By Gary Tyreman, Univa CEO
Most of us know that Open-Source software is continuing to penetrate today’s enterprises with the vast majority using it in some capacity, but there appears to be a new shift in the usage model. At Univa, we are seeing a number of enterprises move towards a supported software model. This is not an indictment of open-source, we have been told, but rather a sign of the challenges an enterprise faces without the availability of suitable support for mission-critical applications. This decision usually derives from senior management who understand that the risk of suffering downtime without the availability of support is unfathomable, renewing the focus on IT’s primary responsibility: supporting business users.
It was this information that motivated us to commission an independent study to obtain insights from enterprise end-users at the director level and above. The key answer we sought concerned the motivation to add support to Open-Source and the factors that would engage that discussion. What we discovered was an interesting mix of both the expected and the surprising.
As expected, our 2013 Open-Source Software Use survey found that Free and Open-Source software (FOSS) is prominent within businesses today, with 76% of companies using FOSS. It’s important to note that we only surveyed commercial entities, and then only more senior decision makers. This was to accomplish two things: first, to receive a business owner’s view, and second, to take the discussion away from one of technical merit and to direct insight into the thought processes around business risks.
What we found surprising was the high number of respondents that admitted to having experienced a problem with using FOSS, with 75% admitting as such. Clearly businesses are relying heavily on unsupported Open-Source solutions today, running the risk of using FOSS without the adequate support to call upon in their hour of need. It’s like waiting to buy fire insurance until after your house is on fire.
In our own business, Univa, we convert Open-Source users of Grid Engine, a distributed resource management software used across the globe for HPC and Big Data environments (a former Sun Open-Source development project), to our binaries and support. Yet, the percentage willing to pay for support that we have experienced with Univa prospects is much higher than averages that have been published prior to our survey. Our FOSS survey supported this fact with 64% saying that they would pay for supported software if it solved their problems moving forward. We’ll cover some of the rationale a little later in this article. First, let’s layout the rest of the survey findings.
View Infographic of the full survey Here
The lack of enterprise-grade support was the largest problem enterprise FOSS users experienced in their company with 27% of respondents raising it as their top concern. Other troublesome issues include usability (24%), maintenance (20%), crashes (19%), bugs (18%), downtime (16%), loss off productivity (16%) and interoperability (16%).
For those companies willing to pay for better quality, the following were listed as the key reasons to do so:
- Stability (25%)
- Enterprise-grade support (22%)
- Ease of use (20%)
- Extra functionality (18%)
- Bug reports/fixes (15%)
- Integrated solution (13%)
- Product upgrades (13%)
- Predictable lifecycles (13%)
The demand for stability and enterprise-grade support suggests that the comfort of reliability and quick access to know-how is the number one factor that leads to invest beyond Open-Source solutions. The key product development departments of a business where most mission-critical software resides – Engineering and R&D – rely most heavily on FOSS (32%).
This supports our experience. Univa’s software typically resides in the departments within large enterprises that drive product and service innovation such as engineering and R&D, which depend on expensive mission-critical applications. Whether that’s in an oil & gas company running reservoir analyses, a fabless semiconductor company trying to tape-out a new design or an automotive manufacturer simulating the crash-test of a $200,000 car, the risk of using FOSS without support is connected directly to the top line of business revenue.
Our product is middleware that sits between the operating system and the application. Since the majority of our revenue comes from the Engineering and R&D Departments’ IT budget, we can also draw a connection to the behavior across the broader software stack. Our customers typically spend 10 to 20 times the amount on commercial application licenses that “require” a supported operating system such as Red Hat Enterprise Linux. It’s also common for these application vendors to certify workload management systems and specific configurations.
This is clearly different than some areas of the HPC ecosystem where applications are home-grown, so the risk and cost-benefit are not the same. In these situations we continue to see un-supported users across the entire software stack.
The key take away from this exercise is that while Open-Source is for everyone, it’s not for everything (or every use case) without an immediate path to quality assurances (stability) and enterprise-grade support, particularly when you consider areas in the enterprise where mission-critical applications require both supported stacks and configurations. What becomes a concern is the business outcome in the event of a failure when the application vendor denies support due to an un-supported configuration. That denial of service could be catastrophic and it’s why it’s better to buy insurance before you need it.
Gary Tyreman is the President and CEO of Univa Corporation the leading provider of enterprise workload and resource management solutions built around Grid Engine.
Survey Methodology
Sample was provided by uSamp, a premier provider of technology and survey respondents used to obtain consumer and business insights. This project was fielded from 3/14/2013 through 3/18/2013, collecting 128 completes from uSamp’s Whiteboard B2B panel. This online survey is not based on a probability sample and therefore no estimate of theoretical sampling error can be calculated.