerings over the past two years, Adaptive Planning will likely not build their own data integration tools, but will partner with an existing open source solution.

In many ways, the chief competitor for Adaptive Planning is MS Excel. That is, the end users choose the status quo of slogging their way through Excel workbooks and pivot tables. Bill emphasized that the Adaptive Planning strength is allowing user collaboration on modeling, assumptions and background information in a secure, controlled way.

SourceSense and Microsoft Apache POI

I'm having lunch with Gianugo Rabellino, who is very excited about the new partnership between his company SourceSense, the leading open source integrator in Europe, with the little known &#59;) software vendor from Redmond Washington in the USA. The partnership centers on a new top-level Apache project, Apache POI, which is the open source file format reader and writer to create, edit and read Microsoft Office formats used in Excel, Word, PowerPoint and Visio, and supports Open XML.

Microsoft provides the detail in a press release, Microsoft and Sourcesense Partner to Contribute to Open Source, Apache POI to Support Ecma Office Open XML File Formats. Congratulations to Gianugo and Sourcesense.

WSO2 Interview at OSBC 2008

I met with Paul Fremantle, Co-Founder & CTO of WSO2 at OSBC 2008. Paul and WSO2 CEO Sanjiva Weerawarana both come from IBM Web Services architecture and founded WSO2 to provide an IBM level of quality of service in a simpler implementation. I was actually introduced to Paul and Sanjiva about an half-hour before our scheduled interview by our mutual friend Gianugo Rabellino. Here's some of the information that Paul provided about WSO2.

WS* for security & messaging is a major focus for WSO2, and WSO2 is filling some of the gaps we mentioned in our "suggested topics" for OSTT, especially in the area of exposing data through WSO2 Data Services and Governance in the WSO2 Registry:

  • RESTful not UDDI
  • ATOMpub
  • Social Governance: Tags, Comments, Ratings, subscribe via any feed reader to tags, versions, etc.

Independent security through WSO2 User Manager or standard APIs and applications

Fifty percent of the WSO2 developers times is spent in Apache projects, and the company is 95% engineering. This shows one of the strengths of Open Source in that the company is not heavy with GS&A costs.

As anyone reading this blogs knows, we feel that Community is the major strength of any Open Source company. The WSO2 Community provides requirements & bug fixes, primarily. One recent example of the value of the WSO2 community is where Hessian provided requirements & testing.

WSO2 is very much a platform company, much like Spring. WSO2 worked with Spring to have Apache Axis2 work as a component within Spring using Axiom (AXis Object Model).

  1. Spring WS requires Contract First
  2. The WSO2 framework for Spring allows POJO Code First

The Axis2 beta was published six weeks ago and a GRAILS contributor has already contributed a plugin to the community that is dependent upon the WSO2 web services framework for Spring.

Business models for Open Source companies are still being developed. WSO2 achieves monetization through subscription, training, developer support [email & IM] and product support. All WSO2 products fall under Apache Licenses.

In addition, WSO2 provides Open Source Development: customers pay to get customizations into the main tree of the code base of Apache projects that are important to them. This is a great way to provide support to your favorite open source project.

We ran out of time at this point. In conclusion, we feel that WSO2 is filling some of the gaps in open source products for SOA, as well as having some unique approaches in extending their business model.

OSBC 2008 Opening Keynotes

The OSBC 2008 started with a Keynote from Matt Asay. Here are some of my "takeaways"

  1. There has been a large uptick in participation at this year's OSBC from members of IT [customer-side]
  2. 2B$ invested in OSS & >2B$ given back through M&A activity in the past year
  3. Some stats
    • 71% of enterprises using OSS are expanding their use of OSS
    • 65% say use of OSS sparked innovation
    • 85% say 0SS hiqher quality
  4. Matt shows off his video editing & mashup skills. The result defies my ability to describe, you'll just have to wait for the YouTube download ;-)

The next Keynote was from Jim Whitehurst, President & CEO of Red Hat. He started with a marketing multimedia that gave some interesting facts:

  • 3000+ certified applications are available
  • Red Hat linux is certified on thousands of hardware systems

Red Hat is seen as one of the largest and most successful open source vendor, and it will be stepping up to its role as an OSS leader.

The career path at Red Hat starts in Services with promotion to Development. This reminds me of my time in aerospace, where a "proper" career path was to start in QA/Test, move to development and then "retire" back into a QA role.

Jim has never seen a company as focused on the customer as is Red Hat.

Jim went over some points that fall under Red Hat's evolving Business Model.

  1. good for customers as increase value every day through iterative innovation with stability through enterprise edition
  2. Red Hat will do a better job of involving large enterprise customers in the community
  3. Red Hat's goal is to become the defining technology company for 21st century through its work with the community

I had to skip the next keynote to prepare for my first scheduled meeting of the day.

OSBC 2008 Starting Up

Clarise and I have registered and we're just getting coffee. :p We've already run into a few folk from the Open Source Think Tank. Matt Asay's keynote starts in 15 minutes. We'll see how different it is from last year. &#59;)

The WiFi is free, the coffee is hot, and the Open Source Showcase room is filling up and the booths are already active.

We'll try to keep you up to date as it happens. But we're not going to try and live up to Coté's calling us a "blogging machine" last year. &#59;D

OSBC 2008 Starts Tomorrow

The Open Source Business Conference starts tomorrow in San Francisco. We have a fairly full schedule of interviews we're conducting and sessions that we wish to attend. But if you want to meet, let us know, and we'll arrange it. A lot is happening in the data management and analytics open source world. We'll be conducting interviews with:

And perhaps some locals, either at or after the conference. We're looking forward to seeing some old friends, like Gianugo Rabellino and Dave Rosenberg, as well. Speaking of Dave, don't forget to sign-up for MuleCon2008 starting appropriately on April Fool's day.

Thoughts on Open Source Think Tank 2008

The third annual Open Source Think Tank, hosted by the Olliance Group and DLA Piper, will be February 7 - 9 this year, at the Silverado in Napa Valley, CA. As the title states, this is a think tank, and not a conference. The schedule shows two CIO Panel Discussions, four brainstorming sessions and many hours of networking.

Attendees have been asked to think about topics they would like to see discussed, and here's what Clarise and I have bandied about.

  1. Our contention has always been that the strength of open source lies in flexibility and community. These two qualities make open source more valuable than the closed competition. How can new business models build on these strengths, and move away from licensing, subscriptions, support and other means of pushing "80% of the capabilities for 20% of the cost" as the message?
  2. SaaS is a new way of deploying and delivering software to customers, and open source leads to new ways of development of software and new business models to commercialize this software. Some companies make use of one and make money off the other, in various ways and various combinations. Mashups and SOA and "Platform as a Service" are leading to even greater re-use and innovation. What new synergies and business models can come about from these concepts?
  3. Open source software, and even hardware designs, cover many areas of business and consumer applications: operating systems, web servers, databases, reporting, productivity, web browsers, and on and on. However, there are still holes, actual and perceived. These holes exist in segments, such as Governance, in functions within an application, and in vertical/niche markets. Will these holes be filled by existing open source companies expanding their offerings, or by new companies? How does the consolidation, acquisition by proprietary companies, and leveling-off of VC funding for open source companies affect the expansion of open source to fill these holes?
  4. There is continuing resistance to open source in organizations of all sizes. From some smaller companies, we've seen two types of responses: one being that they don't have the skills to use, let alone leverage, open source software, and the other being that they gain credibility when they tell their customers that they are using Oracle, WebMethods, Weblogic, Sharepoint and other "brand" names, and that they are afraid of losing this credibility if open source is anywhere except deep under the covers. From large companies, the refusal is often more abrupt and more definite. Arguments of flexibility, security and lower cost often fall on deaf ears.
  5. Some in the open source world have blogged about the dearth of executive talent for open source companies. What talent and background make for a good open source corporate leader? At what point in an open source company's life cycle is general business acumen more important than open source commitment?
  6. It can often be difficult to find the open source project needle in the haystack of the web. It can be very frustrating to find a project that looks like a solution, only to discover that after several years, it never left the "proposal" or alpha stage. Using tagging, ranking and other Web2.0 social networking type of tools on Sourceforge.net and other forges, foundries and code repositories might help to promote and strengthen ideas and projects into the mainstream, or at least the "Long Tail".

Here are some other attendees' thoughts from their blog posts:


Today I had the good fortune of speaking with Doug Moran, Founder and VP of Community, and James Dixon, Founder and Chief Geek/CTO of Pentaho about their OpenScrum methodology. Doug had responded to my question on LinkedIN, "What has been your experience with Master Data Management and SOA, SaaS and Agile, in any combination?"

In creating the OpenScrum Agile software development method, Pentaho faced many of the same challenges that we've been facing with some of our customers:

  1. Extending Agile methods to be used by distributed workgroups, in Pentaho's case this includes every inhabited time zone on this planet for their extended community and several time zones in the USA and EC for the core group
  2. Adjusting Agile methods to work with several products or modules or projects in parallel
  3. Documenting the communication among team members even when "The most efficient and effective method of conveying information to and within a development team is face-to-face conversation" isn't possible - quote taken from the Agile Manifesto Principles
  4. Dealing with the concepts of timeboxes and rhythm
  5. Deciding on who best fills the roles of product owner, scrum master, and the various levels of commitment among the core group, the extended team and the community at large
  6. How best to involve QA/Test - a subject on which Scrum gives little guidance

The OpenScrum methodology deals with much of this, as does our ever evolving N Dimensions of a Project methodology, which has gone from 5D™ in the mid- to late-90's to the 8D™ iteration that we're currently documenting.

The most interesting point to come from our discussion is that Doug and James have come to many of the same conclusions and principles as have Clarise and I, even though our "experimental sample" &#59;) is the Pentaho open source communities on the one hand, and the various internal IT and SaaS software development distributed workgroups on the other.

One of the first things we discussed concerned the idea of fixed timeboxes and rhythm. The idea of rhythm is very attractive, but very difficult to achieve in most situations. In many ways, the concepts that Agile methods address: responsiveness to changing user needs and a changing market, coupled with the realities of changing personnel and other organizational and business changes, make achieving a short-term rhythm very difficult. As an organization matures, perhaps a longer term rhythm, such as the quarterly or "seasonal" rhythms of salesforce.com might be a realistic goal. As an organization first moves into an Agile process, and develops an Agile software development methodology that fits with their evolving culture and ecosystem, sprint timeboxes must be flexible, and achieving a fixed release rhythm isn't practical. For our part, we make setting the timebox part of each sprint planning meeting. Once set, the timebox is inviolable, and any further negotiations must be around the feature set to be achieved in the sprint.

In the same area of time, James and Doug shared their observation that the sprint burn-down chart can result in a false sense of security or panic. No work effort follows the nice smooth regression line of a sprint burn-down chart, but more sophisticated modeling, to plan for the stepped nature of real-life work, is beyond simple spreadsheets. It's not beyond the current math modeling state of the art nor the capabilities inherent to the various analytic and data mining modules of Pentaho. While Pentaho currently has no plans for a "Pentaho for OpenScrum" similar to "Pentaho for Jira", it's certainly food for thought for the Pentaho community, and led to some fun brainstorming on our call.

An issue related to time, and space in this instance, is that almost everyone today works in some form of distributed workgroup. This may include the occasional telecommuter, or, as in the case of Pentaho and other open source projects, communities that span the globe. This would seem to preclude most Agile methods, especially Scrum, which puts a great deal of importance on stand-up face-to-face meetings, specifically the daily scrum. One way to get around this is to use instant messaging or teleconferences, especially if remote whiteboarding can be used - though always one group or another is inconvenienced by the selected time. Another way is to use blogs, wikis or forums to supplement the daily scrum, especially for those who might not be able to attend due to the time selected. One other things that we've done is to record the daily scrum, either audio alone, or with video. The take-away here is that some form of asynchronous communication, and keeping an historical reference, is a necessary addition in the face of the flat world of software development today.

The idea of parallel scrums has been batted around within the Agile communities for some time now. The reality is that sprints must often be done in parallel and some individuals are going to be assigned roles in more than one sprint team simultaneously. The reasons for this are many, but the bottom line was discovered by project managers long ago: small teams with sharply focused goals are more often successful than not. Large, complex or broadly defined goals must be managed as programs or portfolios, not as projects. We've followed separate strategic, tactical and implementation tracks for over a decade, with a modified iterative waterfall approach cycling through these three tracks throughout time. This allows us to easily incorporate parallel implementation sprints for each tactical project. The OpenScrum methodology has some good depictions of their approach to parallel sprints. For another perspective, I can draw upon a buffet-line-conversation I had with a senior project manager [who felt that Agile was "crap"] at a recent SFBAC (San Francisco Bay Area Chapter) meeting of the PMI (Project Management Institute) and that is the fact that Agile Software Development methods are not generally applicable Project Management methods. [My opinion and not necessarily representative of anyone else mentioned in this post.] As such, it's much easier to see how parallel sprints can be planned. With Agile methods, we're developing software, not dams, bridges or space shuttles. Software development is much more of an art than an engineering discipline, and the Agile methods allow for this artistry.

The roles of Product Owner, Scrum Master and Scrum team are very important to Scrum and OpenScrum and 8D™ and still requires refinement. Many folk that we've encountered in our consulting, naturally assume that the role of Product Owner is best filled by an existing Product Manager and that the Scrum Master is a lead engineer. But just as your best technical person is often a poor choice for any management role [soapbox time: forcing great techies into management roles as the only path for advancement and "the big bucks" is just plain stupid today] the lead engineer may not be the best Scrum Master. We feel that the Product Owner should be a true representative from the user community, or at the very least a customer advocate. A Scrum Master must be skilled in corporate communication, escalation and, yes, politics, and empowered to clear obstacles that could prevent the sprint from achieving its goal. In addition to these two roles, the Scrum Team should have no more than seven (7) other [full time equivalent] individuals with technical and business skills, and subject matter expertise that assure the sprint's success. This is a guideline, with lots of ways to be implemented.

Another area of discussion that deserves weeks of attention, not the few minutes that we could devote to it in our call, is how best to incorporate QA and all the various testing efforts that any engineering effort requires. Well-thought-out test scenarios and automation certainly play an important part, but that's not the end of it. And we didn't get anywhere near the end in our discussion. Most Scrum articles, books, etc. kind-of ignore QA. Other Agile methods, such as Test Driven Development are centered around it.

We believe in incorporating QA and Test as part of the core group, and that code review & unit testing, at the least, must be a part of the automated daily build process. But what about regression and User Acceptance Testing (UAT)? If the idea is that each sprint results in usable code, isn't regression testing a major part of getting to usable code? Isn't UAT the proof of the pudding? But are such activities a part of the Sprint, a separate, overlapping/parallel Sprint, or a separate activity altogether? And what about QA of specifications, documentation, data integrity and metadata?

There are also some issues specific to open source companies/projects related to User Acceptance Testing, and user satisfaction metrics in general. It's one thing to count downloads, but how does one measure the comments, or really, the lack of comments in the forums? Open source companies often don't know who has downloaded their product, and don't whether the lack of comments on a new release is indicative of satisfaction or disinterest or even disgust. One way in which an open source company can get some amount of automated feedback is to have the product report an heartbeat, or other "phone home" technique [see Positive Feedback Enablers in the OpenScrum methodology]. Pentaho has implemented an "opt-in" or "opt-out" system for their heartbeat and has had no negative feedback for implementing this system.

As you can see… many questions. The answers are often dependent upon the supporting infrastructure, maturity and culture of the organization or community.

The culture results from the personalities of the pigs [the truly committed], the chickens [the involved], the sheep [who prefer to be herded] and the goats [who prefer to be led]; and don't forget the penguins, cats and lone wolves either. :>> Whether in an open source community or a corporate team, a manager may know the core group, but might only guess at the extended team or overall community. And, as with any generalization, the devil is in the details. This is why our methodology has gone from 5D™ to 8D™ over the years, and why James has developed both the Beekeeper and OpenScrum methodologies. From the initial framework for success, one must apply the principles of Agile even to the development of your Agile methods. Again and again, we see the need to adapt to specific situation, cultures and objectives.

We talked for well over an hour. One area that we didn't have time to explore was the Sprint Retrospective vs. Sprint Review. The OpenScrum methodology only discusses the Sprint Retrospective [among the truly committed] whereas we also allow for the Sprint Review [among all interested parties] in our "lessons learned" processes. Perhaps we can go into this further on another call or in the comments.

As always, talking with the Pentaho folk is lots of fun. We're always encouraged by their inventiveness in their execution as an open source company.

December 2019
Mon Tue Wed Thu Fri Sat Sun
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31          
 << <   > >>

At the beginning, The Open Source Solutions Blog was a companion to the Open Source Solutions for Business Intelligence Research Project, and book. But back in 2005, we couldn't find a publisher. As Apache Hadoop and its family of open source projects proliferated, and in many ways, took over the OSS data management and analytics world, our interests became more focused on streaming data management and analytics for IoT, the architecture for people, processes and technology required to bring value from the IoT through Sensor Analytics Ecosystems, and the maturity model organizations will need to follow to achieve SAEIoT success. OSS is very important in this world too, for DMA, API and community development.

37.652951177164 -122.490877706959


  XML Feeds