DATA CENTERS

  • 11/20/2014
    11:38 AM
  • Rating: 
    0 votes
    +
    Vote up!
    -
    Vote down!

Open Source Vs. Proprietary: Time For A New Manifesto

Philosophical debates over OpenStack vs. vCloud or OpenDaylight vs. Cisco ACI miss the point. For IT, ideological purity is neither possible nor desirable
Get the new issue of Network Computing, distributed in an all-digital format.

Contention between open source groups freely releasing code and commercial vendors capitalizing on proprietary products started the minute software became a profit-generating industry. The latest battleground? Enterprise data centers, with fights focused on cloud stacks, software-controlled networks, and big data systems. The lines are far from clear-cut. Established IT vendors incorporate open source code, APIs, and standards in their products. Startup companies are happy to commercialize public open source projects if they see demand. IT must take a pragmatic approach to software and vendor selection.

It's business, not dogma.

Unfortunately, human (and especially, engineer) nature being what it is, debates tend to take on a moralistic -- and occasionally anti-capitalist -- tenor, conflating open source software with idealized notions of liberty, community, and creative license. The tone was set early on by open source evangelist Richard Stallman, father of the Emacs editor and many other OS subsystems that were eventually subsumed by Linux. From Stallman's Gnu Manifesto:

"I consider that the Golden Rule requires that if I like a program I must share it with other people who like it. Software sellers want to divide the users and conquer them, making each user agree not to share with others. I refuse to break solidarity with other users in this way. I cannot in good conscience sign a nondisclosure agreement or a software license agreement."

Figure 1:

There are also cold, hard business and technical reasons why an open source development model makes sense, articulated by that other famous philosopher of the movement, Eric Raymond. Chief among them: software quality, security, reliability, and speed of development. In Raymond's view, open source replaces Brook's Law, which holds that adding manpower to a late software project makes it later, with Linus's Law: "Given enough eyeballs, all bugs are shallow."

Software vendors counter that without proprietary software and intellectual property protections, there's little incentive to invest and innovate. Both sides make valid points, and the fact is, open source projects like OpenStack, Open Daylight, Hadoop, and even core operating systems such as Linux and Android have been adopted and are funded by large IT vendors.

Windows versus Linux is the longest-running skirmish between open source and proprietary software. It's also now largely moot. As virtual machines become the default application environment -- and our latest InformationWeek State of the Data Center Survey finds 64% of respondents will have half or more of their production servers virtualized by the end of 2015 -- the underlying guest OS doesn't much matter. But as one area of contention subsides, many more arise. Technologies such as private cloud stacks (OpenStack, vCloud) and application containerization on low-power, non-x86 systems could shift the server balance. Open source software is making headway in three critical areas important to people running enterprise data centers: cloud stacks, software-defined networks, and big data platforms.

Let's look at each.

Cloud stacks: VMware still the default

Cloud stacks are perhaps the most visible software category in the data center for which open source presents a viable challenge to proprietary systems. Red Hat CEO Jim Whitehurst rightly characterizes the cloud as a disruptive force. "And in those paradigm shifts, generally new winners emerge," says Whitehurst. If Red Hat and other advocates have their way, OpenStack will be the standard cloud platform for service providers and enterprises alike.

Despite abundant media hype, however, OpenStack faces an uphill battle. VMware is the de facto virtualization standard for almost two-thirds of enterprise IT organizations, and the company is building on that foundation. Although VMware faces competition on two fronts -- Microsoft Azure (proprietary) and OpenStack (open source) -- make no mistake, within the enterprise, it's still a David vs. Goliath situation.

A recent InformationWeek survey of those using, piloting, or developing a hybrid cloud found OpenStack trailing VMware by 30 points. Although OpenStack has made progress, with nearly one-third of respondents using an OpenStack variant, its penetration remains almost identical to Azure's. Results from our 2014 Private Cloud Survey are less kind to the open source cloud stack. While 65% of respondents use VMware as part of their cloud designs, only 18% check in with Red Hat, and a mere 7% use OpenStack.

Figure 2:

The data isn't surprising when you consider that, for many enterprises, a private cloud featuring such niceties as self-service provisioning and automated workload migration is an extension of the existing virtual server infrastructure -- over which VMware has held sway since effectively creating the software category (at least on x86 platforms) a decade ago. And VMware isn't sitting still. The company is aggressively developing its own cloud stack replete with virtual networking and storage, infrastructure automation, workload orchestration, and public-private cloud integration. In fact, in an exclusive interview after his Interop keynote, CEO Pat Gelsinger


Comments

Open Source

Open Source is everywhere, from the servers that enable networks to function, to the networks themselves and back to the servers that piece together data. I feel it would be safe to say that open source is heavily present at the network edge as well, on non-interface devices, etc. 

One reason that makes open source a success is that it not only eliminates lock-in, but the same process enables specialization and division as business can invest in their core competency -- investments that require a stream of revenue to flow in-wards, in the long term. 

Re: Open Source

I agree to Brian here, but one thing always surprise me for Open Source and that is license, how and when license will be deployed, what if we are using service from different vendors, how their copyright will me maintained.

Re: Open Source

One of the main misconceptions about open source software is the fact the software may not be updated and source of support. These things have actually been addressed, though open source still faces challenges.

Re: Open Source

"I agree to Brian here, but one thing always surprise me for Open Source and that is license, how and when license will be deployed, what if we are using service from different vendors, how their copyright will me maintained."

Aditshar, now a day's some of the startup companies and independent vendors are bundling license with support and installation, for open source products. you can seek their help for a better idea.

Hyper-V

Hyper-V has improved considerably, but considering both the numerous complaints many IT people once had about it (and, in some cases, still do) and Hyper-V's relative youth, I think it's really saying something that Microsoft has more hypervisor marketshare than KVM and Xen combined.

Open source equals vendor checks and balances

In addition to Richard Stallman's near religious devotion to the principles of the Free Software Foundation, there are many practical reasons to consider and support open source. With many interests converging on a najor project, such as Linux or OpenStack, there's little likelihood that any one can seize control of it. It's the system of checks and balances applied to software.

Re: Open source equals vendor checks and balances

"In addition to Richard Stallman's near religious devotion to the principles of the Free Software Foundation, there are many practical reasons to consider and support open source. With many interests converging on a najor project, such as Linux or OpenStack, there's little likelihood that any one can seize control of it. It's the system of checks and balances applied to software."

Charlie, there is no doubt that open source movements are gaining momentum across the world. The best part is there is no monopolistic nature with open source products.

Re: Open Source Vs. Proprietary: Time For A New Manifesto

Wow, there's a lot to chew on here. The topics on offer here do inter-relate a lot, and it's about time somebody condensed all these different issues together and looked at how they affect one another. You can't look at upcoming SDN technologies in a vacuum without considering how they'll be affected by incumbent Cloud standards. Most enterprises are inclined to follow the path of least resistance, and that has a lot to do with VMware's (and Cisco's, etc.) success - if they have a lot of VMware infrastructure in place, and VMware makes a product that covers the next big thing, they're inclined to buy that unless there's a compelling reason not to.

I am an open-source idealist myself at heart, but I think you're right on the money to say that the lines are becoming blurred in a largely favorable way, Kurt. I think this has a lot to do with the horsepower that's at the modern IT pro's fingertips. Open Source has a leg up because of it's reduced cost, ease of installation, and interoperability. If you have the power to do something with one click, you'd like to, rather than dealing with red tape from a vendor, feature lockout, or compatibility problems. That said, that same environment is conducive to this kind of hybridization, and we on the receiving end have very little to lose.

I'd like to give a thumbs up to your points on Big Data services as well. I was surprised to see Hadoop with numbers as low as they were, but your explanation makes a lot of sense. A lot of enterprises may have no need of it, and after all, 'Big Data' has no more strict a definition than 'Software-Defined Networking'. It is important to bear in mind that we're talking about real companies in the real world - many of them are infinitely smaller than your Googles or your Amazons. They're not inclined to be told what they should use - they're inclined to seek out what they actually need to solve problems.