I love Amazon’s Web services.  They are game-changing.  EC2 lets you scale the number of servers you’re running infinitely at a moment’s notice, and S3 gives you an infinite amount of storage space.  No systems administrators to pay, no complex infrastructure to maintain, upfront cost to swallow.  Just gobs of cheap scalability. 

I’ve looked into using them before, and the ONE thing that seemed like a glaring hole in their arsenal was the lack of a way to run a database on it.  The problem is that EC2 virtual machines lose all their memory when you reboot them, so you can’t run a database on them.  And there’s no way to use an EC2 instance to store the data, it’s not a filesystem that any database engine recognizes.  So we had this awesome scalable firepower and awesome scalable storage, but no way to really hook them together.

Amazon's new SimpleDB service addresses that.  Now you can run your entire application in the cloud.  This is huge.  It’s not a relational database system, so you can’t do number crunching or analysis on it, but it’s PERFECT for CRUD operations (Create Read Update Delete).  Any transactional application is capable of running on this platform.

Let’s see what this does to an IT budget:

• Eliminates CapEx costs (no servers or datacenters to buy)
• Eliminates systems administrators (no machines to administer)
• Eliminates database administrators (it automatically takes care of storage and indexing for you)
• Eliminates capacity problems (no need to estimate and budget for potential traffic, you just pay for what you use—if your bill is too high you just kick users off)

And, as a result of this, SimpleDB is a startup’s best friend.  This makes bootstrapping a startup a million times cheaper and easier than it was previously—you only have to pay as your user base grows.  Which SHOULD cover your costs, or you probably have a problem with your business model :)

If you’re a system administrator or a DBA, it might be a wise career choice to make yourself familiar with these technologies.  The sheer financial sensibility makes it inevitable that small-to-medium businesses will be interested in running on this platform.  In fact I wouldn’t be surprised if completely new software applications are built on it and sold as traditional premise-based software, but run in the cloud.

Very cool stuff, I can’t wait to play with it.  I read thru the documentation and it looks pretty simple, but I'm currently waiting on Amazon to enable my account.

UPDATE:  It's not all gravy--as with anything, there are challenges and considerations to working with this kind of architecture.  I found a good list outlining the pros and cons here.

BTW Want to know something else that’s game-changing?  Ron Paul raised $6million yesterday—in 24 hours—beating John Kerry’s old record of 5.something (on the day he accepted the nomination)—totally grassroots—with no help from the campaign.  This is a truly incredible, science-fiction type event in politics.

Have you ever heard anyone debate the merits of software sexiness and user-friendliness?  I'd always thought they were virtues on par with showering daily and being nice to your mother, but apparently I was mistaken.  I got a good chuckle out of a debate going on recently about whether enterprise software should be sexy and easy to use or not.  You don't get to poke fun at people arguing for a flat earth or horse-drawn carriages too often, so I just can't resist the urge to join in the fun.

It all started with Bill Gates complaining that bloggers don't write about enterprise software often, instead focusing on consumer software.  Robert Scoble correctly pointed out that enterprise software is generally unsexy and therefore people just aren't that interested in it.  I would add that it's generally not very innovative either.  Innovation tends to flow from consumer software to business software, not vice versa.

So Scoble asks what we can do to make enterprise software sexier.  In response to the question, which any enterprise software company which values self-preservation should be asking itself anyways, ZDNet's Michael Krigsman responds with a sharp "Scoble’s question is irrelevant and meaningless."  We should all be thankful Krigsman is not a school teacher.  Anyway, he goes on to state:

"Robert Scoble misses this point: unlike consumer software, where sex appeal is critical to attracting a commercially-viable audience, enterprise software has a different set of goals."

I'll save you the trouble of reading the post and tell you that his point is that enterprise software is to be stable, above all else, sacrificing the user if need be.  While I can see his point, this doesn't negate the fact that humans actually have to use it, and so if customer satisfaction is a goal (I know it's a HUGE one at Cisco), your software had better be pleasant to use.  And if you actually want to sell MORE software, your software had better have some sex appeal.  This is 2007, Mike--Apple and Web 2.0 are kind of big deals these days.  People like their iPods.  We were laser-focused on the user experience at Latigent, and it's a big reason why we were so successful.

So then Nick Carr, always a good read, chimed in and correctly pointed out that design and functionality are NOT mutually exclusive goals.  Seems like a no-brainer to me.  But Krigman, champion of green-screen and mainframe applications everywhere, is back on the scene to accuse Nick Carr of living in Fantasyland (I'll bet Mike's a hit at parties), saying that priorities, legacy support requirements, product cycle times, and technology limitations prevent enterprise software from being friendly.

And then it gets really funny, as people involved in developing (presumably ugly) enterprise software jump to Mike's defense:

• "it's just a matter of how you look at it, as Vinnie and Anshu say "beauty and sexiness is in the eye of the beholder" and you know what I too am damn proud to be lumped into this one." - Craig Cmehill
• "Yes, consumers get turned on looking at a map overlayed with restaurant locations but my enterprise customers want to see real value." - Anshu Sharma
• "You see, many average joes like Scoble are drawing a long-term assumption, and that is that enterprise software is going to converge with where consumer software is right now and where it is going." - Jevon MacDonald  (Apologies to Jevon, he pointed out that he wasn't actually agreeing with Mike, and after re-reading his post I have to agree.)

I think what we have here is a case of people thinking inside the box.  Wrapped in foam, surrounded by packing peanuts, with no chance of any exposure to the world outside their bubble.  I get the sense that these guys have built enterprise software before, and they didn't get the best feedback on it from end-users, and are a tad defensive.  You see, the key here is:

End-users are also consumers

It all comes down to priorities.  If you make user experience a PRIORITY, and spend some time actually WORKING ON IT, and THINKING ABOUT IT, it will probably turn out OK.  If you let programmers design the interface, it will just make sense to the programmer.  And most programmers are not the most aesthetically-oriented guys.  Being a programmer myself, I learned that a long time ago.  If I just run off and build something that makes sense to me, there is absolutely NO guarantee that it will make sense to anyone else.  In fact it probably won't.  I've learned to live with it.

But fixing that is anything that requires focus groups and design gurus to implement, either.  All you really need is good communication between the people building the system and <gasp> the sales department.  You know, the people who are going to show the product off to the customers who buy it.  They generally have their fingers on the pulse of the people they're selling to, and they're usually not too technically inclined (which is a GOOD thing when you're talking about designing interfaces).  If I scribble a mockup of a screen on a whiteboard and their eyes glaze over, I know that I'd better rethink that screen or be prepared to A) document it well B) train people on it and C) hear complaints about it for years.  It's probably easier just to erase it and try something else.  Better yet, get some key customers involved.  That way you'll know that you have at least ONE customer happy with the end result.

It's amazing to see how much effort some people will go to in order to defend something that obviously can't be defended.  A little internal communication would go a long way in preventing these types of situations.

I’ve been thinking a lot recently about the whole centralization/decentralization thing, and privatization of the Web.  Seems you can’t really get away from the topic right now, every other day some new privacy issue related to FaceBook or MySpace is lighting the blogosphere on fire, splashing cold water on our collective faces and reminding us what a fragile ecosystem the Web really is.  The latest hot button issue is FaceBook’s Beacon advertising program, which basically broadcasts your Internet activities to your friends on FaceBook (yeah, just the kind of innovation we’ve been waiting for…)

I'm not going to go into the stupidity of this Beacon idea--really, is it even worth debating?--but to me this is just the latest in a long line of examples of the dangers of too much centralization.  Give someone—anyone—unfettered access to your data, and they WILL find a way to milk it for as much money as possible.  If you’re betting on the goodwill and good intentions of a corporation formed to make money then you obviously haven’t been involved in a capitalistic society for long enough. 

One of the big fears I have about the future of the Internet, if we don’t decentralize and privatize it soon enough—is the growing power of the few centralized sites that everyone uses.  If you think about it, there are a few select sites that have a LOT of power over the Internet and its citizens, simply because of the numbers of eyeballs they control.  Probably at least 90% of Internet traffic flows through them.  Without them, the Internet would be half as useful as it is today (which is a big part of the problem).  I can think of a few sites in particular:

• LinkedIn, Facebook, and Myspace—to find people
• Yahoo, CNN, MSN—to find news and content
• Google—to find just about everything else
• Amazon, EBay—to buy things

The only real exception to this rule is when somebody physically sees a URL somewhere—on a billboard, on TV, or in a magazine, and types it directly into their browser.  But even then, there are a large percentage of people who just type the URL into Google by mistake since they’re so used to using it.  Using Google as the entry point to the Web is almost reflexive.  I’m sure the browser toolbars don’t help with the confusion either.

A big problem is that people tend to go thru these clearing-house sites and central filters in order to get to the content they’re looking for instead of directly to the source of the content itself.  Instead of looking at my business card and seeing my site address is jasonkolb.com, they’re much more likely to look me up on Google and find my site that way (judging from my site stats, a LOT of people do this).  Or, depending on their age, they may look me up on Facebook or Myspace and find my profile there (which is just a placeholder pointing them to this site).

So far this has worked ok, but what if those filters and middle-men become unreliable?  What if Google suddenly decides that there’s another Jason Kolb who deserves the #1 spot in the search results for “Jason Kolb” more than I do?  (Not exactly “Joe Smith”, I know, but there ARE other Jason Kolb’s out there.)  And then, what if that other Jason Kolb decides to pretend to be me?  Hasn’t my identity effectively been stolen?

And since I rank #1 for the term "Jason Kolb", what if I pretend to be a different Jason Kolb?  Haven't I stolen his identity?

And then what happens if those middle-men and filters actually go down completely?  As TinyURL so effectively demonstrated recently when it went down, if middlemen sites go down, the Internet almost ceases to function.  If Google went down, not only would nobody would be able to find me, the Internet would be pretty much paralyzed.  And if Facebook decides to block searches for my name, as it recently did with Ron Paul  (vote for Ron Paul 2008, Save Our Country!  Sorry, couldn’t resist ;), it’ll make it that much harder to find me.  If FaceBook goes down, suddenly HUGE numbers of people effectively go off the grid and become unreachable on the Internet to many of their friends.   

If everyone were running their own private site, similar to this blog, it wouldn’t be as much of a problem.  Of course, that still doesn't solve the problem of people looking for me via Google, but that's a much more complicated problem to solve.