November 2009


I recently spoke with Drupal founder and Acquia CTO and co-founder, Dries Buytaert. Dries gave me an update on Acquia’s success to date and plans for the future.

Many readers know Acquia as one of the key vendors behind the White House’s recent move to open source.

Acquia has achieved significant traction in its two years of existence. Well, Dries points out that the company spent a year establishing a sustainable business model, strategy and working within the Drupal community. According to Dries, Acquia only began revenue generating efforts a year ago.

Acquia currently counts over 300 customers; not bad after a year. As expected, Acquia’s customer base is currently weighted towards the SMB market, making up 50 percent of the total. However, large enterprises adopting Drupal and Acquia Drupal are the fastest growing segment, currently representing 30 percent of Acquia’s customers. The remaining 20 percent of customers are public sector, higher education and not for profit organizations.

What’s Acquia’s goal? Dries suggests: “To become to Drupal what Red Hat is to Linux.” To achieve this goal Acquia is executing against a three pronged strategy. First, Acquia sells support subscriptions to Acquia’s Drupal distribution which can be customized to customer needs, or to any collection of Drupal core and Drupal extensions. Second, Acquia recently announced Acquia Hosting. This offering was in response to customer requests to host and manage their Drupal implementation. Acquia Hosting is built on Amazon’s EC2. Finally, Acquia is working on Drupal Gardens. Set to release in the first half of 2010, Drupal Gardens will be a hosted offering much like WordPress.com. Drupal Gardens aims to reduce the time between “design to online in hours” as Dries puts it.  The demo is pretty cool, check it out.

Acquia definitely has an aggressive roadmap ahead. I look forward to catching up with Dries and team to hear about progress in 2010.

Follow me on twitter at: SavioRodrigues

PS: I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies or opinions.”

I used to work in market research, so I’m always skeptical of data until I’ve understood how it was measured.  That’s why I was surprised with this headline from TechCrunch: “Apple And Android Now Make Up 75 Percent Of U.S. Smartphone Web Traffic“. More specifically, I was surprised how large Android was in this reported data.  I absolutely believe Android will grow in significance in the future; today is a different story.

To the casual reader, that title of the TechCrunch report, along with the accompanying pie charts suggested that the iPhone and Android accounted for 55 percent and 20 percent of the US smartphone market.  These results position the iPhone and Android in number one and two in US smartphone web traffic.

Digging a little deeper, AdMob clearly explains how they arrive at the data:

“The report is based on the ad requests we receive from our network of more than 15,000 mobile Web sites and iPhone and Android applications.”

I couldn’t find details that split the “15,000” figure between mobile websites and mobile device-specific applications.   Mobile websites that serve information to any device seems like a logical way to measure mobile web traffic. On the other hand, iPhone and Android applications will definitely increase the web traffic counted in the iPhone and Android buckets accordingly. This is not to say that AdMob’s data or methodology is flawed.  Rather, it’s helpful to know what was actually measured and how.

I went back to the October 2008 results and found that over the past year, the number of mobile sites and applications has increased 150 percent from 6,000 to 15,000. I’d love to understand how the additional 9,000 “mobile sites and applications” added in 2009 split across mobile sites versus iPhone or Android applications. There has been an explosion in iPhone applications, so it’s not hard to assume that AdMob is tracking a higher percentage of iPhone applications in its data collection network in 2009 versus 2008.

Another factoid that surprised me, over the past year, the number of requests tracked by AdMob has increased 127 percent from 2.2 billion to 5 billion in the US. The mobile web is still in its infancy and it’ll be interesting to track the number of mobile request in a year. Oh, that and the percent of requests associated with Android phones!

Follow me on twitter at: SavioRodrigues

PS: I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies or opinions.”

But follow the niche alpha geek adoption carefully.

As with anything Google does, opinion ranges from revolutionary to lackluster.  Personally, I think it’s too early to tell.  More importantly, I think the success of Chrome OS won’t be based on the success of version 1.0.  Google has the uncanny ability to generate and maintain interest even in the face of negative initial reviews.

Chrome OS will be limited to netbooks, and more importantly, new netbooks that Google approves. Chrome OS is theoretically competition for Windows and Linux which represent approximately 80% and 20% of the operating system market for netbooks.  But Windows and Linux on netbooks allow a degree of user freedom that Chrome OS doesn’t.  Users can store files, be it pictures, songs, videos, spreadsheets, etc. on the netbook.  These files can be loaded, edited and saved with or without a network connection.  Chrome OS on the other hand, requires a network connection to access user files which are stored in the Google cloud.  This will be an impediment to Chrome OS adoption by average netbook consumers.  Rational or not, the fear of needing to get at files “in the cloud” but not having a Wifi/3G connection will diminish the allure of a netbook that starts in under 7 seconds to regular users.

On the other hand, geeks will be chomping at the bits to pick up a Chrome OS netbook to try out during the 2010 holiday season.  Yes, the “geek” audience is without a doubt a niche market.  So it’s easy for Microsoft or Apple to write off Chrome OS.  But that’s a mistake. As John Gruber wrote in his excellent piece, “Microsoft’s Long, Slow Decline“:

“People who love computers overwhelmingly prefer to use a Mac today. Microsoft’s core problem is that they have lost the hearts of computer enthusiasts. Regular people don’t think about their choice of computer platform in detail and with passion like nerds do because, duh, they are not nerds. But nerds are leading indicators.”

Microsoft’s losses to Apple aren’t based on “regular people” choosing the Mac.  Rather, these “regular people” were encouraged to do so by the geeks in their lives who had made the switch to a Mac years ago.  Consumer technology vendors can ignore the alpha geek niche at their peril.

Truer words of caution couldn’t be said to Apple, Microsoft and Linux desktop vendors in the face of Google Chrome OS.

Follow me on twitter at: SavioRodrigues

PS: I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies or opinions.”

Three weeks ago I wrote that Amazon RDS was going to eat into MySQL’s revenue potential.  I also pointed out that Amazon’s RDS was but a precursor to future Amazon cloud service offerings for other popular open source products.  While that post was centered on Amazon, it wasn’t a stretch to predict that any of the big IT vendors (IBM, Microsoft, HP, Google, Cisco, EMC/VMware, and Sun/Oracle) would offer RDS-like cloud services in the future.

Well, reading details of the Windows Azure platform this week, the prediction badge no longer applies to Microsoft.  According to Microsoft, Azure SQL will support MySQL, and Azure .NET Services will support Apache Tomcat.  Microsoft will also support PHP and Apache Web Server on Azure.  I’ll focus on MySQL, and to a lesser degree Apache Tomcat for this discussion.  I believe MySQL and Apache Tomcat will be the first two products offered as a service on large IT vendor cloud platforms, aside from the IT vendor’s strategic software stack that is.

When Amazon decided to offer MySQL via Amazon RDS, they did so without purchasing MySQL support from Sun.  I’ve confirmed that Microsoft Azure is supporting MySQL on Azure without paying Sun for a MySQL Enterprise subscription.  The logic as to why Amazon could do without a MySQL Enterprise subscription applies equally to Microsoft:

“Amazon’s decision to use the free version of MySQL to build RDS is completely sensible. First, Amazon has the technical skills to support their usage of MySQL without having to acquire the MySQL Enterprise subscription. Second, this decision helps Amazon lower the cost of RDS, which makes RDS more attractive to customers. This is clearly not good news for Sun/MySQL, which is missing out on capturing some portion of the revenue from MySQL users spending on RDS.”

As in my previous post, I don’t want this to be about Amazon or Microsoft versus Sun/MySQL.

The larger point is if Amazon, Microsoft, IBM, HP, Google, Cisco, EMC/VMware, Oracle/Sun offer a simple and supported cloud service for running MySQL, Tomcat, JBoss, Mule or Apache HTTP instances, what reason do customers have to acquire “enterprise subscriptions” from the vendors developing these open source projects?  Until now, the value of an open source “enterprise subscription” has largely been access to support and access to administration and management tooling.  In the case of MySQL, the former is provided by Amazon RDS and Azure SQL as part of the per-hour service.  Again in the case of MySQL, the latter is rendered unnecessary or replicated through Amazon RDS and Azure SQL tools.

If Microsoft can find business justification to offer both a .NET runtime and Apache Tomcat runtime as a service, then I wouldn’t bet on other IT vendors being solely faithful to their strategic software stack in their cloud service offerings.  I expect MySQL and Apache Tomcat will be supported by a plurality of large IT vendor cloud offerings.

Let’s turn our attention to the popular Apache Tomcat runtime. Covalent, now part of VMware via SpringSource, and OpenLogic are leading providers of Tomcat support.  Realizing that the scalable revenue potential is in selling products, SpringSource launched tc Server, or “Apache Tomcat with enterprise management”.  While I applaud the “selling products” angle, I reiterate that enterprise management not enough of a differentiator when up against cloud services from large IT vendors.

I still think that proprietary features in the “enterprise” version of the open source product, which are note available in the “free community” version, will be the path forward for open source vendors. Large IT vendor cloud offerings for open source products will rely on availability of the “free community” version and whatever features are supported in this version.  If there are compelling features outside of administration and management tooling in the “enterprise” version, then customers have a reason to skip the IT vendor cloud service and pay for the “enterprise” open source product.  I know the proprietary features suggestion is difficult for open source vendors to accept, and more importantly, difficult to rationalize in front of customers.  Difficult, but necessary.

In the end, there is no way that all expenditures on open source products such as MySQL or Tomcat will occur through a large IT vendor cloud service.  Clearly customers will want to use open source in their private data centers and acquire support and administration/management tooling from the open source vendor. There is also clearly an issue if IT vendors capture too much of the customer expenditure on open source products via their cloud services, leaving open source vendors behind these products to scrimp on future product development.  A happy medium will be found.  Well, it may not be happy enough for open source vendors as they watch large IT vendors capture revenue around open source products. But that’s competition for you.

Fun times ahead.

Follow me on twitter at: SavioRodrigues

PS: I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies or opinions.”

Reports suggest that Google will partner with phone maker HTC to deliver the much hyped and awaited Google phone; the gPhone as some like to call it.  The Google branded phone is rumored to be out in early 2010 with advertising starting in January.

As PC World’s Ian Paul suggests, the driver behind a Google branded phone is to “own the customer experience” through the gPhone.  While this makes complete sense for Google, what does it mean for Android, and more importantly, handset manufacturers building  Android-based devices?

First, let’s consider whether Google would use Android improvements required for the gPhone as a competitive differentiator versus other Android handset makers.  The Android license doesn’t require derivative works to be contributed back. However, I’ll give “don’t be evil” Google the benefit of the doubt and assume that they will contribute Android improvements back to the Android community. In this respect, the gPhone helps the Android community.

Next, let’s look at brand.  Until now, Android-based handset manufactures have been able to trade on the Google brand.  Competing against the Google brand will be altogether different.  Matt Asay nails it when he concludes that Android is winning over Symbian because of Google’s brand.  Taken one step further, would a consumer purchase a Google gPhone over a similarly featured and priced Android-based phone from Motorola, LG or others?  I would.  Would a consumer pay more for a Google branded phone over, say the Motorola DROID? Yes, that’s why Google’s brand is the seventh highest valued brand according to BusinessWeek/Interbrand.

Finally, can Google actually design and manufacture a superior device versus Motorola, LG and other handset vendor?  That remains to be seen.  Google can hire experienced designers and work with manufacturers such as HTC.  Google expertise in creating compelling, yet easy to use, user experiences could be the competitive differentiator here.  If Google can come up with an innovative way to interact with the gPhone, akin to the iPhone’s “pinch to zoom out” or “flick to scroll”, then the gPhone starts to be real interesting.  These innovative interactions will require API support inside of Android and new hardware designs.  The former may be available to all Android handset vendors, while the latter would not, thereby providing Google another point of competitive differentiation.

In some respects the gPhone, which Google has insisted it had no interest in building, is a kick in the shins, or higher up, to Android handset makers.  Google, friend and foe at once seems like a common trend these days.

Follow me on twitter at: SavioRodrigues

PS: I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies or opinions.”

I don’t want to get into whether or not Microsoft violated the GPL with their Windows 7 USB/DVD Download Tool. That’s already been covered in depth.  I do however want to ask how a situation like this could have even happened, and how it could be prevented.  Not just at Microsoft, but at a typical enterprise that sells products or services which include software.  Enterprises that fall into this category are increasing in number daily and will accelerate as enterprises start building mobile device applications.  Air Canada and Domino’s are early examples of the enterprises that I’m referring to.

We’ve previously discussed the notion of using open source to compliment a development budget. This is true for enterprises and software vendors alike.  As Black Duck Software’s Eran Strod writes:

“There is an abundance of great open source code available that includes components like libraries, stacks, databases, frameworks, etc; it simply doesn’t make economic sense to allocate development resources to build what Savio calls ‘undifferentiated capability’. Why spend money, and time, reinventing the wheel?”

As enterprises start to use open source within products that will be released externally, the need for an open source usage policy becomes critical.  This effort begins with developer education.  It’s helpful to have a set of guidelines with approved licenses and open source projects that developers can potentially build from.  But that’s not enough.  Enterprises need to verify the pedigree of code checked into each build, to the degree possible.  Companies like Black Duck Software and Protecode offer services to help enterprises using open source in their development process.

What surprises me about the Microsoft situation is that I’d be shocked if Microsoft doesn’t have an open source usage policy. Microsoft uses open source in some areas of Windows.  So, it’s possible that a developer just made an honest mistake.  However, that really isn’t a viable excuse, not for a software vendor and not for an enterprise in the future.  The need for education and vigilance is an endless task.

My wife recently replaced her Thinkpad T40 with a MacBook Pro (MBP). To save me time in the future, when a crash or data loss occurs, I decided to get Time Machine and SuperDuper! set up on a Western Digital Passport USB drive (I love those passport drives!).  The problem with this setup was that it relied on her plugging in the USB drive into the MBP on a regular schedule.  The solution was to set up some network attached storage (NAS) on our home network.

I quickly settled on two options.  Apple’s 1TB Time Capsule or a combination of a Western Digital MyBook World Edition (MBWE) and Linksys Gigabit N. In terms of backup, storage and network connectivity, the MBWE plus Linksys router provided equivalent function to Time Capsule.

What to do?

Time Capsule was $30 cheaper than the MBWE plus Linksys router option.  However, since MBWE is running a Linux kernel, the ability to add functionality to the device was almost limitless.  There’s a strong community of MBWE users that have everything from BitTorrent clients, to PHP, to a PBX running on MBWE devices. Even out of the box, the MBWE provided more functionality than Time Capsule, such as a Media Server. These “extra” features that the MBWE provides out of the box versus Time Capsule were not features I’d originally considered part of the purchase decision. But when I learned about them, I tried to figure out if I could add these to capabilities to Time Capsule. While Time Capsule runs a stripped down FreeBSD, it seems that the code is loaded onto a chip and is not modifiable.

While both Time Capsule and MBWE are built off an open source kernel, the user freedom afforded by the two products are vastly different. Time Capsule is not meant to be tinkered with. MBWE is arguably not meant to be tinkered with either, but Western Digital knows that their customers are doing so, and is making it easier to get more out of the device than the functions the device ships with. For instance, the old version of MBWE did not support SSH access to the device out of the box. But since that was the first thing that MBWE users built a hack for, Western Digital smartly responded by adding this function in the new version of the device.

I struggled with the choice. Using Time Capsule meant that Time Capsule and SuperDuper! just worked with the MBP.  On the other hand, there were a few system tweaks I’d have to make to get Time Capsule and SuperDuper! working with the MBWE and the MBP.

In the end, I went with the MBWE solution over Time Capsule. I liked the extra function and ability to add yet more by choosing MBWE over Time Capsule. To me, the freedom, in open source terms, that MBWE delivered was well worth the extra cost and effort to set up.

What do you think? Good choice?

Next Page »