Pennies in the Cloud
By now you have seen the Amazon S3 price reduction. Effective Monday, Amazon has discounted the price per gigabyte for their cloud object store by up to 13%, with the discount inversely proportional to the storage consumed. The official explanation for the discount is that Amazon continues to innovate to drive down costs and “pass along the resultant savings to you.” If you’re a competitor, you’re likely to take away a different conclusion: Amazon’s disruptive pricing isn’t going to stop any time soon.
The Amazon announcement reveals a little known fact in the 2012 battle for the public cloud: competitors that have maintained tight control over the cost effectiveness of their software, infrastructure and operations have an unfair advantage over those that have not. Their technical and process innovations, combined with increasing economies of scale, will continue to allow them to drive the costs of cloud services down while still maintaining healthy profit margins.
Unfortunately for some emerging clouds, in a rush to enter the market they have not maintained strict control over the cost effectiveness of the basic building blocks of their service. One of the most common sins is the licensing of expensive third party software and/or non-commodity hardware for core components, which gets passed on to customers via the cost of goods of the cloud service. While licensing third party software is nothing new in the cloud, what is different is the more common use of consumption-based licensing for lower level services. With consumption-based licensing, the more a customer consumes of the cloud service, the more the cloud provider needs to pay a technology provider. This model is further amplified by the fact that core services – such as compute, storage and basic application services – are the building blocks of higher level cloud services (e.g. Amazon RDS uses EC2 compute and EBS storage to deliver their database service). So as a cloud provider moves up the application stack – and all cloud providers will eventually need to move up the application stack – they carry with them the weight of the margin stacking decisions made for lower level services.
A former colleague liked to say: “the cloud is a game of pennies.” Amazon’s price reduction is a sign that their book selling DNA not only understands they are playing for pennies, but that they have maintained tight control over the cost effectiveness of their technology, process and infrastructure. So as we celebrate the price reduction, competitors take note: it’s not enough to get a feature competitive cloud offer to market – you also need to do so with economics that enable you to compete in the game of pennies.
Great post.
We find that many of the companies who want to use the cloud simply fail to check costs of using the cloud. The first assumption is ‘it will save me money’. This might be true, however the pay-as-you-go pricing model needs to be understood.
- Will it save money over x years.
- When will these savings be made.
- Forecasting costs / savings for a growing business. e.g. Christmas peak sales might mean many more servers are needed from Nov to Jan.
We have developed a tool to look at these questions – I would love to get your thoughts on it: http://www.ShopForCloud.com
Thanks,
Hassan Hosseini
ShopForCloud.com
Interesting questions and post.
I always remember some years ago, a company owed me a lot of money. My finance guy looked at their books and concluded it could be saved as a business. However when I had a turn-around angel looked at their business, he found that every widget them made lost them money. They had a full order book so the conclusion was they would go bust even more quickly if they continued. The only answer was to raise the prices and win profitable business which they did. And yes, I got paid !
Where do you think price reduction leaves vendors such as VMWare, IBM, Oracle and Microsoft. All have solutions generally based on a paid for licences and high end hardware ……..
It does however seem HP, Dell and AT&T are adopting http://www.OpenStack.org which is a free open source cloud code.
It tends to suggest you are right about long term consolidation in the cloud market as vendors run to compete in an economic way with market leaders Amazon and Rackspace.
Your thoughts ???
OpenStack definitely has the potential to have a disruptive price impact, especially if it continues to broaden its scope. But even if it fully commoditizes the software for compute and storage, there is still a lot of room for cloud providers to be highly inefficient, even with open source software for compute and storage.
I’ve been curious to know more about the impact of owning your own hardware. Dell, HP, and IBM all own their own server and storage hardware, which you would think would give them a cost advantage.
I do have a growing suspicion (but not quite an opinion): it’s all about the software.
Nice article.
@Joe – Agreed. It will all driven by software as there are disruptive opensource/proprietary software that will provide a lower cost point with scale. It’s very true that inefficiency of cloud providers will keep the cost up as even with commodity hardware and opensource software, they have to deal with power, cooling and real-estate costs (which will be higher with scale out commodity hardware). So the winner is those who uses the best of class opensource software and has an efficient, green data center.