09 January 2015

Open Source Software for Softwarization

Last post has argued that open source software can be considered one of the key “control variables” of “Softwarization” of network and service platforms. A lot of software is already there, today, but the point here is about decoupling it from hardware (as in SDN proposition), making it open source, and executing said software (e.g., implementing network functionalities) on virtual logical resources (as in NFV proposition) dynamically allocated/orchestrated on (even standard) hardware.

What will be the impact ? And how long would it take to become virally adopted ?

Looking at the past, there have been some examples of successful development and adoption of Open Source Software, producing a strong impact on the market, even in short span of time (e.g., less than 5 years).

Linux kernel is one prominent example: it is a Unix-like operating system released (around 1994) under the GNU General Public License version (GPLv2). The Linux kernel is used by a variety of operating systems based on it, which are usually in the form of Linux distributions. An example of “good enough” but simpler and cheaper operating system. The popularity of Linux virally diffused across developers and users. According to recent analysis from IDC, Linux servers now represent 28.5% of all server revenue (which means, given their lower costs with respect to other commercial servers, a much larger market share).

Apache is another example. Apache has been the most popular web server on the Internet since April 1996. Apache adoption diffused virally reaching about 60% of the market in about 5 years. Today, 385 million sites are now powered by Apache, landing today on a 37.45% share of the market.

Concerning open source OS for smartphone, in Q2 2014 Android rose to about 85% of the global market share. As known, most of Android is free and open source, even if a large amount of software on Android devices (such as Play Store, Google Search, Google Play Services, Google Music, and so on) are proprietary and licensed.  Interestingly, this piece of news mentioned the potential intention of an OTT to deploy SDN-like capabilities and protocols into Android (a way to extend the SD-x model to the very edge of the infrastructure, up to the Users’ mobile phones. Google does not charge a fee for companies that install Android on their devices. Nevertheless every installed copy of Android gives Google potential customers for its services and encourages Developers to build apps for the platform, this creating ecosystems around Android.

Coming to Cloud, SDN, NFV…we have several other recent examples. OpenStack is an open source platforms that can be used to provide Cloud services. OpenStack is being developed by the OpenStack Foundation, originally founded by NASA and RackSpace but now comprising a large group of technology companies. Several companies in the Foundation run public cloud services on the OpenStack platform (comparable with Amazon’s Elastic Computing Cloud EC2). OpenStack has been launched in 2010 and today dominates the market with a 69% adoption rate of IT Enterprises offering Cloud Computing services.

And, no need mentioning the long list of open source SDN controllers (see for example SDx). Moreover, a few weeks ago, ON.LAB has released in open source an operating systems (ON.OS) for SDN-like networks: after the release some Telco Network and Service Providers immediately announced that they will start testing it for assessing a future potential deployment in production environments.

In summary, my take (looking at past) is that if an open source software solution will be developed and released with the support of a critical mass of companies, it is likely that in about 5 years it will reach at least 60% adoption in the related market. And this could create a very strong impact on the related ecosystems.