Building Apps in a Cloud World

I run a monthly meeting called the SoCal VCA (Venture Capital Alliance) in which we have great technology clouds_320leaders present to the best VC’s in Southern California so that we can better understand the trends and develop tighter relationships with our local technology leaders.  In recent sessions we have had Dmitry Shapiro (founder of Veoh), Ian Rogers (CEO of TopSpin Media) and David Sacks (founder of Geni and Yammer, former COO of PayPal).  I hadn’t yet restarted my blog so notes from those meetings are still scratched in my yellow notepads.

This month we had an equally impressive speaker in Michael Crandell, founder and CEO of RightScale, a Santa Barbara based company that helps manage cloud computing infrastructure.  The company has raised more than $20 million from prominent investors including Kevin Harvey of Benchmark Capital and Danny Rimer of Index Ventures.

Michael’s entire presentation is here on Docstoc but I wanted to highlight a couple of slides below from his session and offer some thoughts about cloud computing.

1. Cloud continues to accelerate

When I first started experimenting with cloud computing it was 2005/06 and I had just launched my second company, Koral.  Amazon had just launched its S3 storage (EC2 for cloud processing didn’t exist yet).  It seemed that over night every start-up I knew was using S3 and the logic was simple.  In my first company we had to buy very expensive EMC storage ($500k), high-end UNIX servers from Sun ($40k / pop), load balancers, Sun Solaris operating system and an Oracle DB ($80k).  We were $750k into the hole before we ever had a single customer.  Such was life building a SaaS company in 1999.

Fast forward to 2005 and we were building with $4k PC’s, LINUX operating systems, cheap storage devices and a PostGres database.  But it was still $60k when you consider 5 servers, load balancers, storage, hosting, bandwidth commitments, etc.  Enter Amazon Web Services (AWS) who would charge us for storage only as we used it (e.g. scalable with our business).  We initially only used it for back-up storage and test processing.

In 2008 I invested in my good friend Stuart Lander’s new business called PublicSpend, whose long-term objective is to bring transparency to government spending on local vendors.  The architecture was designed by rock-star SaaS technologist Ryan Lissack , who is now with  He convinced me that we could build a company without buying any servers, load balancers or storage.  And simultaneously I noticed that a number of start-ups that were pitching me had done just that.  So it seems that the era of CAPEX is ended for a number of companies and Nicolas Carr got things mostly right in his book, “The Big Switch“.  And while the economy has slowed down and start-ups have curtailed spending I expected to see a slow down in Cloud spending by start-ups.  But if the graph below is any indication it seems things haven’t slowed down at all.  The red line is a good proxy for start-up spending with AWS. (note: there is more text after graph – some problems with don’t allow me to properly resize it)

Cloud Growth 3

2. Clouds are not inherently managed

What we were doing in building on AWS was buying “infrastructure-as-a-service”, a term I first heard from Michael Crandell.  We were buying storage and processing power but we still had a lot of technical work to monitor the services and be ready to provision as demand increased.  This is the problem that RightScale stepped in to fill as depicted by the graph below.

The blue line represents purchasing in a CAPEX world where you buy equipment before demand.  The dotted line represents predicted demand and the red line represents actual demand.  You’re either over-provisions or under-provisioned, leading to increased expenses or unhappy customers.  But what a truly “managed cloud” service offered was a feature called “auto scaling” that automatically monitors usage patterns and can provision new server instances if you get flash crowds as happened with this Free Chocolate website by Mars.  The beauty with auto-scaling is that  not only will new instances by provisioned for you for when TechCrunch writes that great article about you but you can also auto-scale down when the flash crowd leaves.

managed cloud rationale2

3. Everyone but Amazon seems to be asleep at the wheel

What has surprised me the most about Cloud Computing is just how asleep at the wheel everybody but Amazon seems to be.  Where are the great cloud services from Microsoft, Google, Sun, IBM or RackSpace?  I know that all have made announcements as have a host of other providers like GoGrid.  But literally every company that has pitched me who uses a cloud uses Amazon’s AWS.  Clearly that’s not great for the market and for innovation but it’s great if you hold AMZN stock.  And it’s probably not great for RightScale.  Ultimately Amazon will move up the stack with more management offerings and the most compelling feature of RightScale when that happens will be the ability to “abstract” from any individual cloud implementation so that you can more easily transfer across clouds.  But for now RightScale’s product offers significant enough value in offering a true “managed cloud” rather than just “infrastructure-as-a-service” that you get from Amazon.  Michael made it clear that he thought the competitors were making progress but the momentum feels as intractable as trying to supplant iPhone for mobile browsing.

4. Private Clouds on the horizon

One of the more interesting developments I noticed in the past 6 months was the evolution of the idea of private clouds.  When I first learned of the company Eucalyptus Systems I have admit I didn’t immediately get it.  Why would anybody want to manage an internal cloud?  But the more I learned about it I realized that there is a compelling case for large companies who have lots of excess server capacity or storage and want to be able to offer it out to other geographies or departments rather than have excess capacity.  And if you could do so in a way that allowed you to use some internal clouds and then also call external clouds with total abstraction and therefore one set of instructions to manage it would help to migrate to external clouds over time.  And now it seems that RightScale can also sit on top of these “hybrid” clouds and help manage their co-existence.

My conclusion is that Cloud Computing is not only here to stay but it is accelerating.  Today it seems to mostly be the domain of start-up companies or test projects for larger corporates.  But as we experienced with the success of it was the small, nimble players that adopted it first.  And as we all learned when reading Clayton Christenson’s monumental book, The Innovator’s Dilemma, eventually the higher-end of the market will trade down given dramatically reduced costs and as Cloud Services continue to improve.

Built apps in the cloud?  Views on AWS or RightScale?  Views on AWS competitors?  Love to hear from you in the comments section.


9 Responses to Building Apps in a Cloud World

  1. David says:

    Great post. The next logical step would be to bring cloud-style billing to the client-side. People seem naturally inclined to pay a reasonable sum for what they use in the cloud, so why not in the browser too?

    BTW, you have a small typo: “Nicolas Carr got things mostly write…”

    • marksuster says:

      Thanks for catching the typo. re: Cloud style billing, not 100% which type of billing you refer to but there are a host of really good new Cloud Billing applications including Zuora and Vendicia.

      • David says:

        Thanks Mark, I knew about Zuora, but not Vindicia.

        This is perhaps not the place to expound these views, but (briefly) I think most subscription models create too much user friction and apprehension.

        I blogged here on the problems with subscriptions (within the context of freemium) and why pure cloud-style usage billing makes more sense.

  2. As an active participant in the cloud computing industry, we are seeing very similar trends. Through most of 2008 we spent a large amount of our sales effort directed at education. In 2009, we have seen a significant shift as the awareness of cloud computing has grown and individuals are more interested in how they can leverage the cloud. More discussions are focused around when they will shift and what they will shift to the cloud, rather than if they will shift.

    • marksuster says:

      Thank you, Alicia. It reminds me of SaaS circa 2001 where we had to evangelize a lot in our sales process. The big corporate buyers we cautious and nimbler start-ups used the technology. I’m convinced corporate computing power will shift to the Cloud – the bigger question is when!

  3. Yes, well the good thing is that conversations are getting easier, there’s still a lot of education going on in the sales process, but from what we’re seeing I think the shift is on its way and soon!

  4. Great post Mark. I would also agree that there is still a lot of education required but really, I would encourage your readers to try things out in the cloud. It costs users nothing to next-to-nothing to test a few servers and get their feet wet. The thing that shocks me the most is that there is all this buzz about the cloud but 80-90% of the “buzzers” have never even tried it. What are they waiting for?

    RightScale has a 100% free edition of the platform and all you need to do with an IaaS provider is put in your credit card. Just remember to “turn the lights off” when you’re done.

    • marksuster says:

      Thanks, Matt. I agree – people should give Cloud a shot. But history tells me corporates will move more slowly.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: