Anyone who’s been in this business for more than two or three nanoseconds knows that we phase-shift from time to time – and some of those can be pretty darn disruptive if you’re living through them.
For example, in the late 70s everybody was arguing about just which network architecture was going to win out in the enterprise – SNA (that’s a new bonus term from the “where are they now” pile … answers tomorrow) or DNA (still wondering what that is?) – and then it just happened. Some obscure protocols (tcp/ip) springing out of a federally-funded research project took over the world. Kind of like Kudzu, but a lot more useful. Once that mindset was in place, a few other helpful additions – ftp, smtp, http plus their buddies – and we had an actual internet.
Phase shift … and I guess it worked out pretty well.
To make that particular phase shift practical (and you can argue about what precisely enabled the phase shift, and what just went along for the ride – but that’s not my point here) there had to be some big breakthroughs in hardware, operating systems, and software development. Those breakthroughs came along, some fortunes were made, some lost, and some people were left to mumble over a drink about the good old days.
The bubble burst called bogus on the first round of net economics, but not on the technology. (as a side note the bubble bursting didn’t get rid of all the kudzu – some of the clumps were smart enough to sell before the economics sorted out, buy sports teams, and give advice way outside of anything that they know anything about – but that’s most definitely a post for another day!)
The technology kept improving, and now there are actual, shiny new business models that may be sustainable for awhile.
All of that innovation – business models and technology – is setting the table for the next phase shift. In fact, I think it is very clear that we are well into the first part of this phase shift.
Redshift – Truly Googlactic Scale
In my last post we talked about the world according to Sun, at least post the next phase shift. A world of five computers, mostly powered by Sun, where everything is a service, and all entities trust all of their applications and data to those five entities. A world with a kind of geeky name – redshift computing – that I kind of like (true geeks will admit the affliction in public!).
As several folks have pointed out to me, the real value in this vision is that what we’re really talking about here is scale – humongous scale, googlactic scale, really truly ginormous scale – and some things that will have to happen to enable that world to exist. However, the part that’s not nearly so obvious is who wins in this world in general, how to get to this world, and whether Sun will even have a meaningful role. They may, and they’re working as hard as possible to ensure that they do, but the fight has already moved on.
Yesterday David Berlind posted again on Redshift, this time a contrarian view from salesforce.com. Well not contrarian on the need for scale, or even particularly opposed to what I think is perhaps the most dubious part of the vision as it stands – that there’ll only be five computers – but contrarian to whether Sun, or for that matter any particular hardware provider will matter much. Platform evangelist Peter Coffee of salesforce.com makes the case that the real value will not be in the hardware, nor in any of the infrastructure software, but in the application platform. Of course, he thinks that salesforce.com and their application platform service (apex) will be one of the surviving dominant players – occupying two of the surviving five computers – but that’s probably just the result of a bit too much mountain dew in the morning!
Are You My Platform?
In any case, what I do agree with from this point of view is that what we call a platform is changing. In fact, it is at the very heart of the phase shift.
Used to be that the servers or mainframes that you bought was the platform. Then it became the operating system, then all of the infrastructure software – application servers, databases, web servers, server servers (just kidding on that one), and so forth – along with the the operating systems and the hardware they ran on.
Now we are seeing two new worlds arise. The first are fairly complete ecosystems (platforms) that are focused on particular applications. In many cases the applications (generally as services) came first, then generalized into platforms. Obvious examples and contestants include google (in many parts), yahoo (in some parts, such as flickr), amazon, myspace, even salesforce.com and facebook.
The second world are partial ecosystems (part of these new platforms) that have everything but the app itself (such as amazon ec2 and s3), as well as many variants on the “naked scale” theme. Some of these are services, some are within an enterprise, some are in some of the strangest places, but that’s also not the point.
In each of these worlds the notions of scale and reliability and all sorts of other things – including, to the chagrin of many folks, security – are assumed to be delivered properly by the platform service provider (or PaaS, the name proposed by salesforce.com). The reality of that is a discussion for another day, but let’s keep moving for today …
I think what is clear is that in this phase-shifted world the real value proposition is the application itself, which by definition includes its constituent data, the ability to get at that data, and the ability to slice and dice it to your heart’s content.
The real battleground will be over who can most directly enable that world and actually deliver on the scale, reliability, security, & accessibility that customers demand, all while doing it cheaper than any competitor. Cheaper in all dimensions.
Who wins here? Sun or salesforce? Redshift or PaaS? Is there a role for the current hardware and operating system providers? What about the enterprise?
More soon.
——————
Postscript. When I first posted on redshift Andreas Schwarz of Sun pointed out that salesforce.com was actually a Sun customer, and had been so from the beginning. I corrected my original post to move salesforce.com from the commodity camp to the Sun camp. Turns out that I didn’t need to do so – I was just a little ahead of the curve. As a result of this discussion Mark Benioff (ceo and founder of salesforce.com) pointed out to David Berlind (look towards the end of this post) that the key phrase here is “had been”. Turns out that salesforce.com is developing a taste for commodity, and now prefers “linux on Dell” – they’re “phasing out the Sun gear”.
Oops.
Technorati Tags: application fabric, application virtualization, data grid, datacenter, redshift computing, scale, scale data, scaling db, scaling Web 2.0, service virtualization
Comments on this entry are closed.