Ha well said
Thanks Clinton, private conversations make for better conversations sometimes. I will make sure I rationally answer the cloud questions here though too and try not to pick fights.
Ha well said
Thanks Clinton, private conversations make for better conversations sometimes. I will make sure I rationally answer the cloud questions here though too and try not to pick fights.
News: Foursquare, Quora and other sites go down in Amazon EC2 failure today.
They are back online now or on their way back.Scores of well-known websites have been unavailable for large parts of Thursday because of problems with Amazon's web hosting service.
....
Amazon EC2 is the retailer's cloud computing business. It provides processing power and storage to companies that do not have their own data centres.
No reason has so far been given for the outage.
Are these "cloud" hosts more prone to downtime and outages?
Show your support - Like us on Facebook
I don't pick arguments Clinton, that would imply a desire to get into arguments and in fact I'd much prefer it if people just agreed with meOriginally Posted by Clinton
JJMcClure - Rich, as his friends call him - has access to the VIP Lounge and that's despite all the arguments he picks with me!I read stuff, I sometimes say what I'm thinking.
Love to mate but I don't think I can add anything since you're far more qualified to argue the other view than I am. It's been a very informative discussion so far though.can you start us up a conversation please Mr McClure ? Clinton is way too good at doing that already
I recently had a total nightmare with Streamline and Fasthosts, two of the ISPs I use for hosting websites, I think I posted about it. I couldn't access some of my sites for weeks, literally. It took me over two weeks to be able to backup a database on one site and move it to a new host and I probably spent a full day sorting it out overall. It turns out that Fasthosts and Streamline are both part of 1and1 and there were multiple failures throughout the system, and that's with one of the biggest ISPs in the country.
Obviously no online service is going to be 100% secure or reliable but as I said in my OP, neither is my PC or the external drive I use for backups. You're taking your chances wherever you store data and I think that the more dispersed it is the less vunerable it is and the easiest way to do that is online. If I had to choose between my PC and online storage, given that the security issues pretty much even out, I'd choose online simply because I can access it from anywhere with an internet connection and if there was no internet anymore I'd be out of business anyway so it wouldn't matter.
IN DEFENCE OF THE CLOUD !
Yeah ! Amazon EC2 had issue recently and predictably every man and his dog with an axe to grind (LOTS of 'technical' people especially) are using it as an opportunity to have a good dig at the cloud. Remember I said it was a marketing term ? Expecting high availability from a marketing term because you believe that the market term means high-availability is ALWAYS going to leave you disappointed.
We have learned some valuable lessons though, the biggest of which is that cloud based systems (lets call them conventional hosting systems) can in fact fail. Of course I knew that we all knew that and I will slap anyone who tells me cloud can in fact fail with my stick which has "if you want high availability, you must implement high availability" written on it.
Its very unreasonable to expect that if you put your website or whatever into the cloud, that you will automatically get high availability just because its cloud. Sorry, imagining high-availability to be magically available just because its in the cloud is a recipe for trouble.
Serious companies using EC2 implement serious high-availability strategies and architectures, which is what made EC2 customer NetFlix able to continue delivering services even though almost every other customer on that same cloud had major issues. NetFlix clearly sat down and worked this out waaay before the issues occurred, they had correct process in place to make sure they were ready for when they got hit and they were. Netflix was still open for business when everyone else's websites were down and DELIVERING HD VIDEO, how much awesome do you want ?
Only those businesses that failed to perform even mediocre availability design on their cloud (EC2) based systems have experienced prolonged outages as a result of this week’s event. This is only as would be expected Amazon does not promise to make your application highly available. What Amazon EC2 provides is a rich set of tools that allows anyone that is serious about building a highly available application to do so.
The main criticism of Amazon is its inability to re-route capacity between its data centers t ensure that websites of its customer were still up, but again why is it Amazons job to do this ? The people behind this criticism are talking about Amazons availability zones and those of you who have played with EC2 will know that there is NO AUTOMATIC mechanism to re-route capacity between zones unless your infrastructure has been specifically designed and implemented to switch resources between availability zones.
If you want your application to survive the failure of an availability zone, you must frequently back up (snapshot) your disks so that you can re-instantiate your system in a surviving availability zone. In this week’s outage, all but one of Amazon’s US East availability zones were functioning normally within about four hours. Only those customers with systems in a single availability zone in US East could not reliably access their data in their Elastic Block Store (EBS) volumes. If those customers had simply performed regular backups (snapshots) of their volumes, the outage would have been confined to a few hours, not 40+ hours. No, none of this is automatic unless its designed to be.
What makes matters worse though is the press and even CEO's of 'cloud' companies talking shit about Amazon and the cloud when they should be talking to their own engineers about the outage. It they spent enough time understanding their own infrastructure to realize that they should be talking to their own engineers about why they failed to design a robust multi-zone backup solution on EC2, rather than imagining capabilities for EC2 that do not and have never been asserted to have existed just because its "cloud".
I for one do hope that the results of this debacle will be more careful and comprehensive engineering of solutions deployed to the cloud, but I think that instead the media coverage and customer responses will dictate the conversation and that the blame will not be placed where it belongs, but on top of Amazon and the general idea of the cloud instead.
It is not deserved and anyone who holds this up as an example of why the cloud can be bad needs to be slapped around the face for not knowing what they are talking about.
THIS is why I get angry and very rude at people who say things like "Cloud is bad because gmail once went down". The devil as always is in the detail and most people prefer not to inconvenience themselves intellectually enough to understand this properly before speaking.
Remember the old phrase "Computers do not get things wrong, humans do." ? Same thing applies with the cloud.
Now if you were to destroy my California DC with a Tomahawk missile and send my desktops down, they would pop up in Texas shortly afterwards, if you then took out texas, they would pop up somewhere else until I ran out of DC's, but by then we would have simply live migrated our infra over to the Amazon/Rackspace cloud, its not hard to do if you are ready for it and prepared.
If I tell you I can guarantee you 100% uptime I can, but its going to cost you. The best I can do for a reasonable price is 99.9 an thats good enough for most people, but remember that this uptime guarantee does not in itself guarantee high availability, you need to plan and design you infrastructure for that and again it costs money.
Clinton (April 23rd, 2011)
Here is a real life cloud scare story for you : https://forums.aws.amazon.com/thread...65649&tstart=0
Like I said, you cannot assume high availability is a given in cloud solutions the engineers who were using the cloud to monitor thousands of patients ECG signals made a HUGE mistake when they assumed high availability was a given.
This is scary when you think about it.
Interesting article : http://www.cloudsigma.com/en/blog/20...essons-learned
The cloud and outages, five key lessons.
Clinton (April 23rd, 2011)
How can any company make me such a guarantee if this old woman lives down the road from me?If I tell you I can guarantee you 100% uptime I can, but its going to cost you
Backing up to an online repository adds one extra layer of risk - a failure in the access route to that data store. This risk comes in many forms: if my internet connection goes down, my ISP goes bust, my router goes faulty, there's a power cut ....
With a local copy of the data I can still access my own data whatever happens to my ISP. I can even fire up my laptop without external power if I need to urgently access a file.
I appreciate JJMcClure can't work if he's not online and feels access to his data would be pointless if he doesn't have an internet connection. That's not the case with the majority of data users.
Show your support - Like us on Facebook
Ha
Its a funny story but a very real threat, in response to this and your point I would tell you that any decent cloud product would allow for local caching of data in the event of a failure of the internet. Like Dropbox does as do any number of cloud based storage solutions.
JJ is dead right again. At least i am understanding it that by the internet being down, you also mean the websites which are the basis for your own businesses will be too and therefore this is why he feels access to data would be pointless if he doesn't have an internet connection.
What we fail to realise though is that everything is internet these days, take our phones for example. If you work in any big corporation everything is VOIP, all those sexy Cisco phones on the desktops all need internet, the copper wire that used to connect our PSTN telephone network has gone the way of the dinosaurs and it doesn't matter if your great aunt has had the same phone number and phone for 40 years, BT went and switched the entire national backbone over to VOIP already, like it or not we are al 'skype' users now.
So if the internet goes, then so does everything with it. Phones, banking, credit card authentication, everything. We will be back in the stone ages.
Happily though the internet is not one thing but a connection of many things. Its often more precisely called the interwebs for good reason and we are far past the point of relying on the good old copper cable and some days I really do lament its passing.
Anyway, onsite caching of data if you really must have 24/7/100% access to your data at all times. Thats the ticket !
I would give you 100% desktop uptime using something similar, in desktop virtualization we call it an offline desktop instance.
How do I connect to this local caching facility you've got if my telephone line gets blown down in a winter gale?
I appreciate the point about a lot of businesses coming to a standstill if the internet goes down. The article I linked to provides some scary stats. Too many businesses (and people's livelihoods) rely on this internet thing not going down for even a short time.
But ....and one in five (companies) said a week without being online would be the death of their company.
If the internet goes down "everything" doesn't go down with it!
If I lose connection now here are just some of the things I could continue to do:
- Start organising and sorting the thousands of family photos I've got, perhaps even caption the best ones and put them in a separate folder
- Play my music (all stored on hard disks)
- Continue writing my next article / blog post (which I always save in draft form on my hard disk till it's ready to go online)
- Get down to some video editing and improving my skills at it by editing our home videos (stored on disk)
- Play a computer game (no, I don't need to access any games server)
etc
You get the picture. Lots of people are reliant on their data, but not reliant on the internet. Even those who are not reliant on their data (family photos etc) are better off storing it locally in case that old woman brings her shovel to your town.
Would you agree that for people who can continue to function offline storing data locally is more practical and safe?
Show your support - Like us on Facebook
Posting Permissions
Reply With Quote
Bookmarks