scale fail: cloud addiction

I've always been wary of anything that is being hyped.  Until the hype dies, I'm prejudiced against it, and I won't give it a fair chance.

This is not just for technology issues -- I do that to movies and books also.  I still won't bother to read Da Vinci Code, although I finally did see Forrest Gump years later, when I became convinced it was OK.

But it's much more true for technology.  Here, unlike a movie I have not seen or a book I have not read, I can actually expect to have an intuitive feel for the truth already, so the prejudice lasts longer and very rarely reverses.

----

The "cloud", to me, has always been just that: a bit of water and a lot of hot air.  I can certainly see some uses for cloud computing in small and medium enterprises -- the smaller the better.  An extreme case is an individual running his own web-based business -- finding a cloud provider is ideal for him in terms of bang for the buck.

But I've always believed that the larger you get, the more you lose by going to the cloud.  At some point, the economies of shared infrastructure disappear simply because as you get bigger and bigger, you are less amenable to sharing.

Josh Berkus (of Postgres) wrote a very fantastic 2-part article series called "Scale Fail" for LWN.  Part 2 of this, at http://lwn.net/SubscriberLink/443775/a17084926dbefa54/ , has a section called "Cloud Addiction", which is well worth a read.  Here're some extracts:

Several of our clients are refusing to move off of cloud hosting even when it is demonstrably killing their businesses. This problem is at its worst on Amazon Web Services (AWS) because Amazon has no way to move off their cloud without leaving Amazon entirely, but I've seen it with other public clouds as well.

[...]

[restrictions on memory, processing power, storage throughput and network configuration inherent on a large scale public cloud, as well as the high cost of round-the-clock busy cloud instances] are "good enough" for getting a project off the ground, but start failing when you need to make serious performance demands on each node.

[...]

That's when you've reached scale fail on the cloud. At that point, the company has no experience managing infrastructure, no systems staff, and no migration budget. More critically, management doesn't have any process for making decisions about infrastructure. Advice that a change of hosting is required are met with blank stares or even panic.