Exactly. I can see that people want low level sockets. However having high level ones already available for a large common service may aid in prototyping.
Yep. As one of the biggest cheerleaders for raw sockets around - they don’t replace good high level libraries. We could wrap up raw sockets in nice friendly libraries in lua for http and so on - but for something really vertical like AWS - something Amazon themselves provides iOS libraries for - you gotta wrap the library if you want it to work right.
Having said that - I won’t use this. I may be old and curmudgeonly, but I don’t trust “the cloud”. Heck, I don’t like to depend on ‘the net’, much less stuff on the other side of it. (This is one of my main uses for Codea! Being able to do something “useful” when the net isn’t available…)
Virtualization and commoditized compute and storage infrastructure that is fungible…all services are headed that way inevitably. Too many cost and operational advantages over the big iron era.
The cloud is lovely; and like all other useful tools, there are those who will use their shiny new hammer to hit every nail they see, even if it’s not a real nail. For bursty on-demand service - cool beans.
But I work in finance (kinda sorta) - we’re not going to put your banking data on hardware owned by others, end of story.
Maybe not this year or next, but I believe it is inevitable. The services will mature, as will the reliability. And the economic advantages will tell in the long run.
Plus, the sidekick platform had other problems, not just a cloud architecture. As for Netflix and the other amazon stalwarts, people seem to be living with not quite five nines of reliability happily enough for their discounted pricing. I guess it’s great when services can remain in beta indefinitely!
There is absolutely no denying the sidekick platform has issues, nor should be ignore Microsoft’s culpability in the fiasco. Still doesn’t mean I trust cloud to not go down as far as I can throw it. Nor, frankly, do I trust it’s security either, but I think that’s just common sense.
I asset there is deep wisdom in “if you want it done right, do it yourself” - yes, even if it’s more expensive. The only time I’ve been able to rest 100% sure that my important services were as important to me as they were to the people running my hardware is when I am running my hardware. If my cloud provider can shrug and say “well, things happen - we’ll refund your money”, they’re not properly motivated enough to handle my important data. (My non important data? Sure - cloud away. I have no issues posting my .codea stuff to posterous, because if they disappear, no real harm done. But - I’ve run too many services where if I go down, I lose business (and money, and my reputation), permanently, to trust much to anyone not under my direct supervision… the trick is, knowing what can go down [and so can be trusted to a third party] - and what can’t)
Netflix can afford to be down at times - Lord knows they’re doing a good enough job ruining their own good name with their business plans. But Tax information? Nosir. You lose someone’s tax return, or screw up their internet banking - you lose a customer, permanently. (I work for the Financial Institutions division of Intuit, not that I’m here as a representative [insert standard disclaimer here] - we take this sort of thing very seriously…) The netflix usage model and uptime requirements are just a tad different than mine
Hey, I used to know some people who worked for Intuit … Small world.
Anyway, I dislike “the cloud” as well. Less for the data security (though that is a worry) and more because I have a fast machine with lots of storage right in front of me so why not use it?
Don’t get me wrong. A fast personal machine in front of you is a must-have. But there are all sorts of modes of use that require ubiquitous access to information, email, media, phone books, etc… The whole dynamic of picking up a new consumer electronics device which in most cases is yet another gadget to be configured and synced to your life is now different than a couple of years ago. Now instead of spending 4-6 hours loading An OS, mail archives and programs, you simply type in your account and off it goes to set up and synch automatically.
Now you might say, Okay, that’s just outsourcing my IT to anyone that can run a server, and as long as I am careful about who I pick to run it, what’s so “Cloud” about that? What is there to hate? Truth is, this is how Intuit and all the banks operate. The only difference in a cloud system is that instead of the bank running a server farm of discrete machines, it could virtualized servers and abstract the software services layer away from individual machines. It’s still a client server architecture, but one with completely different economics. Individual companies (and IT managers) can still run their individual virtualized clouds and apply whatever level of redundancy and fault protection they want to spend money on. But in the long run, as the virtualized systems mature and become reliable, they will inevitably take over from the discrete hardware because of the enormous savings and the increases in rate of innovation.
So rather than saying something like ‘I hate Clouds’ I would say," I hate trusting Joyent to run my cloud services." But I have to say, we’ve been using amazon’s cloud for six months or so and it has been absolutely fantastic. The short interruption was annoying but ultimately didn’t really result in any noticeable list business or other support issues. And for small startup businesses that can’t afford to invest up front in a big IT infrastructure, it’s lowered the cost of rolling out a national scale service by a factor of more than 1000!
This is a HUGE deal, where some of the traditional Telco and cable companies are just shivering in the cold, not realizing that people in general are willing to live with slightly lower reliability in return for MASSIVE cost savings. VoIP calls, for instance, have taken over from the POTS lines of the past despite atrociously worse call quality and not so good reliability. Why? Because it costs 1/10 the price to consumers. All businesses are headed this way, like it or not.
I’m old school - “physical posession trumps all security”.
Use the cloud for non-important stuff. Keep important stuff local. Been burned too many times on that to change easily now. (the correlaries are “make sure you have a damn good net connection” and “doing it right is not cheap”)
On reliability I’m with Bortels. If you want to do it right, do it yourself. But doing yourself is an investment and has its risks. You put a bunch of work into it upfront and who knows if your company is even still there a few years down the road, where if you really did it right, maybe you’ll see the payoff. The cloud lowers the barriers to entry. And if things work out then later you go big and develop your own thing.
For retail users I think it’s less of a big deal. I hate to deal with my ipad or whatever trying to sync stuff all the time. I like to know what’s where and please don’t touch it. I put it there for a reason.
As for security, well security is a theater - if someone wants to take you down, it’s not very hard. Even when you spend a lot of resources protecting yourself, security is weak at best. Witness wikileaks or stuxnet. Security measures make your users feel at ease and that you’re doing something for them.
A private cloud, in my garage, would be awesome. (the bad guys would never find it among all the junk!)
Oddly enough - we’re using AWS now for testing
There’s a difference between a DOS (bad) and a compromise (find a new job). You can prevent data compromise. It is not easy. First step is to host your own stuff. (second is to find people who know what theyre doing - its not a job for amateurs…)
For a little app just starting out cloud is the way to go.
When you have the funds to invest in house/custom can be the way to go.
If you want something that is just standard cloud is the way to go. Most in house systems are actually less secure and/or more expensive than a cloud. This is due to poor management and/or funding. Think small business or a small local government but there are examples of a few notable larger organizations that our poor also.
That being said even though most in house is worse, there are still an easy million places it is better. Where it is better, it is a market differentiator. If you have a good thing going don’t change it, a good IT department doesn’t just happen and likely never does.
And with all things the real answer is a grey area and a hybrid approach. Outsource what you don’t have time and money for, focus on the rest. Just make sure to listen when moving the pieces. Too many times pieces are forgotten in the cost analysis including the need for flexibility.
There are good sci-fi novels that refer to anything from the past 200 years as the unknown era, the cause being the rise of paper then bits.
When the zombie apocalypse occurs, make way to the google server farms. Not only will they contain most knowledge, they tend to be self-sustaining for water and energy.
