Blog Post

Open the Clouds With Portable Stacks

Stay on Top of Enterprise Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

Currently, moving from one cloud to another is easy, and having multiple clouds to choose from gives customers the ability to utilize a range of features and service models to meet their varying needs. But proprietary next-generation databases, by locking customer code to specific clouds, remove the benefits of market choice, such as customized service experiences, competitive pricing and — most importantly — increased adoption.

To ensure continued advancement of the cloud, the industry needs to turns its support to an open cloud by using database technologies such as Cassandra and Drizzle Drizzle, which are portable to any public or private cloud.

Finding the New LAMP Stack for the Cloud

Many suggest that standards are the key to encouraging broader adoption of cloud computing. I disagree; I think the key is openness. What’s the difference? In the standards approach, a cloud would look and work just like any other. Open clouds, on the other hand, could come in many different flavors, but they would share one essential feature: all of the services they’d offer could run outside of them.

Such is the case with Drizzle, the fork of MySQL built for the big data needs of the cloud era, as well as the open-source Cassandra project, a next-generation database of the NoSQL variety and the engine powering the massive data needs of Twitter and Digg. These database technologies are the future of the webscale business — the next generation of the LAMP stack that helped drive down the cost of creating new startups in the first phase of the web.

The Million-user Problem

Five years ago, only Amazon, Google, Yahoo and a few others had to worry about millions of users and the data they create, which they dealt with by building a set of custom next-generation data technologies. Today, hundreds of companies are facing the problems associated with scaling their databases to a million or more users — and in another five years, thousands of companies will.

The first wave of web technologies that currently power most web applications (LAMP and .NET/Microsoft SQL stacks) are not adept at solving the million-user problem. Some companies have solved it with their own platforms — such as Google’s AppEngine, Amazon’s SimpleDB and Microsoft’s SQL Azure — but each is locked to its respective cloud.

The web needs an open set of tools to solve the million-user problem. Imagine a set of tools that could be run on any cloud. Drizzle and Cassandra are the leaders in this race to create the technical foundation of the million-user stack, and Twitter, Digg, Reddit and many others along with Rackspace are contributing to these tools in order to keep them advancing.

Open Not Free

The time is now to commit to an open cloud. Google and Amazon should make their data storage technologies open. That does not mean they must be free. Microsoft has hinted that its Azure technologies will eventually find their way into the core SQL products. This is good, as it will give the web another set of tools unconstrained by a particular cloud provider. As cloud providers and web companies continue to search for the next generation of technologies for building out webscale businesses, they, too, should support the efforts of Drizzle and Cassandra to help scale the web in order to meet its full potential.

Lew Moorman is the president of Rackspace’s cloud division and the company’s chief strategy officer

For an in-depth look at all things related to cloud computing, join us at our annual Structure conference on June 24th and 25th in San Francisco.

14 Responses to “Open the Clouds With Portable Stacks”

  1. Lew’s right, and Rackspace definitely practices what it preaches. But the norm is to intentionally open-wash one’s product or service, and worse, to never get called on it.

    Apple and Facebook are so large that there’s been a real debate; customers know what they’re buying into. Totally different story among anyone smaller: the topic rarely comes up and there’s no accountability.

    We’ve seen this firsthand. One of our competitors encourages customers to create apps that are tied to them (and are likely to be forever). The company uses the word “open” because the API is documented, journalists don’t call them on it, and even customers rarely connect the dots. Imagine encouraging customers to create a “Rackspace app” that could run nowhere else, and then calling it “open.”

    After watching that progression, I can say that it’s easy to open-wash using loud marketing and well-groomed relationships. In the commercial world, I think legitimate open-ness – multiple vendors or source access to a viable implementation, a standard, actual interoperability/portability, people who care about such things – will be rare without real consequences.

    Here’s the single easiest way to start. Ask suppliers and interviewees, “As a customer, how would I change to another vendor?”

  2. Clear disclosure: I work for Rackspace.

    Amazon, Google, Microsoft, they all have rights to create their own proprietary solutions. However, imagine what would have happened if there were three major web hosts back when the web first started, and they all had their own proprietary databases that you could only use with their respected services. Think about it–how much do we owe to MySQL and the fact that it is open source and is ubiquitous? Where would the web be without it?

    All we are saying is that we would rather put our dollars and our sweat behind technologies that are free and available to anyone, not just people using our service. We feel it is a better solution for us, our customers, and for the web as a whole.

    There are many non-relational database solutions out there that solve different problems. Instead of consumers picking what cloud they want to use, they should be able to pick the database that works for their needs and then use it with any hosting provider out there. Locking in consumers with proprietary technologies seems short-sighted–like everyone deciding they are going to build their own roads when we can all pitch in and get where we need to go a lot faster.

  3. For cloud mobility it is not just the database thats the issue, it is the proprietary API’s on the various clouds which are required for performing essential tasks such as load balancing, logging, messaging. Currently there are no open standards for these essential services.

  4. Samantha Atkins

    How open is MySQL? Last I heard there was a dual license with fees for business usage. Presumably this would include any cloud based business. Yet I can run MySQL server in a cloud and even used the distributed version Amazon provides with no apparent license issues. So I am unclear on the status.

    What of Postgresql? This is a very strong database with its own replication and scaling projects. It is FreeBSD licensed all the way. Yet it is not used as much as MySQL. Technically I don’t see why not.

  5. john – No question that traditional tools can solve these issues, but there is a reason that google, amazon, digg, twitter, reddit, yahoo and many others are using new technologies. The scale of the web is creating unique challenges and there will be new tools.

    Kristina – I do work for rackspace and am clear about that and our intentions. We did not get a unique advantage from LAMP and we will not get one from the new set of open, standard technologies – whichever win. We are helping advance a set of tools that we think are well positioned, but ultimately, the market will determine the winner.

    Adrian – We are not a technology stack company. That is the point. Should clouds and development stacks be tied or not? We think not. But, we also understand why others would compete on the ability to get access to new tools. We will see what is the better model.

    • Lew, let me be clear, I’d love to see cloud inter-operability. Right now, however, the software layer is considered by the major players to be their competitive edge. I’m sure many of Amazon’s largest customers have probably privately requested to have the AWS-stack run their private cloud, at the very least for compliance reasons. If anything, Amazon’s Virtual Private Cloud is some kind of weak compromise towards that end.

      I think “cloud” is still quite a few years away from commoditization. This is a good thing as there’s still room for innovation, and also for smaller players to offer inter-operability solutions.

      Compute is a long way away from being a utility. And hey, Ellison already humourously criticized the cloud as “everything we already do”. Perhaps Rackspace should consider people being in the market for “Oracle-as-a-service”. Just food for thought.

  6. Let’s see, Rackspace, a company with solid infrastructure chops but no core software competency thinks that everyone should become a commodity provider of… infrastructure? Color me surprised, but don’t expect everyone to rush to take you up on what would become a swift race to the bottom.

    • JLW, that is not what I am suggesting at all. Honestly there are a lot more ways to differentiate in the cloud than people realize. My argument is just that technologies tied to clouds will not thrive – I find it hard to believe that many will bet their company on a technology that could only be run on one cloud. We need ubiquitous core technologies. How they are offered – features, experience, service will all still be ways for clouds to differentiate.

  7. Understandable position with regards to Google and Amazon of someone representing an organization without the technical know-how to be competitive.

    I hope you gave GigaOM a nice Christmas present to deserve this free soapbox.

  8. Cassandra is a somewhat useful database for a small set of problems, but it is hardly a leader in the space. The author of this article works for Rackspace, which is a major supporter of Cassandra and Drizzle. The bias is understandable, but it’s a bit like asking a political party which candidate won the debate.

  9. “The first wave of web technologies that currently power most web applications (LAMP and .NET/Microsoft SQL stacks) are not adept at solving the million-user problem.”

    Pretty sure that’s inaccurate.