What could possibly go wrong?
If you thought cloud computing was complicated now, just wait until next year, when pin-striped traders will buy and sell contracts in the stuff. At least, that's the plan of the Chicago Mercantile Exchange, which announced on Monday that it had signed a "definitive agreement" to build a commodity exchange dedicated to the …
Brilliant. Alien territory ... where one needs to know everything based upon experience tempered with invaluable hindsight to develop unbelievably successful foresight.
And as you say, what could possibly go wrong in an imaginative sub-prime fiat currency generator market place, which is surely what options and futures and derivative market trades are, with traders completely ignorant of the actual product and technology they be selling to punters expecting to make a fast return buck for nothing they have done other than believe in the hype of an imaginative sub-prime fiat currency generator market place, which is surely what options and futures and derivative market trades are, with traders completely ignorant of the actual product and technology they be selling to punters expecting to make a fast return buck for nothing they have done. Such is a kind of flash magic which carries the risk of releasing dark voodoo.
I recall a Punch cartoon from the 19th century in which a group of expensively dressed men clutching prospectuses are all assuring anyone who will listen that they are all totally above board and have the buyers' best interests at heart.
Of course, they all have their fingers crossed. There is nothing new, says the Preacher, under the sun.
Cloud resources are NOT as mobile as guys like this think. Most companies depend on MPLS or dark fiber connections into the data center of their choice to guarantee quality of service. As a result, it can take months or more to be able to reroute traffic efficiently to a secondary party.
Computing resources are far too easy to simply build for yourself as well. Call a company like Cisco or HP (or if you're really really desperate, DELL) and ask for a data center in a box and they can ship it. If prices fluctuate too heavily because of market trading, customers will just go back to building their own data centers which will be extremely harmful for the planet.
Also, computing resources drop in value almost daily. How would the handle this? There's not a single system that makes any sense for this type of trading. Each time Intel or AMD release a new CPU, the value of CPU cycles drops because cheaper hardware can process more information. RAM drops fast as well.
Yep... this is pretty amazingly stupid. I should build my data center out in preparation of the fallout. I could probably make a fortune by just picking up customers who want predictable service and prices.
I don't think you've understood how trading futures work (CME doesn't trade equities, only options and futures - the article title doesn't make that clear). Futures are all about predictable service and price. Futures only requires 2 things:
1. A provider willing to guarantee that for a particular period of time in the future he will provide a specific service for a specific sum of money
2. A consumer willing to take that service at that cost
Everything else that happens in the middle is fluff.
If I was running a cloud service, and was able to sell a contract for £10 to provide a service to whoever held that contract from 1st Jan 2015-1st Jan 2016 at a price of £5/month then assuming I'm reasonably confident that I can indeed provide that service for that price and still make a profit it's guaranteed income in the future for me (plus gets me a bit of cash in the drawer right now).
If I'm a commodity broker who sees that the cloud provider is selling a contract at that price, and believes that the going market rate for that service during 2015 will be, let's say, £10/month then I'd happily buy it, because just before 2015 I sell the contract for £50 to someone who *actually* wants the service (i.e. is in the right location or whatever), at a profit to me (£40) and an overall saving for them (£110 instead of £120).
To address your points:
1. At the point that an individual actually wants to make use of that service they will buy that either from a broker with the right contract for the contracts value (unless they had the forethought to purchase such a contract early on) or from the cloud provider themselves for the market rate; the overall prices will be comparable, the only thing that changes is who makes the profit.
2. The contract value is somewhat divorced from the actual service cost. Wheat futures are traded well in advance of farmers growing and the price of the actual item varies wildly depending on the weather, etc., etc. - but the contract is to buy at a set price no matter what, so the value of that contract will increase/decrease over time as conditions change (i.e. my contract that I bought for £5 for a tonne of wheat at £5 when the real cost of the wheat is closer to £50 can be sold on for anything up to £45).
3. Please do build a data center. Would you consider accepting some money now to sell a contract of service at a set price that you won't need to deliver on until you have built it, thereby allowing you to build it? If so, you've just traded your first cloud derivative on CME. :o)
PS. I'm not 100% convinced it will work just because a pig is a pig - you don't have lots of choices, how many legs it'll have, how many ribs, etc. Cloud contracts can be very variable. If there's a market for such futures contracts then maybe it'll bring providers into conformity, but that in itself is unhelpful to those people who want to tailor their cloud solution for their specific requirements - like buying an off the shelf web hosting package with a lot of crap you don't want just because you wanted ssh access.
Absolutely brilliant. For the exact same reasons as retail electricity, billing for cloud services will become unpredictable and volatile due to the whims of this 'commodity market', having been disconnected from the actual costs of providing the services and exposed to the greed of market traders and their relentless drive for $profit$.
...which will trap companies into HAVING to buy forward contracts off the exchange to protect themselves from movements in this new market.
Which will therefore lock in the flow of cash to the exchange and the brokers. Economic success (the cloud coin is being successfully clipped, and somebody is getting paid for providing a service which is only necessary because ... they are providing it.)
Now you know why bankers get paid so much.
Come off it. Futures markets exist because they take the gambling out of the equation for the consumer/provider, and give it to those who deal in and have experience in that area (the bankster).
It's primary advantage is that it allows a certain amount of certainty about the future, which very quickly diminishes risk for those all those involved, from investors, suppliers, consumers, and the entire chain in between them all - as described by posters above.
Cloud computing, aside from being ill-defined (but can probably be offered as CPU cycles, or storage), is no different. If trading it offers no advantage to building it yourself, as you suppose, then the market will dry up, and all the speculators will be left high and dry. Hell, Amazon AWS pretty much does all this already - you can buy cloud at spot prices or at pre-determined prices in the future, and they, together with their customers, are doing well.
Glad someone posted how Amazon AWS is alerady offering different pricing models. If a company has to process Super Bowl ad driven web responses, they can manage the risk of not being able to scale out on demand by paying more. A university number crunching batch run might be willing to release CPU cycles and finish a few hours later than planned to stretch the budget. Within a cloud vendor such decisions could be made quickly. Reminds me of how setting batch vs interactive priority allowed the AS/400 to achieve near 100% capacity utilization with sub second interactive response time for online users.
The CME contract would be targeting longer range time spans; however I wonder if there is enough natural variation (catastrophic storms, terrorist attacks on infrastructure, memory prices, rare earth pricing, megawatts, etc) to justify a contract market. The idea is intriguing. The success of the idea would depend in part on how quickly an application or a chunk of scale out could be provisioned to run in another vendor's data center. I didn't see a description of provisioning time granularity in the article.
To the person remembering Enron - their concept was to create a market for network bandwidth after having successfully creating a market for electrical power. Markets for power contracts exist today, I believe the bandwidth on demand market is dead due to vast over supply of the early 2000's.
There was an essay in "The Atlantic" written in 1979 during the second Arab oil embargo against the US for support of Israel, where the question was raised how to break OPEC's power. Apparently setting up oil contracts market through the New York Mercantile Exchange was part of the answer. The Arabs kept taking dollars, they stopped their embargo strategy.
"To the person remembering Enron - their concept was to create a market for network bandwidth after having successfully creating a market for electrical power. "
That would be this AC here, I assume (Hmm..). My intention was to draw attention to the fraudulent aspects of Enron (there are credible accounts of entirely fake bandwidth trading desks, for example), and the spectacular incompetence of their auditors, not to question the basics of such markets - which, properly constituted (and operated by mostly honest people), can be extremely useful. However the early days of any market are a goldmine for the less than honest, as are markets which are either obscure or less than transparent: it is hard to imagine Bernie Madoff, or indeed John Law, managing to do what they did in a fully transparent market with competent regulators and auditors that actually bothered to perform audits.
If it works then there's good reason to believe it will be a big boys game - those with the clout and muscle to predict costs, maintain margins, volume and play the game. What's the impact going to be on smaller providers (i.e. everyone else that isn't Amazon, Google, Microsoft, Rackspace & IBM...) - can they survive? I've written a blog published today on Compare the Cloud asking these questions of our community.
With trading services such as power etc you are trading the delivery of something, e.g. Electricity. With hosted services you are sending them stuff to process not the other way?
With servers although they are trading contracts for usage of server tin, aren't they actually trading who gets the contract to hold & process your data?
I can see this going very badly.
Biting the hand that feeds IT © 1998–2021