It sometimes seems as if the whole world has gone cloud crazy - well at least most of the vendors, pundits and many in the media. If we listen to the evangelists, the days of the enterprise data centre are numbered and players like Google, Amazon and Microsoft will inherit the earth. Even David Cameron, the illustrious leader of …
Comparision would be great....
Some situations showing different costs would be good - although that could be quite complex depending on needs - ranging from small business, to large enterprise business - to people like me - someone hosting a handful of websites for friends and family,etc. (and like to have a server to "play" with) who's bandwidth and resource needs are next to nothing.
It did take me a long time to work out which solution was best for me (price vs all the options) - if some form of "here is a platform, we only charge per month for what you use" was out there, that'd be great (as mentioned, I hardly put a strain on resources (simple LAMP stack) and a couple of gig's of disk space).
I think that there are major shortcoming in the current hosting model that are not addressed at all in this article, and are specifically relevant to Cloud computing (IaaS). For example, when I price out a dedicated server, 90% of the available systems do not have RAID redundancy for data storage, and even if they do, they are still subject to a single point of failure in the system itself because they run on a single server. To me IaaS is not about temporary or burst requirements (as nice as these features maybe) it is about redundancy in hardware...
If I have 100 servers clouding up my virtual host, instead of a single partitioned server, I've got hardware redundancy and a far greater assurance of up-time and reliability. This is the real crux that the Cloud solves, and is (almost) non-existent in the current hosting market. In other words if the Cloud can competitively price against a dedicated server, then it is more than competitive in the current hosting market, because of the real gains in redundancy and improved reliability.
It's the utilisation... dude
Disagree, because classic hosted servers need to be sized for some imaginary "Peak Load", resulting in very low average utilisation on the equipment.
The elasticity cost overhead associated with cloud computing minimises when enough users average out the loading.
Thus at some usage point a cloud will be cheaper to run than classic hosted.
I expect this to be very soon - rising power costs, cheap RAM, Core i7 chips, tuned virtualisation technology, combined in-house & external clouds, cold standby of cloud resources, ...
By way of example GMail costs about 1/3 that of other email solutions, in large part to efficient hosting. Customers will naturally head to the lowest cost equivalent service, and that will just happen to be a Cloud provider.
Re: Other Shortcomings
The discussion about reliability is fine and dandy for hardware, but there lots of other levels not addressed beyond hardware (LAN, WAN, geography, etc.). I do think the built-in redundancy is a major consideration, but that's just another area to include in the comparison.
I strongly agree that too many folks are comparing dedicated hosting to cloud and bypassing the obvious comparison with shared hosting models (at least obvious to me). Anyone truly considering a cloud model should have to long look at virtual servers as well (especially those hosters offering addons like VMotion, clustering, etc).
It's nice to have an occasional article that has a strong IT angle and no need to wonder when the Playmobil version will be published. Now, back to more Swedish lesbian love children!
Amazon will rule the World
Your capacity is elastic, your data is automatically DR-ed to at least one other site and you've got loads of professionally built machine images you can copy for your builds. Then you've got huge tech companies devoting top brains to security (they get it wrong; their dead). I'm struggling to see any flaws.
Agreed: Not out of business, but will change us
No, the Cloud movement (aka. centralisation or IT-services-over-the-wire) will not put hosting companies like us (Memset) out of business, but it will change who our main customers are.
For example, we are seeing a lot more business from companies who want to use our infrastructure as the back-bone for their SaaS - we can run the IT layer much more effectively than most operators, since for almost all applications the actual required hardware is fairly small in terms of number of machines these days (thanks to Moore's Law).
Further, although Amazon EC2 & co. are appealing to a niche market (those few online services that have very peaky loads) the vast majority of people we speak to that want some hosting for just a single virtual machine (or a resilient pair), to carry on running a legacy application that used to be on a physically dedicated server which is now out-dated hardware.
Personally, I think the threat of 'Cloud' will be to the premium hosts like RackSpace who have made lots of money exploiting the view that "good hosting is expensive". We, on the other hand, are embracing the commoditisation of hosting and have already taken the fat out of our prices (hey, any excuse to turn the knife in my arch-rival, mmkay? ;)
Kate Craig-Wood (aka. Famed skydiving hosting-biz queenpin)
RackSpace have a cloud offering.
As for the authors premise:
"Will cloud put traditional hosters out of business?"
No. big hosters who have demand for it, will create their own "cloud". ( see rackspace example )
Small, shared hosting providers who do not have demand for it, will continue selling $5 per month plans.