SAN FRANCISCO – Marc Andreessen is probably best known as the co-founder of Mosaic, the first widely-used Web browser, but he’s also the co-founder of Netscape, Loudcloud and now Ning, his latest venture that provides a platform for social networking sites.
Marc Benioff, CEO of Salesforce.com, characterizes Andreessen as the person who “practically invented the core technologies we are using mainstream today.”
Benioff is probably best known as a primary founder of the software-as-a-service movement, delivering applications over those core technologies he attributes to Andreessen. Now he is among the proponents of cloud computing, making those core technologies a platform as a service.
The two spoke in an exchange Jan. 17 at Salesforce.com’s Tour de Force, a day-long conference for developers, where Andreessen said he believes the platform-as-a-service wave is just beginning.
MB: We saw the emergence of the PC, Windows, Linux and the horizontal aspects of platforms – and database, analytics, applications companies. Are we going to see a whole new range of platform companies?
MA: Yes, I think there’s going to be a whole range, thousands of platform companies.
There’s a common definition of what a platform is: it means you can program it…This shift to an Internet, the cloud, or whatever you want to call it is a very, very big deal from many different perspectives. I think there is going to be literally thousand of new companies in this approach. We’re just starting to see cascade of that.
MB: In our world, we are seeing the first layer [of an on demand platform] is a global trusted platform – the data management layer is there, the integration layer, the Web services APIs, or meta data API. On top of that is logic, or workflow, and on top of that maybe some kind of user interface and packaging mechanism. Is that how you see [PAAS] getting laid out? Is that how it will evolve as stack?
MA: That’s the general stack of how we thought about it in the old days. The new analogy to the old way of thinking about the platform stack [is to] add a few things on top of that. Clearly significant is that [an on-demand] infrastructure relieves developers of worrying about infrastructure. Some of our networks are growing 10, 20 percent a week
So as traffic takes off, you don’t want to worry about the underlying stuff, the run time.
That’s a very, very big deal.
The second part is that the platform doesn’t have to be just about running code, it can be about accessing other services. Running financials, tying to Google, tying to Amazon. It can be live services. It’s not just code, its access to services and to data that are actually live connections.
The other thing is that by definition it’s a service that can be used simultaneously by thousands or millions of people. So it’s inherently social. People can collaborate on building an application, but on using an application too. All the sudden you have live data, live services, and live people and the whole thing is connected. That’s a much more fertile ground for innovation versus a server sitting on the floor. It’s just code. It’s inert.
Is PAAS a Complement to Existing Platforms or a Replacement?
MB: Do you see PAAS complementing existing platforms out there? Or as a replacement strategy? How do you see this evolving?
MA: If you’re Salesforce.com or Ning, you’re using a lot of piece parts from the old world to build the new – servers, Linux, an operating system. We even use Oracle as our foundational layer for the multi-tenant database we’ve built.
From a developer standpoint, we’re representing a new style platform to our users. They never deal with an operating system, SQL, a piece of hardware. There’s never any awareness of what router they’re using, how storage is configured, how the OS is configured. It’s purely programming at HTML and the scripting layer. For developers, that’s a very big shift.
Sun is going to be an arms supplier to new companies building platforms like Salesforce.com and Ning. In the fullness of time we’ll see how everyone else acts.
There’s a lot of potential growth. The roles are shifting of who customers are and ultimately what developers can do
MB: What does this do for industry at the end of the day? What’s the big value add?
MA: A lot of people in the world today – if you tell them you have to buy servers, fire walls, etc. [to build an application] and if they had to exist in rural Wisconsin where I grew up [it wouldn’t happen].
Today, someone is able to go on Ning, on Amazon or on Force.com [and build apps]. Even if you look at Silicon Valley there are more and more start ups – more start ups that I am investing in – that are launching with the concept that they will never have to deal with an infrastructure.
I’m seeing two types of start ups: start ups building platforms, which is interesting; and start ups building apps, where all they have is a bunch of laptops to get on the Web to start building applications. They can run a transaction through PayPal and advertising through Google Ads. They can run their internal system on Google Apps and they never buy anything but a bunch of laptops.
There’s going to be a lot more people in the world going completely virtual.
MB: Do you see that happening now? Or are we going to see these emerge in new countries, new regions? Or will Silicon Valley dominate?
MA: Silicon Valley has critical mass and all the talent used to doing all these things, but I am seeing more and more start ups that are completely virtual. As an investor, the capital startup – if all the sudden able to be completely virtual, running on other platforms – then all the sudden the need for up-front capital is very light. So the financial risk of investing is much lighter. The [implications] are huge.