Cooking With Technology

Microsoft’s Pay-as-you-Go CloudOS

Microsoft is quite  apparently determined to not miss the possible shift to cloud computing. On top of October’s Microsoft Developer Conference announcement that the company will be launching “Azure” (Microsoft’s answer to Amazon Web Services), they also announced a next-generation operating system that will be cloud-centric. Overall, it seems like a logical progression for a company to make, and Microsoft really seems to have a handle on this emerging technology. However, the third development of Microsoft’s investment in cloud technologies is something consumers may want to be wary of.

It has surfaced through CNET (and echoed on CNN and other mainstream outlets) that Microsoft has applied for a patent relating to metered computing. The essence of the patent is a “pay-as-you-go” model for computing, where the amount and intensity with which you use a computer directly relates to how much you pay to use your computer with Microsoft’s operating system. They also discuss the notion of selling packages of functionality to the user – one performance level for gaming, one for office productivity, another for web browsing, and so on. Microsoft openly admits that users may end up paying more to use their computers than under the existing model, where consumers purchase the operating system at a fixed one-time cost. What are the advantages to Microsoft’s users, then?

Summary: none, for the average user.

It seems as though segregating functionality and services that have been historically monolithic is becoming a popular trend for businesses (telcos fighting to dictate which web sites you can visit and violating net neutrality is another example). In the past, using a computer has been a process of 1) buying a computer, 2) buying an operating system to put on it (or more commonly having one come with the computer), and 3) supplying the power to run the computer. What is done with it once installed is up to the consumer, and they may use it as long as they want for whatever purpose without incurring more charges. Microsoft’s speculative model is much like how you pay for power – you pay for what you use. The difference here is that providing power to you costs the utility company money in terms of infrastructure and fuel (or in the case of Washington state residents, dam maintenance), but once Microsoft has sold a product, the costs they incur by the consumer’s use of their own computer is nothing.

If Microsoft is planning on using this model for a cloud-centric operating system based on Azure, then the model makes more sense, as Microsoft will have legitimate infrastructure costs whenever processing power or disk space is allocated to someone. However, modern computers are becoming increasingly more powerful and decreasingly expensive as time goes, and I doubt users like my parents have ever had a need for more processing power than what is afforded to them locally.

The devil then rests in the details:

Is Microsoft going to apply the “pay-as-you-go” model to users utilizing only their own local resources, or only to people using a cloud-centric OS? The former would be very bad for the consumer, the latter would be understandable. If only a cloud-centric operating system is released by Microsoft in the future, are they going to ship it with consumer protections enabled by default (ie, default to not use cloud resources)? This would be okay, as extended functionality that would incur more costs would be an opt-in consumer decision. And finally, is Microsoft really going to separate Azure utilization into packages based on processor usage? It is this author’s opinion that if you are using cloud resources, you should just use cloud resources – price-packaging processor or disk space is ridiculous. A clock cycle is a clock cycle, regardless of whether processing an instruction for Firefox or Far Cry 2.

I sincerely hope Microsoft executes whatever plans they have in a manner that increases consumer confidence in the cloud. They have an opportunity to turn “cloud computing” from a buzzword into something average consumers utilize, and faltering here could be a setback for the software industry as a whole.

Don’t mess it up, Microsoft.

Storing Data in the Cloud

I want to open this up with a question posed by j9er on twitter:

j9er Taking cloud computing to its ultimate outcome, does ayone care that, say, Microsoft would have all your data? Or, Google?

My concise response: yes, I care. But the three primary reasons go beyond the 140 characters twitter would allot me.

Unrestricted Access

The big thing that keeps me from being completely comfortable with a third party storing my data for me is whether I would have unrestricted access to my data. Currently, users are even having problems with google disabling access to its user’s gmail accounts. So say Google or Microsoft have all my data, not just my e-mail and some google docs: how are they going to guarantee that I can access my data how I want, when I want, and wherever I want?

Security

Beyond that, security is an issue. On my home system or my personal servers, I can guarantee their security, as I am the sole administrator of my nodes and I am security conscious. Additionally, my personal data is not coveted by malicious hackers. The amount of effort it would take for someone to gain access to my machines externally is definitely not worth the information they would gain access to. However, when the information of millions of users is aggregated into one location, that data becomes extremely attractive. So the second question becomes: how can the company guarantee the security of the cloud, and the security of the information within? (On that note, I highly reccommend the cloudsecurity blog, where they attempt to answer these questions.)

Privacy

The last concern is one I am sure most people familiar with the topic have thought about. Storing information with an organization involves a great amount of trust, and when it is done improperly people raise their voices. (A good example is any time the CIA/NSA/FBI want to merge databases. If you do it without consent, it is an improper collection of information.) Storing information in the cloud therefore needs to be consensual, guarded well, and exposed only to the owner and applications owned by the organization, but never reviewed by humans, and with all personally identifying information scrubbed.  The final question therefore is: how can the organization storing my data guarantee my personal privacy?

I care about my data. Especially who has it, who may have access to it, and how it may be used to identify me. Until the questions can be answered and those guarantees made, I won’t be comfortable storing all my data in the cloud.