Community

It's Pronounced "Duffle Tee See See"

Posted on by tarus

I want to point out a post Alex Finger (OGP) wrote over the weekend.

It has been said that you can divide open source people up into three groups: Those who write open source code, those who pay for open source code and “free riders”. I think that is total crap, and Alex is one of the reasons why.

Alex doesn’t write much code, yet he has been a great contributor to OpenNMS over the years. His career is in IT management (as in managing people and solutions), and he’ll soon take a new position as a CIO. What he brings to the table is a view on open source from upper management.

It is no secret that OpenNMS is usually introduced into an organization from the bottom up (usually, but not always). We don’t have any full time sales people and most of our leads come from people who have already downloaded and installed the software. I don’t see anything wrong with that, but sometimes it is nice to be able to talk to the top people in charge, especially when you want a check written. In order to be successful when talking with those at the top of a management structure, it helps to understand them, as they process information much differently than the technical people I’m used to.

To me an open source community is a lot more than just who writes code. Everyone who uses the software can contribute. But until this weekend I did not have a good model for understanding the process of how one becomes a member of that community.

Alex has come up with something he is calling “the DUFLTCC Cycle”. It’s an acronym that stands for Download, Use, Fail, Learn, Teach, Change, Commit.

Check out his post for the full details. I especially like the “Fail” step. A lot of open source software requires a steeper learning curve than commercial off the shelf products, and usually everyone gets stuck at some point. It’s those that use those failures as a learning experience that, should they overcome them, tend to become some of the more vociferous proponents of the project.

I often paraphrase The Matrix when discussing my open source experience in that it is one thing to know the path and another thing to walk the path. Once you walk the path the use of open source seems to be a no-brainer. But how do you get that across to people who have never even seen the path in the first place? It’s ideas like Alex’s that help, and work like this is as important as any code in a community like ours.

Let the Sun Shine In

Posted on by tarus

It’s been cold and snowy for most of my stay in Milan, but today the temperatures warmed and the sun came out.

We held a seminar on OpenNMS and open source for over 20 of Italy’s best and brightest IT decision makers and professionals (I know this because they were interested in OpenNMS, so it goes to reason that they were intelligent, amazingly witty and very attractive to the opposite sex).

I was limited to a 20 minute presentation and 10 minutes of questions which is extremely hard for me to do (since the guys in the office say it takes me more than 20 minutes to introduce myself) but I think it went well.

I started off with a discussion of software business models, and then talked about the various permissive and restrictive open source licenses. I couldn’t resist doing a slide on open core software since it is my own personal goal to make sure that open core is seen as the commercial software business model it is versus an open source one, and this being Europe I think it was well received. While I think it is quite possible for open source and proprietary software to live side by side, I don’t think this is possible within one company (well, at least a small company – IBM might be able to pull it off).

The reason is simple: If there are proprietary features that drive software revenue you can bet that they won’t ever become part of the “community” edition. In fact, I bet that any contribution from the community that threatens that revenue stream will be refused. The goals of an open source community and a commercial software company are hard if not impossible to align.

I then talked about the OpenNMS business model. Since our mantra of “Spend Less Than You Earn” allows us to exist year after year, there is no danger of OpenNMS ever going away. With our active and growing community we will keep improving OpenNMS and thus provide pressure to our commercial competition. In our target market of large enterprises and carriers, solutions are driven by knowledgeable professionals both within these companies and via outside consultants, and by making OpenNMS the best tool for the job we expect to see widespread adoption both through the commercial side of OpenNMS as well as from the community.

It may take ten years, but I fully expect OpenNMS to one day be the default platform for any large scale management solution.

The people in the room today together spend more than 10M€ a year on network management. They have suffered through expensive solutions that never delivered on their promises and they have had few options but to switch to another expensive solution.

While downloading OpenNMS doesn’t instantly fix their problems, combined with the right hardware, services and perhaps some custom development it can immediately start to reduce costs while increasing functionality. Once in place OpenNMS does not require expensive maintenance contracts and can represent a much lower cost of ownership than a commercial product.

But most importantly OpenNMS represents freedom. The fact that the code is 100% open moves the power from the vendor to the client. This seemed to be important for the people in the room.

I managed to get all this and more in my allotted time, and I think it was well received. We have a number of large projects going on in Italy and, while challenging, it gives us a chance to shine.

Just like today’s sun.

Another Reason to Use Open Source

Posted on by tarus

Even though OpenNMS is a open source project, we do sometimes receive support from commercial vendors. For example, when we do demonstrations where Internet access isn’t available, we often use Gambit’s Mimic software to simulate a network. In exchange for temporary licenses we place a link to their website in the footer of the OpenNMS wiki.

Back in 2007, Johan Edstrom, one of our OGP members, really liked the IntelliJ IDEA IDE. While most of our developers use Eclipse, he was just more comfortable with the IDE from JetBrains and since they offered free licenses for people who work on open source projects, he wrote to them and asked for one. Here’s the response he got:

We are pleased to support the Open Source community and we look forward to seeing your project’s progress. If we can be of any additional service, please don’t hesitate to ask.

Also, while it’s not required, we would be very appreciative if you would add an IntelliJ IDEA banner to your project’s site in support of IntelliJ IDEA.

[snip]

This is absolutely not required. This is a no-strings-attached license, but we would be very grateful for any help leading people to info about IntelliJ IDEA!

Again… welcome!

Best regards,

Ilia Dumov
Product Manager
JetBrains, Inc

So I went ahead and added the banner to our site and all was right with the world.

The license was good for one year, and so in 2008 he renewed, again with no problem. But this year he received quite a different response:

Hi Johan,

We can’t give you free license because it will be used on your paid support services engagements. You can buy commercial licenses if you want.

All the best,

Victoria Dumova
OS Support Program Manager
JetBrains, Inc

I never tire of pointing out that the OpenNMS Project is independent from the commercial services company, The OpenNMS Group. Johan is not an employee of the commercial business, so he doesn’t perform “paid support services engagements”. In fact, his involvement in the project has not changed in the last two years but for some reason he didn’t qualify for a free license this time. Furthermore, since 100% of OpenNMS is free and open software, what would it matter? The IDE is used to help develop code, and all of that work gets released back to the community.

I must wonder what types of projects qualify as open source for JetBrains. Only those that make no money whatsoever? I can see the next e-mail: “Sorry, your license will be used to get donations on Sourceforge so you have to buy one.”

My point here is not to bash JetBrains. To complain about not getting a free license is like complaining that the bisque in a free soup kitchen is a bit salty. But it does illustrate the dangers of commercial software.

In two years we went from “This is a no-strings-attached license” to “You can buy commercial licenses”. The rules changed. Luckily for us we can probably get Johan to use Eclipse now, but what if we depended on IDEA? We’d be screwed.

With open source software the power lies in the user, not the provider. With commercial and open core software companies the revenue model is to sell licenses, and thus to maximize profit these companies are motivated to increase license revenue. This may mean selling to you licenses at a discount to get you using the product, only to change the rules a couple of years down the road.

If you are a decision maker in your company, I think you owe it to your employer and your shareholders to question any commercial software purchase. Are you willing to base your operations on software that may double in price without warning? Maybe the vendor will go out of business, leaving the code in limbo, and what will you do then?

Transitioning to open source is not easy. Although the software is free, there is a cost in time, perhaps consulting services and in getting your staff up to speed on the product. But in the long run the cost is worth it, if just to lose the reliance on outside vendors who, as this situation demonstrates, can be very fickle.

How Much is that Project in the Window?

Posted on by tarus

It’s snowing here today. For many who live in cold climates you are probably saying “big deal” but it is rare to have snow where I live and this is the most significant snowfall we’ve had in four years. By the end of the day we could have up to six inches (15 cm).


Sortova Farm in the snow

Since it happens so rarely we don’t have the equipment to clear the roads quickly so everything pretty much shuts down. At the OpenNMS Group we’re used to working from home several days a week so it will be business as usual, but I did cancel two meetings today and it frees me up to watch the coronation at noon.

It also allowed me some time to browse through my news feeds, and I came across an interesting post by the 451 Group.

I like the 451 Group, mainly because they realize that OpenNMS exists (grin). In addition, they seem to have a good understanding of the open source marketplace.

Today Matthew Aslett wrote that Intalio has raised funds to acquire 8 to 10 open source companies in the next year. It was interesting to me for a number of reasons. Most open source business models that I’m aware of have being acquired as the exit strategy, but many of them take the VC route to get there. Most VCs are looking at a big payout, say north of US$100 million, but the number of companies that can spend US$100 million on an acquisition is small. However, the number of companies that can spend, say, US$10 million, is much larger, so it would seem that the safer way to get acquired would be to either bootstrap the company (like we did) or get a small amount of angel funding to get started and to keep the company small and focused. Plus the model of acquiring smaller companies worked well for JBoss.

Another reason I found this post interesting was Intalio’s list of requirements for a possible target:

Small (25 employees is a maximum, less than 10 is preferred, mainly engineers)

The OpenNMS Group has six employees, and half of those were added in the last 18 months. We are all technical – there are no full time sales or marketing people.

Open Source or ready to go the COSMO way

Duh. We’re free and open to a fault.

Exceptional technology that took many person-years to develop

OpenNMS has been around since 1999. A tremendous amount of work has gone in to building it, both from the commercial side as well as the community side.

Architecture compatible with the one built for Intalio|BPP

From what I can tell, OpenNMS would be a good fit. We’re written in Java utilizing a lot of the new enterprise technologies such as Spring and Hibernate.

Support for industry standards (J2EE, WS-*, etc.)

Got that. The newest OpenNMS code adds REST-style interfaces, we support most industry standards (SNMP/WMI/TL1/HTTP), and there is always lots of XML.

Active user base (the larger, the better)

With 50,000 unique visitors to http://www.opennms.org and 5000 downloads each month I think we can check that off.

Committed customer base (the larger, the better, but small is OK too)

We’ve had over 100 customers in 18 countries, with over 50 current support customers, and our first commercial support customer (from December 2001) is still a client.

Committed employee base (location irrelevant, we’re in 13 locations already)

I’ve never worked with a better group of people. I think “committed” is too weak a word to describe their dedication to both the company and the project. While most of us are near RTP in North Carolina, Jeff is in Atlanta, so the fact that location is not important is a plus.

Profitable or break-even

Yup, got that. Since our business model is “spend less than you earn” we’ve been profitable since our first day in business.

Little or no debt

Heh, since we’re bootstrapped the opportunities for debt have been pretty limited (banks are a little wary of loaning money to companies with no liquid assets). So while that has caused us to grow more slowly than I would like, the company remains closely held and well positioned to ride out the current economy.

So it looks like we’re a pretty good fit, which gives me a little more confidence that what we are doing is the right thing (remember, I’m making most of this up as I go along so any reassurance is nice).

Since snow gives me time to reflect a bit, I’ve been thinking about what kind of company I’d be willing to allow to acquire us. They’d have to be doing cool things and provide a great work environment. They’d also have to have a serious interest in investing in the OpenNMS project itself, versus just buying it for the name or to get us out of the way. Their business would have to fit in with our mission statement of “Help Customers – Have Fun – Make Money”.

In exchange they would get the best bunch of guys on the planet. Seriously, each day these guys amaze me. They’d get a solid customer base and a mature, profitable product line.

One problem would be that they’d also get me. I’ve often wondered about my role should we ever be acquired. When I search Monster for “loud-mouthed, opinionated, free and open source bigot” I get no hits [Update: heh, actually I get eight]. My job at OpenNMS has always been to hire people smarter than me, which I’ve succeeded at beyond my wildest dreams, so next to them I start to lose my relevance.

I think the role I’d be best suited for would be Community Evangelist, or Vice President of Openness. I hate the term “community manager” because it seems to refer to the open source community as a resource to be manipulated, versus something to be nurtured and grown. Would Community Gardener work? God knows I’ve got the fertilizer (grin). I’d love to see if I could get as excited, and thus get others excited, in projects that compliment OpenNMS as I do with OpenNMS itself.

Fun thoughts for a cold day.

OpenNMS 2009 User Conference – Europe

Posted on by tarus

I am very excited about our first ever user conference. While we’ve had our annual Dev-Jam developers conference going on five years now, this will be the first event aimed at the end users of OpenNMS software.

It will be held on 14 March in Frankfurt, Germany, which will be the weekend after CeBIT. We are currently finalizing the details of the venue and cost. Our guess is the cost will be around 200€ and will include catering and a really nice OpenNMS polo shirt (although not a green one).

The format of the conference is to have several training seminars in the morning, break for lunch, and do more “real world” presentations in the afternoon.

The event is being sponsored by the OpenNMS Project, The OpenNMS Group and Nethinks. You mayremember Nethinks as the focus of a rant of mine from awhile back. We are now getting along famously and they have done a lot of the work to organize it. The other great contributor has been Alex Finger (OGP) who is taking the lead from the project side of things.

I’ll be there to talk about the project and to demonstrate the upcoming features in 1.8 (such as native WMI support). Jeff is coming to talk about OpenNMS and Asterisk, both from a “how to monitor Asterisk with OpenNMS” viewpoint as well as integrating OpenNMS notifications with the phone system [be sure to check out Jeff’s presentation at Asterisk World in February]. I believe a couple of other OGP members will be there as well, so it should be a great introduction for those who don’t know much about OpenNMS as well as a chance for current users of OpenNMS to learn how to use it better and to get their questions answered by the team that develops it.

In order to gauge interest we have a “pre-registration” form set up. If you think there is a greater than 50% that you can be there, please pre-register to help us plan a little better.

If I still have your attention, there will also be an OpenNMS seminar in Milan, Italy, on 4 February at the Sun Microsystems office. This is much more of a business oriented seminar. If you are interested in attending that, please drop Antonio or me a note.

I will be in Milan to help teach the OpenNMS training course that we are holding there. We have very few seats left so please register or let me know soon if you are interested.

Matt Raykowski, OGP

Posted on by tarus

Some days I absolutely love my job, and today is one of them. Today we inducted Matt Raykowski into the Order of the Green Polo (OGP). The OGP is the governing arm of the OpenNMS Project and it is made up of those people who have made (and continue to make) major contributions to the application.

This is what makes OpenNMS so great. Building on Matt Brozowski’s highly scalable collection engine, Matt Raykowski added native WMI support so that OpenNMS can now gather Windows performance data. I want to emphasize the word “native” since there is no proxy or additional software (such as the SNMP Informant agent) required to be installed.

There is still some ways to go before the feature is ready for the stable branch, but it should be included in the 1.8 release later this year. However, it is in trunk and thus people are encouraged to test it by either building from source or by installing the nightly “snapshot” packages. I am looking for some people with large Windows networks who can help us test its scalability. My guess is that it will be orders of magnitude greater than the other solutions that are out there.

When I go on and on about the ideals of true “open source” it is work like this that illustrates exactly what I mean. While there is a myth that just by calling your project “open source” thousands of qualified people will give up nights and weekends to work on it, it is possible to build, over time, a small but dedicated group of people who do just that. But to make that happen requires a level of trust that the work won’t be commercialized, that it will remain open and continue to be appreciated and developed for time to come.

This is the biggest flaw in the “open core” model and exactly what makes “open source” so powerful. Sure, the open core guys get contributions, but I doubt they get such powerful features. Even if they did, they’d probably want to make them “enterprise only” and try to sell them. At OpenNMS we encourage work like this and have built a platform to make such things possible, and a major part of that is keeping all of the software free and open.

I’m humbled that Matt trusted us enough to spend the months it took to create this feature, and I hope in at least a small way his membership in the OGP provides some compensation. I’m sure his dedication will continue to make the project even better, and move us closer to becoming the de facto network management platform of choice.

Awards and Honorable Mentions

Posted on by tarus

I’ve never been much for industry awards or the opinions of the various trade rags. I mean, when NetInfoCommWeek decides to compare open source network management platforms, we are never included, even though most of the time a Google search on “open source network management” has OpenNMS as the first hit. This despite spending zero money on marketing.

I’m not sure I’d even want them to check out our software, since quite often it is more of a beauty contest, and the “amazing” test lab with nearly 25 devices doesn’t exactly lend itself to demonstrating the strengths of OpenNMS, which starts to shine in the 2000 to 20,000 device range.

But I am often delighted when people who actually eat, live and breathe network management mention OpenNMS favorably. For example, when TechTarget interviewed 1300+ users and asked them what was their preferred network management platform, OpenNMS came out ahead of OpenView and Tivoli. This was from the viewpoint of people who actually use the software, not overworked and deadline driven reporters.

Last week we got a mention on Doug McClure’s list of 2009 Predictions as a possible contender for a Business Service Management (BSM) Lite platform. That was pretty cool, considering that the whole project is bootstrapped and community driven.

Doug’s is an opinion I value (I first met him at barcampESM and since he is in Atlanta Jeff gets to seem him occasionally), and his mention of our project was rather humbling. And I didn’t even blink at the word “Lite” since OpenNMS has a long way to go before we can replace all of OpenView and all of Tivoli.

We started this project to build a sustainable, free and open replacement for the major commercial management platforms. So of course we started on the most basic functions: the need for discovery and an inventory database, monitoring, event management and data collection. As was seen in the TechTarget award, this is enough for many folks.

However, we’d definitely like to move into the area of BSM, especially as more and more companies start looking at the utility or Software as a Service (SaaS) model. No one at the executive level of the company cares about the bandwidth or errors running through a router; all they want to know is that the widgets are rolling of the assembly line.

Another area I’d love to explore is the idea of runbook automation. One company that is doing a lot in this area but you rarely hear of is IPSoft out of New York. They claim that they can resolve 56% of problems automatically. The idea of capturing and automating the knowledge of experienced network and system admins is exciting to me, and seems to play into the main ideas of BSM.

But for now, as we enter into what may be a difficult time for all businesses, we will focus on remaining profitable and our mission to “help customers, have fun, and make money”. It appears to be working, as we reached a nice milestone last week. Our first customer, who joined us in December of 2001, renewed their support for the seventh time. The fact that someone who saw value in our product and our services so long ago still does means more to me than any magazine or trade award could.

Holiday Cards

Posted on by tarus

The OpenNMS office is full of festive cheer, made even merrier by some recent cards that have arrived.

One was from Tobi Oetiker, although it is in Swiss German so I’ll assume it’s positive (grin).

I got to meet Tobi many moons ago, and his RRDtool project drives a lot of open source tools as well as a number of closed source ones.

We also got a cute card from Mike Huot. He lives in the frozen mid-west and sent us “A Minnesota Christmas”.

Both have been enshrined on the amazing, wonderful Wall of Cards.