Yeah, put your critical business apps in the cloud
Google App Engine - the development and hosting cloud that serves up third-party apps and websites - was on the fritz for a good six hours this morning. According to a public note from Google's App Engine team, the service began experiencing problems at around 6:30am Pacific. "We're currently seeing elevated Datastore latency …
Seriously thinks this shit is a good idea -- and *why*?
Not saying that Google are perfect or that they actually follow their "do no evil" motto, but....
How much exactly does cost to put your apps somewhere else and get the same level of scalability, stability and security? I'm tired of hearing statements without supporting evidence that you can do better, cheaper or faster. I know that there are some other key issues that you must be aware of when you commit to GAE, but as with any business decision those should be weighted against the cost.
Please substantiate your dismissal of GAE or shut up.
... aside from putting all your eggs in one big basket with everyone else's eggs.. then waiting for the basket holder to drop the basket.
Cloud computing is questionable at best and reckless at worst, your uptime is not yours to maintain, your data is less safe from others (legal or not) and your costs are arguably higher!
True, as many readers here are likely to be admins that see a threat to their continued careers in outsourcing to anyone's cloud, there will be many that think the whole idea is just pants. But, it is a good example of the possible issues - highly-redundant systems/datacenters are only as good as the app itself, and losing six hours of business is not a good selling point. Imagine if that had been a billing system for someone like a concert ticket seller, they often see big spikes in sales when they announce a popular concert such as the Michael Jackson ones, where all tickets can sell out in less than six hours.
@ AC Friday 3rd July 2009 06:58 GMT
AC Said: "Please substantiate your dismissal of GAE or shut up."
I've nothing to prove to ANYONE about ANYTHING here, AC, but as someone who is a potential consumer of cloud services or as someone who IS called upon to make recommendations of a particular strategy or technology, I'm sure you would reasonably agree that the onus is NOT on ME but THEM to prove anything regarding scalability, reliability and availability,
What I have seen and KNOW about this subject leaves me deeply unimpressed on all these counts and this story merely serves to support my current belief framework.
So what's in this story that makes you feel you can ask anyone to shut up? Its arse-backwards statements like yours that give computing a bad name in management circles.
@ Matt Bryant
Matt, I'm not an admin so I have no vested interest in my comment. Nothing would give me greater pleasure than to have a totally reliable, available and secure networking platform with massive throughput with cast-iron uptime but I am a realist and I know that this is very unlikely to happen in my lifetime.
I speak as a developer with substantial networking experience and I am all too familiar with the mathematical shortcomings of Shannon's principles. This is a discussion where these limits haven't even been mentioned (and are unlikely to be in any meaningful way) so I am as much entitled to my opinion here as you or AC here, thanks.
systemd'oh! DNS lib underscore bug bites everyone's favorite init tool, blanks Netflix
Biting the hand that feeds IT © 1998–2017