Slashdot Mirror
Will Cloud Services One Day Be Traded Just Like Stocks and Bonds?
Brandon Butler writes "Today, cloud computing resources are bought and sold in a fairly straightforward process: A company needs extra compute capacity, for example, so they contract with a provider who spins up virtual machines for a certain amount of time. But what will that process look like in, say, 2020? If efforts by a handful of companies come to fruition, there could be a lot more wheeling and dealing that goes on behind the scenes. An idea is being floated to package cloud computing resources into blocks that can be bought and sold on a commodity futures trading market. It would be similar to how financial instruments like stocks, bonds and agricultural products like corn and wheat are traded on exchanges by investors. Blocks of cloud computing resources — for example a month's worth of virtual machines, or a year's worth of cloud storage — would be packaged by service providers and sold on a market. In the exchange, investors and traders could buy up these blocks and resell them to end users, or other investors, potentially turning a profit if the value of the resource increases."
Let's take something useful, and let the parasites make money off of our work...
Just like the stock market.
General Relativity: Space-time tells matter where to go; Matter tells space-time what shape to be.
What the hell would be the point of... any of that.
I have a hard enough time understanding why anything below the 1st sale market works, but what practical purpose does this serve.
Just sounds like yet another way for people to skim money off something without actually providing anything valuable. The benefits to the consumer given in the article seem pretty damn thin.
Also does the cost of computing really go up that often? When was the last time your VPS provider increased the cost of what you were paying for?
And finally, this all assumes providers are all interchangeable. I don’t see any motivation for that to happen. Providers want to build lock-in (or brand loyalty) like any other industry, which they do by offering provider specific tools and features.
I don’t consider myself a hippy, or a communist, but the more I see stuff like this, the more I think we really need to re-think the whole money concept. It seems to have outgrown it’s use as an abstract bartering tool and driven a massive amount of human potential into pointless and non-beneficial activities.
I'm not working in the financial industry, but isn't the derivative market covering this one already?
Like buying an option to cover potential extra computing power during the holiday periods.
In the exchange, investors and traders could buy up these blocks
Step 1. Get most of the major internet websites and businesses onto cloud architecture.
Step 2. Add middlemen between cloud providers and users who can arbitrarily increase the price of computational resources once they're locked in.
Step 3. Profit!
#fuckbeta #iamslashdot #dicemustdie
This essentially describes the Amazon AWS Marketplace where you can sell your unused reserved instances.
Neither the reporter nor the editors understand the difference between a commodities exchange and a securities exchange. The article is just a random collection of words the writer thinks would look nice next to each other.
No.
. . . just what life is missing right now . . .
Schroedinger's Brexit: The UK is both in and out of the EU at the same time!
Wall street has ruined America. So let's do the same thing with computing. Brilliant.
That's the crucial "if". Unlike most other commodities that businesses seek to control and restrict the supply thereof (such as stocks), processing power is expected to keep going up per Moore's Law for several years yet. Anyone investing now is not going to make money. After Moore's Law runs out, however, then it will depend on the total supply of processors that are built and connected as "cloud power". People would have to stop adding more to the Cloud for the Law of Supply and Demand to start increasing the value of that resource. And the only way that could happen is if brand-new businesses have no way of adding servers to the Internet --the design of the Internet itself would have to be changed. Therefore what average folks need to be on the lookout for is attempts by anyone to do just that --redesign the Internet to become a limited resource, rather than a resource to which just about anyone could add more processing power.
wheat, corn and other commodities are called commodities because corn is mostly corn no matter which farm you buy it from. food processors buy from lots of suppliers and mix it all together.
and commodities are mostly traded for price protection and risk reasons. if you invest say $1000 per acre to grow corn you want to be fairly sure that you can sell it for more than that when its ready to sell. that's what the commodity markets do, they match buyers and sellers who want to lock in their prices before the commodity is delivered to reduce risk. the speculators are a tiny percentage of the market
cloud services are not a commodity. amazon's cloud is different from salesforce which is different from google's cloud which is different from ADP
...that someone is trying way too hard to literalize the meme, "I'll give you 100 internets if..."?
No, but the data stored in them will be.
I remember a bunch of douchebags who managed to convince non-technical business leaders that bandwidth could be traded like this. They set up a whole trading market, pumped a bunch of money through it...I even worked for someone who managed to get us in to do a vulnerability assessment of their whole operation.
After we were done, the upper management of this company (the douchebags with the trading capability) came in, and shut down the meeting where we presented our findings...after which, they sacked the IT people who brought us in. Why, you ask? Because the whole thing was a sham, and the upper management was afraid it would get found out. The douchebags were Enron.
This sounds very similar to me.
For your security, this post has been encrypted with ROT-13, twice.
There are tons of companies reselling Amazon and other webservices, and they make good money by adding value like preconfigured images and other services like backup software that backs up to S3. I don't see why this is news now.
Oh wait, I get it, this guy wants to make good money without adding any value at all. Good luck with that, if he tries to corner the market in AMZN.S3 "stock" (capacity), Amazon's shareholders would happily vote for issuing more "stock" by buying additional capacity, and profiting heartily by selling storage on them too. Unlike stock, there's no intrinsic value to be cheapened by issuing more capacity. Without collusion, the speculator will be unable to raise prices on their block of capacity.
If I have been able to see further than others, it is because I bought a pair of binoculars.
This would actually be useful if it worked, but it faces the same problem as most would-be futures contracts: fungibility.
Fungibility is the property of one thing being like another, which allows them to be traded without worrying about which of the two you get. Stocks and bonds are fungible - you don't care what the serial # on your share of INTC stock is. They're all the same. But it's not clear that blocks of cloud computing or storage are fungible - is an hour on a Azure VM worth the same as an hour on a Amazon Web Services VM? Probably not.
This problem seems fatal to me.
Cloud computing is a commodity, not a stock or bond. And the answer is yes.
The main obstacle is the lack of a common standard for cloud resources. It's only a commodity if it's interchangeable: wheat is wheat no matter where it's grown, but AWS and App Engine are very different things.
Really, when has the cost of storage or compute time ever gone up over a 12 month period.
Why would I go to a market to buy a service (with a middleman markup) when I could just buy it directly from a provider? Do we really expect to have shortages of computing power that would benefit from a secondary market redistributing computation around?
Enron (remember Enron?) tried to do this for network bandwidth. Didn't work.
A futures market requires a standardized, fungible product, like oil or electric power. This is hard when the manufacturer or service provider controls the product definition. Rarely has there been a successful futures market in a manufactured good or service.
It's been tried. There's a futures market in cold-rolled steel sheet. But there's no futures market in cars or office space or air travel. Some airlines have sold options on future air travel to big customers; pay something up front and lock in the price. But those aren't tradeable.
I believe shares in CPU and bandwidth etc will rather be sold in microseconds than months and years. And it all will be done automatically, by algorithms.
It will be heaven for botnets and a return to ugly mainframes.
Are they? I mean, if the price today to run my workloads on Service A is $25 and on Service B $20, is there enough compatibility and flexibility to simply "vmotion" my workloads to whoever has the best deal that day? The same thing could apply to storage.
My basic understanding is "no" -- an Amazon VM instance isn't directly portable to Rackspace or some other service and the connectivity isn't necessarily fast enough to move the associated storage around that easily, either.
But will it get that way in the future to where you could just move your services around to whoever has the best deal, even if the timelines are weekly or monthly rather than daily?
I suppose there may be ways to do this now, but it might require a lot of intermediate layers that run directly on the respective compute platforms while the workloads are more platform independent.
Isn't all cloud storage pretty much equal and isn't there pretty much unlimited space by definition? These people are acting like they're selling real estate.
For anyone who needed proof that the cloud bubble is maturing... let me introduce the vultures.
No
The first half of Greg Egan's Permutation City is set in a future where computation is a commodity. Storage and bandwidth, the more important parts of today's cloud services, are ignored in the novel.
When cloud computing will become a non differentiable product, there will be a lot of analogies with the current electricity markets: electricity cannot be stored, or can partially be stored, and hence, like computing must be provided at the moment of request!
If aggregators, dealers, and other "middle men" don't offer you anything you want, don't use them. Simple.
Note that the grocery store, gas station, and just about every other business you use is a middle man. If the grocery store doesn't offer you any advantage over ordering items shipped directly from manufacturers and producers, you can make that choice. Sometimes, I order things direct. Most of the time, it's more convenient and cheaper to go through an aggregator / retailer like Walmart.
If you want some of the services of a middle man but not all, you have that choice too. Sam's Club and other warehouse stores sell cases at low prices, just like buying direct. Internet distributors are another in-between option. Yet, most of the time we prefer the services of a middle man, a retailer.
More on topic, I have bought, and continue to buy data services through a middle man. The backbone providers sell 10Gb connections. They aren't interested in the 50Mbps I want to buy. My retailer IS very interested in my 50Mbps account and they work hard to keep me happy. If there's a problem with one of the backbones, they have the expertise and the pull to get it fixed.
Haha, you don't want to participate in culture in its exact form it exists right now, go live by yourself!
That's not an insight, it's lazy, and it ignores any institutional support that large middlemen get.
The monetization of literally *everything* continues, who in their right mind (besides the mentally ill financial participants) be interested in such a thing?
"If any question why we died, Tell them because our fathers lied."
If aggregators, dealers, and other "middle men" don't offer you anything you want, don't use them. Simple.
What of monopolies? What of ISPs, car dealerships, and yes, even grocery stores and gas stations?
It is a 'buy from middlemen or get nothing at all' world out there. Avoiding paying your tithe to the rent seekers of the world is anything but simple.
I had this discussion with someone else a few days ago. The thing that makes "middle-men" an undesireable thing is the value they bring to the table. Walmart brings me a lot of value, I can go to their store and buy the produce of several different farms, and several manufacturing plants. Getting everything I get from Walmart would require me to travel all over the country, so I'm quite happy to pay Walmart for their service.
Putting cloud services on an exchange doesn't create value. Currently I go to Amazon and buy cloud hosting directly from them. In the future, will I have to bid on cloud services on an open exchange, where I don't know what the prices are beforehand? Who will regulate the exchange? What if I buy some server time from the market, and the server crashes: who is liable? What is wrong with the current system that putting up an exchange will fix?
Is 1563649 a prime number?
One of the problems that has to be solved is the version-combo problem.
For example, a given project may require version 1 of the database, version 1.5 of the language, version 2 of the OS, version 3 of the middle-ware, version 3.5 of Apache, etc etc etc.
Most cloud services don't support enough versions to handle specific matches, making swapping vendors difficult. You have to change versions to move your project, which often creates bugs and reprogramming effort when you move.
Also, there is no incentive for cloud hosters to make swapping/migration easy; because if they make it easy, you will leave them. Easy come, easy go. They thus pull various tricks and gimmicks to lock you in.
For example, they may give you a nice GUI/web console to make changes to configurations, but there is no accessible file version (or standard) of those config settings such that one has to manually re-configure any new hosting environment to match the old one.
A de-facto "file config police" has to be created somehow to create, enforce, and test such standardization, and vendors will resist using FUD.
Table-ized A.I.
Wow, if this did happen, companies would start bringing machines in-house again. The cost of a few hours of bandwidth on the spot market would pay for a rack of servers. Would be a great hedge against this sort of thing, locking in a fixed-cost server rack over a 5-10 year period.
Enron
One if Enron's hare-brained schemes was to develop a market in bandwidth . It was one of the things that steered me away from investing in what was then the hottest stock in the market
I worked for The Weather Channel for several years w/access to core forecasting systems. they were always talking about finding ways to diversify their revenue from just ads. when I heard about this (weather futures) it donned on me that this is one data stream to which I had access w/lower latency than Goldman and in theory should be able to insert some code in some key places to front ru... er, I meant "offer premium weather quote service for" the market. I knew if I were ever caught I'd be fired on the spot - not for committing a crime but for pocketing the proceeds instead of them going to executive bonuses...
now that they're owned by Bain & Blackstone I wouldn't be the least bit surprised if they haven't already implemented this...
You can buy shares in Amazon, Rackspace, etc. So if you want to profit based on value of cloud resources just buy or sell shares in the companies which provide the resource.
Computing resource isn't like money.
You can't use it when ever you please. If everyone wanted to use all their "compute blocks" at the same time, a provider could grind to a halt. They have finite resources. If they never allocated more than they could provide at once, they would never be fully utilised since the "cloud traders" would be holding on to some to trade.
Super retarded idea!
Actually, this would be more like the futures market. Speculators can speculate, but infrastructure providers can sell their capacity out a few years and have a more reliable ROI.
Say, you have space for 10 machines. So let's say that is 500 "standard VMs" (eg. 1 CPU core, 1G, 10G disk) worth is selling at $5/VM/mo up to 3 years in advance. Assuming price is stable, that means the provider can sell access to the machines for next 3 years for $30,000. Sells the futures, boom, cash in the bank. Futures market allows guaranteed sale for providers and guaranteed access for users. It allows for efficient market.
Now who uses that for how much, who cares? The provider just needs to keep the machines up and running for 3 years. This may include keeping an extra two servers for hot standby that could be sold on spot market if they are idle.
Standardized market for VMs like this would be great. The only caveat is there would need to be a definition of standard VM. Without such definition, there could not be an efficient market.
Another note is some big players would definitely not want an efficient market in VMs. Efficient market would allow smaller players to compete with more recognizable brands - market like this *kills* brands. For example, do you care which company or farmer you buy your wheat from? Your oil? Your RAM? It''s all the same.
Am looking for funding for the next Enron...
What do you mean "of you don't want to participate in culture in it's exact form as it exists right now"?
Right now, you can buy from a boutique retailer who buys from a distributor, you can buy direct from the manufacturer, or many choices in between.
I bought my last pair of glasses from 39dollarglasses.com. They are the same glasses the retailer in the mall will sell me for $160. The differences include - the retailer will measure the distance between my eyes for me, help me find a pair that looks nice, adjust them for me, and charge more. Both choices are "culture as it exists right now". Right now you can buy direct from the manufacturer who is 1,000 miles away, buy from a discount store, or a boutique shop. You can have it any way you want. Why do you insist that I also have to have it your way, that I'm not allowed to getvalue added by a dealer? What posesses you to need to take away the last bit if freedom I have left?
So, you have a FHT server farm, right in the exchange, because if you are across the street someone will take your lunch money (well technically your chance to steal someone else's lunch money). But because of the antiquated nature of the markets, they close down for two thirds of the day... Can't use those same servers to game the Nikkei - you have to be right on top of that too. So now you have a world wide set of servers which are 60% idle... Maybe you could sell that time.
But you're someone who thinks FHT is a good idea, so you posit selling your spare cycles like a stock.
I think I know where this idea came from...
-Jeremy
One of the many delusions that Enron fostered and fed upon was that the 'market' was the universal solution to every problem. Sorry to say, but this is just another attempt to monetize something and then institute a futures market where the insiders can game the system and rob the rest of us. We need to step back and ask, what is the purpose of markets? If it is to perform price discovery in an open manner across a large pool of interchangeable suppliers? Or a means for the few to manipulate availability to increase their profits? There is too much of this already, with very few benefits to the average consumer. Shoot them now...
Anyone owning a computer with enough ram can download linux and create their on cloud full of VMmachines. A commodity by definition implies rarity and value The ability of anyone - with the right skill set - to do this 'decommodifies' the end product. . Isn't that were most applications start? Some programmer is fed up with some companies extortion prices or it just misses the mark functionally, so it's developed in house. Every application starts as the need to 'scratch an itch'.
Two questions: Are we going to wind up developing the equivalent of a "USDA Certified Grade A Cycles" sticker? And what is the cloud computing equivalent of Taco Bell "meat"?
I feel like commoditization might provide a level of anonymity to allow both a low grade of service (faking processing with either less accurate processing or known-faulty equipment) and a security risk (While a collection of cloud services are mining your customer data for you, how many are copying it off for later perusal?)
"Because Science" is one step from "Because old book". Try "Because of my experiment testing my falsifiable assertion".
Can't wait for his fad to pass. Hosting all of your financial and customer data and everything else that companies typically care about keeping private in The Clown is arguably a little better than sending it out for processing and storage to some nameless third party in another country, but not a whole lot better.
Vonage tried this and failed - they instead became a VoIP provider instead of a market maker for phone minutes as a commodity.
http://en.wikipedia.org/wiki/Vonage
If futures didn't have value, people wouldn't buy them. How do you think this will work? A bunch of people come knocking on your door saying buy our futures or else?
My retailer IS very interested in my 50Mbps account and they work hard to keep me happy.
I believed everything you said up until that point.
"First they came for the slanderers and i said nothing."
What I think might happen is that regulations might be put in place to prevent cloud services providers from selling directly to their customers, as is currently the case with car dealerships. I believe that would be an unlikely scenario, so the more probably outcome is for the people who set up the exchanges to quickly go broke.
Is 1563649 a prime number?
What if I buy some server time from the market, and the server crashes: who is liable? What is wrong with the current system that putting up an exchange will fix?
If the market works like any other capitalist market, risk of server crashes would be an exchange you get for server time at discount. If, for instance, 100 hrs of server time costs you $100 at par, then server time trading at $.5 per hour ($50 for 100hrs) carries some risk of downtime. Again, if the market works correctly prices should regulate themselves; Server time will only be bought and sold at what the market can bear - a price based on the risk and return of the investment.
This is (a vast simplification of) how the bond market works. Bonds with higher rates of return often carry higher risk in not getting your initial investment back than those trading with lower rates of return (all other things being equal).
As for why this market would exist, I can only speculate that the demand for server time and the supply of servers will so vastly increase that hosting companies (like Amazon) will only find it profitable to market large chunks of time to brokers, who in turn sell smaller pieces to people like you and me who need the server time/space.
Such regulations aren't an underlying problem of the futures market itself. Regulations like that always make markets less efficient. But those aren't at issue here, since nobody has proposed creating such, nor submitted a bill to that effect - it's pure speculation on your part. As for your assessment of the likelihood of success of such a venture, who is it that you think will become bankrupt? The future writer (likely a broker) who charges his client - the suplier? The exchange, who charge everyone? The consumer, who could just as easily buy directly? Again, futures have a benefit to supplier and consumer both. It's possible that no one will see any reason to use these contracts. But many people think otherwise, and you're hardly an expert, are you?
Many people think otherwise? One dude with a blog thinks otherwise.
What is the benefit to consumers of this?
Is 1563649 a prime number?
They can buy when prices are low. They can lock in their cost early and set their own prices accordingly. They can plan ahead and ensure they have enough supply for future needs. These ultimately come down to reducing risk. Generally speaking, reducing risk costs, and taking on risk makes money (at least it has positive expectation).
AWS already has a marketplace where you can purchase free capacity at rates much lower than what you'd normally pay. Smart devs are making use of these "spot instances" to save money on their compute costs. Before we can have a more widespread market, a cloud standard needs to emerge. If I can get cloud resources cheaper in Rackspace today than I can get them from Amazon that largely does me no good because the processes associated with working in those two different clouds are not the same. If I could take the same toolset I use to run AWS apps and forklift those to Rackspace, THAT would be interesting in that it would really start to focus competition on price.