Amazon Web Services: No Open Cloud Manifesto for us
Summary: Update: Amazon will join Microsoft as two big cloud computing players not signing on to the Open Cloud Manifesto.The manifesto, which has raised a ruckus following a Microsoft blog post, is set to be released Monday with IBM as the ringleader.
Update: Amazon will join Microsoft as two big cloud computing players not signing on to the Open Cloud Manifesto.
The manifesto, which has raised a ruckus following a Microsoft blog post, is set to be released Monday with IBM as the ringleader. Given the hubbub it was only natural to wonder where Amazon Web Services, one of the premier cloud computing players stood. Here's what an Amazon spokeswoman had to say:
We just recently heard about the manifesto document. Like other ideas on standards and practices, we'll review this one, too. Ideas on openness and standards have been talked about for years in web services. And, we do believe standards will continue to evolve in the cloud computing space. But, what we've heard from customers thus far, customers who are really committed to using the cloud, is that the best way to illustrate openness and customer flexibility is by what you actually provide and deliver for them. Over the past 3 years, we’ve made AWS available via multiple platforms, multiple programming languages and multiple operating systems – because that’s what customers have told us matters the most to them. We’ll continue to pursue an approach of providing customers with maximum flexibility as the standards discussion unfolds.
In any event, we do believe that standards will continue to evolve and that establishing the right ones, based on a better understanding of what is needed, will best serve customers.
Translation: It's a bit early to be tossing a manifesto around about cloud computing standards. This should be a fun conversation.
Also see: Cloud computing and the return of the platform wars
Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.
Talkback
Isn't the main advantage of "The Cloud" being Cross Platform Comp.
Beware of proprietary cloud lock down!!!
1) What happens if my cloud goes out of business?
2) How do I get my data out of the cloud in a usable format?
3) How do I shift my applications to another location if I decide I like the software but not my cloud provider?
4) How do I integrate my multi-vendor cloud services with my internal applications and provide a common view to your users?
The short answer is that it is almost always best to deploy open source software whether locally or in the cloud to prevent proprietary lock down. Virtual Servers managed using industry standard tools hosting open source software is a great way to go....
True to an extent
However, if you're writing a cloud-based app, that's nearly impossible. Amazon and Microsoft are both creating platforms in the cloud, which require re-writing your app. Whether there will be an option in the future to run their cloud platform in your own datacenter (like IBM's "Blue Cloud") remains to be seen.
Big Iron
Cross platform, by the way, results in extremely silly and facile applications. I've got a bunch of crap which was developed to run on most any OS and it's always junk, built to the lowest common denominator.
Then and Now
IBM 3270 Emulation, Character-based, SNA, Twin-Axial Cabling, Monochrome Dumb Terminals.
Now:
Internet BroadBand, TCP/IP, Thin Client, Ethernet, Xen Virtual Desktop Infrastructure. The user of VDI doesn't sacrifice functionality--all of the Destkop features remain accessible, except they reside in the Data Center, where they can be maintained efficiently in lower-cost consolidated Virtual Machine Farms.
BIG BIG difference. A whole new ballgame.
Tell us another story Mr grd003. Please? ;)
I'll take two, please...
That's OK. You are probably afraid of changing the IT status quo
I think you're missing the point
Now transition to cloud computing. What if you only need 4 PCs for one month's time. Amazon charges about $25 per week for PCs to be running 24x4. $25 x 4 PCs = $100 per week. $100 x 4 weeks = $400 total. Now what if you buy physical PCs instead? 4 PCs x $700 per PC = $2800. Plus, you have electricity costs for running them for that one month's time. Then, you have to figure out what to do with them after that - you now have depreciating assets that you don't need. That is just one of the many logical uses for cloud computing.
Cloud computing isn't something you buy, it's something you rent. It's a service, with contracts that can be cancelled at any time, on demand. It's much different than mainframes. Would I put all of my user desktops in the cloud? No. But there are many logical uses for it.
how do they access the cloud????
If I get a free 1 GB internet line and a free KVM setup for each user and free cabling to connect it to my free in house computer system using my free VM/Citrix setup then you might have a point.
Otherwise I'm going to solve your one month problem by calling the local rental company. For under $1,000 I'll have 4 machines installed with 1GB networking. I'll require no in house time to do this. My local 1GB connection will be so much faster than the 10MB internet line that I'll be saving about $100 a day in payroll on the faster system. So my approach will cost $1,000 and yours will cost $2,400 for the month.
Bzzt! Wrong
The reason corporate America "went to Personal Computers" is because IBM released one, and (originally) it instantly became a status symbol (due to the IBM logo on it). This was back when software was scarce, (and the only widely available languages were the 1980's-era BASIC dialects, with all their line-numbered spaghetti code) and these machines were on the desks of high level business execs who couldn't program (and frankly, didn't have time to do so, either), and most of them couldn't type, either -- they went to college when English majors typically made a lot of money typing papers for OTHER students, charging PER PAGE. Few business, and even fewer management majors ever learned how to type back then. They EXPECTED (and did) have people around the office to type documents for them.
Then, after the Big execs ALL had their IBM PC boat anchors on their desks, collecting dust while looking [i]impressive[/i], the NEXT way for an exec to prove his superior status over other execs was for his secretary to have her OWN PC on her desk...soon that became commonplace by the early 1990's... so then, the NEXT way for an exec to prove his importance and status was to have as many subordinates as possible with PC's on THEIR desks, too (By this time, it was the mid 1990's).
And at that stage, what happened, was the non-centralized control you praise so much... also known as ANARCHY and CHAOS. The COST of this chaos (data loss, and mis-coordination through incompatible file formats, malware, etc.) was SO high (in terms of viruses from people installing whatever crap they found on the internet or brought in from whatever computer shop they found)...and most of all, it was COMPLETELY UNWORKABLE...
How unworkable? Unworkable enough that companies every year part with the significant overhead of having one Windows admin for every 10-20 desktops.
Corporate departments never had such high overhead among mainframe terminal users. For one, the staff only needed to administrate and maintain a few centralized machines, and voila, company-wide software upgrades take place IMMEDIATELY.
In contrast, the fleets of desktop machines are a waste for the typical desktop user who does nothing more than data entry/date retrieval and document preparation in some resource-hog WYSIWYG editor.
No... all in all, both corporate America and small businesses would be better off if most desktop machines were thin-client workstations, with software either running off of the central host OR better yet, auto-mounting NFS filesystems for the software, and the configuration files on the same centralized machines, accessed via files mapped to the machines via NIS (Network Information Services, which was, years ago, called "yellow pages").
Interestingly, SUN Microsystems developed both NFS and NIS, and that was also in the late 1980's, when their trademarked sales mantra (on every advertisement) was "The Network IS the computer."
Regarding cross-platform... there's trade-offs. The good--data storage and accessing concepts are programmed according to long-established ideals of the computer science field, and then platform specific libraries translate those ideals to the (generally borked on Windows) mechanisms of the underlying platform. The advantage -- both the vendor and the customer are free from platform lock-in. The bad: more CPU cycles. I'll put up with that trade-off any day.
game changer
I'm in the same school as the response from Fred. Here is one person's experience. My wife is a programmer and mathemetician. She started in military intelligence when big data center filling iron was king 7090, 6600, early 360,etc writing in BAL and Fortran. She wrote a numerical control system for a strucutural steel company to run 100 foot steel cutting tables. The program ran on a 370. It took four people doing data entry working from drawings to produce mylar punch tapes to keep one production table busy. She got to re-write this program on the first PDP-11 LSI which was set up as a single user system and had a CRT with some graphics capability. Now one worker could keep four production tables busy.
When 100% of the computer was used strictly by the local single user productivity went up by a factor of 16 to 1. After this sunk in she quit that job and we started our own company, bought our own Unix based mini from Texas Instruments and have been self employed ever since.
In the late 70s we were selling CP/M software for $3,495 a copy to the vertical market we have served since we went into business as well as doing custom programming for other local businesses, for example doing remnant management for a large carpet company. There was no way these applications would ever have become available on a mainframe.
Rubbish
Taken at face value, you think the readers here are to young to
understand the following statement - which is quite arrogant. But
allowing for the translation that you think we are too young to
remember when it occurred, I've got news for you - electronic
calculators were an interesting novelty about the time I left school.
Yes, I learned to use a slide rule. Yes, I am certainly old enough to
have [b]experienced[/b] the transition from mainframes to PCs. And
so are a great many other readers I expect.
"[i]we went to personal computers in the first place[/i]"
I know exactly why: because the vendors of mainframes and minis
were seriously ripping-off their clients and PCs were seen as a cheap
alternative. PCs were originally implemented as word processors and
for spreadsheets, they were more flexible than mainframes. But
sharing files was usually done by swapping floppies or teaching
people how to use FTP. Not fun.
"[i] why any personal computer is better than any mainframe (server, if
you must, cloud and all).[/i]"
"Better" is comparative, yet you provide no basis for the comparison.
PCs are very expensive to maintain and support. They have far less
processing power than mainframes of the same era. How are they
"better"?
"[i]When the computer is mine it is always cheaper, always more
flexible, always has the things I want (when I want them) and not what
someone else thinks I need, and their central control is always limited
in availability and features.[/i]"
That laughable. The client sites I have worked on lock down their PCs
to the extent that I can't even change the desktop image or screen
saver. Install any software I want? No way! They are locked down as
tight as any mainframe-based environment.
"...[i]I lived for many years under the tyranny of big iron.[/i]"
Yes, monopolies are bad, aren't they? It doesn't matter what type,
they all screw their customers until something ground breaking comes
along and changes the game. You should feel happy that IBM stuffed-
up their potential monopoly of desktop PCs.
"[i]All of you need to slow down for a minute and think through this
issue. [/i]"
I think you need to stop assuming we are all as petulant and short-
sighed as you appear to be.
"[i]Cross platform, by the way, results in extremely silly and facile
applications.[/i]"
Cross platform what? The key is having data that is not locked into
proprietary formats. The ability to freely share data and documents is
what matters, I don't care if applications themselves aren't cross
platform, though it has been a holy grail for decades - ANSI C was the
first attempt I came across, BASIC was similar, then there was Java.
But open standards for data formats are the key, who cares about the
application?
"[i]I've got a bunch of crap which was developed to run on most any
OS and it's always junk, built to the lowest common denominator.[/i]"
There will always be low quality software, if you choose to collect it,
that's your problem. One of the first CAD packages I used was
Microstation, at the time it ran on VMS, UNIX (several flavours), Mac,
MS-DOS, PC-DOS and many others. It was almost identical on all
platforms and continues to be one of the top CAD packages today. It
is great software. Cross-platform applications don't have to be bad,
the MS Office suite is cross platform - it's probably one of the worst I
have used but is still quite good.
It will be interesting...and time will tell.
Oh hang on a sec - that construction company just cut the cable line to my company...I will have to dial up to finish that document/sales order/retrieve that file get... that I owe you by noon today for the contract - ooops.
The glass is half empty....in two, three years...
Get out of your pessimistic doldrums and look at it from a different perspective.
Cloud Computing has been around FOR YEARS. It's now being labeled differently: used to be called Virtual Private Servers. Only now, it's on a broader scale.
Amazon is being innovative, taking their existing infrastructure and using it to make money combined with new technologies, namely Xen Virtual Machines, Linux and some really smart programming.
This isn't like the internet bubble. It is a growing 'profit center' and is viable because it is 'cost driven'--namely fueled by a down global economy.
Economy is a player
There you have it. Thanks adsanders!!
RE: Amazon Web Services: No Open Cloud Manifesto for us
Lots of reasons
Another, more in line with your comment, is business resumption / disaster recovery. With the cloud, you get "unlimited" datacenters. Yes, you can store your docments on a server in a datacenter. But what if that server or datacenter fails? You lose connectivity? What if you are providing a 24x7 service? You have possible SLA violations. With cloud, you have seemless access to multiple datacenters to prevent outages, and you don't have any of the heavy infrastructure costs of setting them up. Also, you can do interesting tricks like "follow the sun" with your applications. Move them to the datacenter nearest the country that is currently during it's peak business hours.
Of course, the cloud providers are still a work in progress. Amazon and Microsoft have both suffered outages. But Amazon is in production with real contracts, and they provided refunds to their customers.
Greater > Lesser
To be completely unbiased, I see it like this:
If I were a large company with a good or better cloud environment, I would not want to participate and instead stay closed and proprietary (=more money).
If I were small and upcoming of course I would push for an open solution (=more money).
What surprises me more is when the large corporations take part in things like this that *don't* have an immediate reflection on their bottom-line.
Running scared and trying to grab control