October 11, 2015 at 12:09 am
Comments posted to this topic are about the item Deploying tech to the cloud
October 11, 2015 at 6:46 am
To be honest, I don't see anything in your article that should be different between how an application should be developed for the cloud or for non-cloud apps except that the cloud app needs to be simple, which is a serious disadvantage of the cloud.
--Jeff Moden
Change is inevitable... Change for the better is not.
October 11, 2015 at 8:11 am
Exactly! There is nothing magical about the cloud and in fact to get the most out of it requires an attention to detail that would benefit any environment.
I feel that the cloud has been a useful catalyst to promote disciplines and thrash out new(ish) ways of working
I particularly like the idea of Netflix's Simian army. Randomly switching stuff off, adding latency etc. If your system keeps running then your design is robust
October 12, 2015 at 7:30 am
Has anyone run numbers on if this way of doing things costs more or less and where are the break even points?
412-977-3526 call/text
October 12, 2015 at 7:31 am
This was our philosophy when we built Chaos Monkey, a tool that randomly disables our production instances to make sure we can survive this common type of failure without any customer impact. The name comes from the idea of unleashing a wild monkey with a weapon in your data center (or cloud region) to randomly shoot down instances and chew through cables -- all the while we continue serving our customers without interruption.
http://techblog.netflix.com/2011/07/netflix-simian-army.html
Unleashing a wild monkey with a weapon into the data center sounds like a ton of fun. It would be worth the cleanup afterward just to see the look on the network admins faces when that crazy sh!t goes down. However, it's probably less disruptive than having 40 million Netflix users attempting to simlustaneously stream the premier episode of a new series. 😛
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
October 12, 2015 at 11:22 am
Chaos Monkey etc sounds like madness but the more you think about it the more it makes sense. At one time garbage collection sounded like the ravings of a lunatic but now is a "so what" topic.
We've tested some of our stuff just randomly destroying instances as Chaos Monkey would and its recovery was almost seamless. Certainly nothing that a quick F5 wouldn't cater for. I'm delighted it worked but not so happy that it was our main source control repository that was the victim of the test.
When I heard what had been done I felt the same icy hand of dread around my heart the morning I had forgotten that I had had beetroot for tea.
October 21, 2015 at 7:39 am
Actually I kind of got lost.
November 4, 2015 at 5:40 am
Great article showing that advances for one specific use helps to highlight that some previous techniques and practices were suboptimal e.g. installing a system with a specific configuration that is a simple and repeatable process shouldn't be as hard as it is for many systems.
Gaz
-- Stop your grinnin' and drop your linen...they're everywhere!!!
Viewing 8 posts - 1 through 7 (of 7 total)
You must be logged in to reply to this topic. Login to reply