My Own Pirate Radio

Amazon’s S3 service was down this morning, from around 4:30 AM PST through 7:17 AM PST.

There are lots of upset people on the developer forum. Services like Twitter, SmugMug and Tumblr that host resources on S3 suffered some pain. ZDNet posted a predictable "sky is falling" piece.  AlleyInsider has a more balanced view. There’s more on Technorati if you want it.

Unfortunately, very little official comment on the forum so far from Amazon. That’s a mistake I’m sure they won’t repeat. I sympathize. It’s tough maintaining the presence of mind to post public status reports when everything is going haywire around you. Nevertheless, you need to, when running a service like this. 

Some voices on the forum are saying this proves you need a fallback option for those days (hours) when S3 is down. Unless you’re running a service that’s essential (i.e. literally vital) to your customers, that’s crap.

The whole point of services like S3 is that they are reliable and cost-effective enough for you to build your business on top of them, using small increments of cash (buying 1GB at a time) instead of big increments (buying 1 server at a time) that result in overbuilding.

Will Amazon run its services at 100% uptime? No. But I bet their uptime will be proven better than almost any smaller company’s uptime.

Should you backup your data and store it somewhere offsite, for disaster recovery purposes, auditing, and so on? Yes.

Should you build another whole operations center — hardware, electricity, bandwidth, security, staff — and leave it idling just in case your main ops goes down? No!

There’s a great analogy here to other utility services such as roads, transit, electricity, water, and medical care. Essential services try to build out a small amount of reserve capacity for dealing with emergencies. Non-essential services cannot afford it, and so when they fail, they fail completely. And this is the way it should be. Can you imagine what all this infrastructure would cost, otherwise, and how much resource we would waste in building and maintaining it?

Perhaps there will come a day when we’re simply unable to survive for even a few desparate hours without Twitter serving up avatar images, or Blackberry zapping us with new email messages. But until then… them’s the breaks.

That’s Microsoft’s stock, plummeting $2.00 after the announcement of their bid to buy Yahoo!. 

And there’s YHOO, which until the bid news had been muddling about at its lowest point in the last few years, especially after recent announcements of layoffs.

Ouch. Gotta hate acquisition bids… the bigger company’s stock price almost always gets hammered as investors worry about integration cost.

Personally I think the proposed merger would be a good thing. If the combined company could manage to hold on to Yahoo!’s customer base and integrate their online offerings — and those are two very big if’s — it would stand a real chance of competing against Google in search. More importantly, it could be great for customers: advertising clients would get a better ad platform and hopefully better pricing due to economies of scale, and end users would get a web portal and email experience that… well… doesn’t suck. (Sorry, I’ve never been a fan of MSN or Hotmail’s UI.)

Google is making noise about monopoly power and trying to stir up the regulators. That’s predictable, especially with Eric Schmidt at Google’s helm. And surely the US government will investigate, and that will take months. But I don’t think that’s going to be the real barrier.

The toughest hurdle here is culture. Or, to put it more bluntly, ego. Microsoft has always suffered from a particularly bad case of Not Invented Here. Asking MS product teams to merge their search, advertising, and portal businesses with Yahoo!’s will be extremely difficult for some to swallow. Asking Yahoo! product teams to integrate their internet properties with Microsoft services will be a similarly tough sell. For this to work, some difficult top-down decisions need to be made by execs in both companies, because you can bet that few people in the trenches will willingly step aside and see years of their hard work made redundant.

Let the games begin.