When Oracle decided to fork Red Hat Enterprise Linux, it did so under the banner of offering improved support and lower costs for its customers. Oracle Enterprise Linux has failed to set the world on fire, perhaps in part, as Gartner highlights in a new research report ("Red Hat vs. Oracle Linux Support: When and How Does It Matter?"), because its cost and support advantages appear to be overstated.
Oracle Enterprise Linux
(Credit: Oracle)It's hard to compete on supporting code when you write and influence much less of it.
But Oracle can charge less, right? Yes, it can. Gartner reports: "Some 90 percent of users that have gotten a price quote on Linux from Oracle claim that Oracle has lower support fees due to aggressive discounts, with most users saying the fees are 50 percent below Red Hat."
A 50 percent discount sounds great, right?
Maybe. That discount comes at a significant cost, for several reasons that Gartner points out:
- Although Oracle claims that it maintains lockstep on RHEL [Red Hat Enterprise Linux] code, and applications need not be separately certified to Oracle if certified for RHEL, in reality, users bear some risk to assure quality of service. [Translation: Oracle Enterprise Linux (OEL) customers have to invest extra budget to ensure their applications will run seamlessly on OEL.]
- OEL begins with the source code of a RHEL version as its base....However, if a priority fix is required for a problem in RHEL code in a complex configuration, then there is the potential for temporary forking....Thus, Oracle users are likely to have a different patch from a Red Hat-supported configuration, at least temporarily....Of course, users need to go through regression testing to make sure that the temporary fix has not created potential problems. This puts Oracle in the position of quickly reacting to problems of severe priority; however, users must monitor the fixes closely. [Translation: OEL customers must do regression testing and closely monitor OEL fixes to ensure they'll work.]
- Users with RHEL contracts will face significant challenges in reducing support costs by selecting specific Red Hat-supported servers and not others. [Translation: Because of Red Hat's pricing/contractual terms, it becomes difficult to mix-and-match support for some servers while not for others, so it may be hard to find cost savings this way.]
According to Gartner, it is this last item that may be driving the most interest in OEL, because Oracle doesn't adopt Red Hat's "all-or-none" approach to Linux server support. Even so, Gartner only finds 20 percent of Red Hat customers surveyed are annoyed by its contract policies. That's a significant number, but not overwhelming.
Given Oracle's aggressive discounts, including its willingness to buy out the remainder of a customer's Red Hat contract, as Gartner reports, one would think that it would be flying high. But it's not. OEL has roughly 2 percent market share, compared to Red Hat's 60 percent market share. Enterprises aren't stupid: if Oracle's Linux story were compelling, they'd be buying. But they're clearly not.
Not that Red Hat is in the clear. Its biggest weakness against Oracle and others is that Oracle can sell a much more complete story to the CIO. Red Hat, to the extent that it talks to the CIO at all, is a tactical decision. It's an important tactical decision, but big contracts are awarded to companies like SAP that promise dramatic competitive differentiation through optimized business processes.
The best Red Hat can do is improve performance and lower costs on server infrastructure. This isn't insignificant, but it's not something that gives Red Hat an impregnable hold on a CIO's attention and wallet. While I've only seen one defection to OEL in the past two years within Alfresco's customer base, I have seen an increasing number shifting to alternatives like CentOS.
In sum, while Oracle's Linux play isn't yet serious competition to Red Hat, it does establish just how limited Red Hat's position is.
Follow me on Twitter @mjasay.
Over the weekend I struggled to get Flash working on my Ubuntu 9.04 Netbook Remix build. I turned to Twitter and Google for support, and was dismayed by the response.
Some, like Canonical's Jono Bacon, were very helpful. The rest offered the somewhat standard Linux supporter's line:
"What are you talking about? Linux is as easy as Windows."
After trying to get Flash installed for hours - using Firefox's extensions directory, the command line, and everything else I could find - this wasn't super helpful.
Not that I'm much more helpful when others ask me for assistance. I'm a Mac user, and my typical response tends to be,
"What are you talking about? Mac OS X is even easier than Windows!"
In either case, telling the drowning woman that it's really easy to swim ("Just type "apt-get swimming lessons") or, worse, yelling at her for incompetence ("I can't believe you're telling me that you don't know what to do with a .dmg file! It's soooo much easier than Windows!"), really doesn't help to win people over to Linux or the Mac. Learning a new operating system is like learning a new language: talking louder doesn't improve communication.
Here's a hint for the Mac and Linux faithful: you won't convert Windows users by talking down to them. Focus on the positive aspects of your own operating environments and then demonstrate empathy and patience while showing newbies how to get around on Linux or Mac OS X.
Using such means they, too, will come to have the same superiority complex that we Mac and Linux users have.
Follow me on Twitter @mjasay.
The Web is buzzing with this genius workaround Microsoft devised for an Oracle problem while using Excel. I had no idea my great aunt works in Microsoft's technical support team, as this is just the sort of solution she'd likely devise:
Method 2: Move Your Mouse Pointer
If you move your mouse pointer continuously while the data is being returned to Microsoft Excel, the query may not fail. Do not stop moving the mouse until all the data has been returned to Microsoft Excel.
NOTE: Depending on your query, it may take several minutes to return the results of your query to the worksheet.
My question: when can I stop? I've been moving the mouse pointer around for a few days now, but the problem persists. ;-)
Follow me on Twitter @mjasay.
Vinnie Mirchandani writes provocatively on his Deal Architect blog that Oracle's penchant for acquisitions may be hampering its ability to innovate in-house and write code:
In existing Oracle internally developed and acquired products in the last five years, Oracle's enhancements have been anemic.
Oracle, in my opinion, has forgotten how to develop code. Its top executives are deal makers, not technology visionaries. Worse, when it comes to their acquisitions, they cannot retain or easily replace the entrepreneurial talent...The rapid pace of acquisitions has also had a significant impact on Oracle support.
Customers report frequent and confusing changes to Oracle's support policies, as so many products go in and out of stages of "lifetime support." Little has been done to rationalize support across products--other than, of course, raise maintenance to 22 percent.
Arguably, it simply does not matter: even through acquisitions, Oracle has managed to deliver increased value to its customers.
Even so, a big question is looming as to whether Oracle, which spends just 10 to 15 percent of its budget on research and development, can keep up with competitors and, in particular, open source.
"Open source?!" you say, "that's just a big commodifier of others' innovations." Not so. In fact, if you look at the budgets of most emerging open-source companies, we spend significantly more on development than Oracle and, importantly, more of that R&D budget goes toward real innovation, not reinventing the wheel. Indeed, that's the whole premise behind open-source development: efficient reuse of code.
That's not the whole story, however. As Mirchandani points out in a follow-on post, Oracle customers are troubled by its support morass. Such customers are likely to be enticed by open-source offerings, which make support, not license fees, the centerpiece of their offerings.
Oracle has made a lot of noise about its Unbreakable Linux and has suggested that the impetus behind its Red Hat knock-off is that its customers couldn't find Oracle-class support at Red Hat.
Well, Oracle is probably right, but not in the way it intended: Red Hat's support and value proposition is apparently much better than Oracle's, given that Red Hat continues to leave Unbreakable Linux in the dust.
Oracle is a great company that continues to make smart moves. However, it is becoming more and more like Microsoft: less about innovation and more about distribution. That's not a bad thing. Not at all. But Mirchandani's comments point to a few areas, like support, where Oracle still needs to get its execution right for the strategy to pay off in the long term.
Follow me on Twitter at mjasay.
TechFlash is reporting that EMC has purchased SourceLabs for an undisclosed fee. The unanswered question in TechFlash's report is why EMC would buy SourceLabs, a provider of support tools for Linux and other open-source software.
It's not that SourceLabs isn't a good company. I have followed SourceLabs since its inception, meeting with founder and CEO Byron Sebastian back at OSCON (in 2003) before the company was founded in 2004, and spent some time in the SourceLabs office in 2004 getting a demo of its technology. It was cool back in 2004, and has improved since then.
In addition, SourceLabs has managed to pull in some big-name customers like Fidelity and Merrill Lynch, demonstrating that it offers real value to real customers. Unlike some competitors like Spikesource, SourceLabs focused early on big enterprise needs and arguably did a better job of tailoring its products to meet those needs than its competitors, notwithstanding its share of financial struggles, which TechFlash details.
No, my question is what EMC, largely a provider of storage solutions, gets from a relatively broad-based open-source support technology company. TechFlash points to Swik.net, SourceLabs' open-source news and information repository as a source of value for the company, but I'm guessing that EMC didn't buy SourceLabs for Swik.net, given that the company had been contacting potential buyers just a few weeks ago to gauge interest in buying Swik.net, likely because EMC wasn't interested in that part of SourceLabs' business.
No, I don't think EMC is interested in an open-source community site, but it's clearly interested in the core SourceLabs technology. I'm struggling to understand the fit. Is EMC hoping to tap into expertise in various open-source technologies? If so, to what end?
I've asked SourceLabs' executive team for comment but, in the interim, anyone care to venture a guess as to what EMC is hoping to get from SourceLabs?
Red Hat has set the standard for world class software support, consistently earning top marks with CIOs for its efforts. On Thursday, however, Red Hat outdid itself, introducing a new product support plan called Extended Update Support. In a nutshell, Extended Update Support enables customers to run their mission-critical systems for longer stretches of time without having to take production systems offline to update them.
From the announcement:
Extended Update Support allows a customer with a large mission-critical deployment to reduce server administration and management costs by standardizing on a single update release of Red Hat Enterprise Linux for up to 18 months--all while preserving stability and data security.
As Red Hat explains, most software companies allow customers to standardize on a minor, "point" release for 6 to 9 months, or at most 12 months. Through its Extended Update Support program, however, Red Hat is letting customers pick a Red Hat Enterprise Linux build and stick with it for up to 18 months, up to three times the industry average. That means less downtime and less need to re-validate software stacks running on RHEL.
The Register provides some additional insight:
While Red Hat commits seven years of support for a major RHEL version, the dot releases within the versions change about every six months. Within those dot releases, the company ensures application compatibility because it doesn't change the runtime environment, the area where the Linux kernel interacts with applications. So even if there are patches for security or bugs and whatnot in the dot release, customers do not have to go through application testing and certification, which can take many months, as long as they stay within a RHEL version.
This is a great service to Red Hat's customers, and provides further evidence that Red Hat's subscription model helps it to be more attuned to customer needs. Red Hat isn't selling an upfront license: it's selling the continued value of an ongoing subscription. By tuning that value to actual customer needs--in this case, the need to disturb production systems as little as possible to reduce risk and save money--Red Hat ensures renewals.
Subscription models align vendor interests with customer interests. Red Hat's Extended Update Support for Red Hat Enterprise Linux is setting the pace. It will be interesting to see who follows.
Sometimes "free" is not so free.
I recently discovered this when a large, global system integrator (SI) deployed Alfresco Labs, our free and unsupported product, for a large client in Europe. The SI wasn't a partner of ours, and as the client soon learned when its deployment stumbled, the SI wasn't capable of providing enterprise-class support on the product. Yes, it knew the product well enough to deploy it and get paid over $50 million for its trouble, but when the deployment hit a glitch, guess to whom the SI came crawling for help?
It's not just my company. I know of another global SI that has deployed well over 100 Mule ESB instances, without buying support through MuleSource for its clients for a single one of them. If something goes wrong with those installations, the enterprise clients are going to end up paying a premium for the SI to figure out how to resolve the problems on the client's dime, never mind potential indemnification issues.
Not all SIs act like this, at least not all the time. My own company works closely with Satyam, SAIC, Booz Allen Hamilton, and others, and Accenture sells supported instances of the Spring Framework, but this is the exception to the rule for the large SIs, many of which seem happy to deploy open-source software for their clients without buying support or production-grade versions of the software.
Such SIs seem to believe that life has started raining free lunches.
This is a myopic way to do business, as the large SI in my initial example found: in that example, spending $50,000 (in the midst of a $50 million project) would have saved the SI the embarrassment and cost of trying to support a product that experience proved it didn't know nearly as well as it thought it did. The SI risked the success of a $50 million project to boost its margins by $50,000, only to find that one problem with the software ended up costing it and the client far more than $50,000.
If you're an enterprise looking for a strong SI on a project, here are a few things to consider:
... Read moreAs usual, ZDNet's Mary Jo Foley delivers an interesting bit of news from Redmond. Microsoft, it would seem, launched a significant new support offering as of August 18.
(T)he new offering, Premier Ultimate, is aimed at the company's largest customers. Via the new program, Microsoft works with customers to create a three-year roadmap, including a suggested set of services. Reactive support becomes something provided on an "as needed" basis, said Charlie DeJong, general manager of Support and Health Services for Microsoft.
Customers who sign up for Premier Ultimate get unlimited problem resolution support (with some unspecified possible restrictions), plus IT health assessments, account management (both on-site and dedicated) and on-site support, 24 hours a day, seven days a week, according to Microsoft.
This sounds like an impressive support offering, in my opinion. The unlimited part isn't all that great, given that most other vendors offer this (including all of the open-source companies with which I'm familiar).
No, it's the consultative approach to support that perhaps breaks new ground. Good for Microsoft for doing this.
Forrester finds that European enterprises cite support as their biggest reason for not adopting open-source software. This has persisted for years, with support (or, a lack thereof) consistently listed as one of the top reasons that enterprises throughout the world avoid open source.
The ironic thing is that open-source companies primarily sell support, not software. So...while proprietary-software vendors sell licenses with support as an afterthought, enterprises don't seem to question that they're going to get support. At the same time, open-source companies sell support with licenses as an afterthought...and enterprise buyers worry that they won't get support.
Huh?
Ultimately, I think this is a question of the relative maturity of the open-source market. Buyers equate the size of their vendor with the availability of support, and most open-source vendors are still tiny compared to Oracle, SAP, IBM, and Microsoft.
What they may be missing, however, is that a dedicated open-source vendor that makes its money selling support may well be a better source of support than a large vendor for whom support isn't its primary revenue stream.
I don't want to suggest that open-source support is always better than that provided by proprietary vendors. Oracle, as just one example and despite widespread grousing, consistently wins awards for its support.
No, I'm just suggesting that stifling your company's open-source adoption because of a perceived lack of support is silly and outdated. Welcome to the 21st Century. Open-source vendors provide support as good or better than their proprietary peers. Really.
Microsoft does many things wrong, but one thing it does very well is long-term support for its products. As a case in point, Windows for Workgroups, born to the world in 1993, is finally being retired on November 1, 2008. Fifteen years...Most software companies haven't even been in existence that long.
The only way that open source can replicate this, given commercial open source's young age, is through "perpetual escrow." Even so, if I'm an IT buyer it matters to me that Microsoft, IBM, etc. have been around a long time, and will be around for an even longer time.
Now if I could just get vendor longevity and open source at the same time, in the same place, I'd be golden.
