IT Ops & OSS: The Road Less Traveled Becoming the Beaten Path?

April 2 2012 | By | in Open Source

It’s much more common today to see OSS in the corporate and government environments, but why? What causes an organization to convert from the follow the leader approach, to the road less traveled? Recently, I read a research paper on Organizational Adoption of Open Source Software* that provided some possible insight on this topic. With the recent announcement that Red Hat has surpassed the billion dollar mark and Floyd’s great article about the need for Red Hat and OpenStack to dance, I thought I would keep the OSS topic alive with factors that drive OSS adoption. Even though I found the data set  they used to be flawed**, invalidating their conclusion, the paper did ask some interesting questions I want to highlight.

Is the selection of proprietary software or OSS subject to network effects?

This is an interesting question and in my opinion, will depend on the size of the organization. What I have personally found is that smaller organizations with limited IT staff will not diversify the organization’s technology to the same extent that larger organizations will. If the organization’s staff are heavy OSS users, the IT support systems will also follow that technology trac. However, if the organization’s users are proprietary software heavy, the IT department tends to lean to the same supporting systems. So on a small scale, the answer to this question would be Yes, the network effects do decide between proprietary software or OSS deployments. On the other side of that coin, large organizations tend to diversify their technology deployments, regardless of the network effects.  In these cases, I’ve found the IT department was much more diverse in their technology prowess and had the human capital to support such environments.

Is OSS adoption associated with financial operating efficiencies?

Of course it is! Most companies I’ve worked for and with are focused on the total cost of ownership of any software product they deploy. Most software adoption talks start with the TCO question, and not just software cost, but human capital. Larger, more diverse and stable organizations are open to deploying the OSS solutions. These organization are able to capitalize on their IT departments diverse skill set.

How is IT usage intensity affecting OSS adoption?

As I have alluded to in the two previous responses, this one department has the highest impact on the adoption of OSS for an organization. It should go without saying that if this department doesn’t have the skill set necessary to support OSS, then its adoption will be very limited, if at all.

According to the research paper, in the last three years, Fortune 1000 companies have moved to deploying more OSS systems year after year. In the three years they tracked data, they showed an increase of 24% in adoption of OSS applications and a 9% increase in OSS operating systems. The trend shows that OSS has been established as a legitimate technological choice for organization of all sizes. It has been established by large organizations that best practices include a diverse IT department and diverse software platforms. The timing of this research could not have been more relevant than at this moment in time, the year an Open Source company amassed a billion dollars. Mark this as the beginning of the end for proprietary technologies being revered as the sacred cornerstone of technology for large scale operations.

*Diomidis Spinellis and Vaggelis Giannikas. Organizational adoption of open source software. Journal of Systems and Software, 85(3):666–682, March 2012. (doi:10.1016/j.jss.2011.09.037)

** Why do I say the data was flawed? They based their research on four distinct areas Web Server Software, Web Server OS, Client OS and Client web browser. They obtained this data by a combination of parsing web server logs and network probe tools such as NMAP. Though these methods will provide what seems like great data on the surface many variables are in play here that will present invalid data such as load balancers, proxy servers, security templates and client privacy tools. Nowhere in the research did they say they took these factors into account.

Photo from Ann Douglas


[adrotate block="1"]