Posts Tagged ‘cloud’
Cloud computing is all the rage these days. It is, in short, computing as a service instead of as a product. This has serious advantages but also has a negative side that may not immediately be obvious.
The concept is pretty simple; somebody keeps some physical computers up and running and provides computing time on those computers to others, rather than everybody buying their own hardware. Sometimes this service is provided by a different part of the same company or government (a private cloud) and sometimes it is sold as a service (a public cloud, like Amazon EC2).
From the perspective of deployment and resource management, the cloud is a great idea. It’s easy to go from nowhere to having a couple of hundred machines running. No need to buy more hardware, fiddle with install CD’s, or worry about replacing broken parts. Everything can be done without having to leave your desk. And it’s possible to scale an application as demand requires; adding a few more servers can be done in less than a minute, and getting rid of them is probably just as fast. It’s easier to get redundancy. Quick repurposing of machines when demand requires it. Anyway, great stuff: the engineer in me is thrilled.
But it’s not all rosy. Cloud computing is just a technology, and technology isn’t inherently good or bad – it’s how you use it. There are plenty of valid use cases for the cloud, but it also enables some worrying behavior.
For starters, it makes it possible for companies and governments to outsource the physical machines and lower software layers they run their applications on. This means that (private) user data gets shipped around to not just one external party – the one running the application – but also to the cloud provider or providers that are providing the hardware it runs on. Their security or downtime can affect a large number of customers (and their customers). The cloud providers may be in a different jurisdiction (or often, multiple jurisdictions) with different legal systems, opening their customers up to legal issues regarding data retention or foreign government surveillance. And while having data spread across different companies and different continents was not impossible before the cloud, it certainly has become a lot easier and more common.
Handing over control over my personal data to so many companies (and governments) makes me uncomfortable, but it’s so darn convenient. Is there no way we can have the best of both worlds, some sort of hybrid model?
Excellent interview with Eben Moglen about the problems of centralization on the internet – probably the best explanation I’ve seen – and various other topics, including WikiLeaks and copyright reform.
The age we live in is the first with an overabundance of data. It’s had a tremendous influence on our daily lives.
Cheap electronics and digital processing have made it possible to gather and organize all data imaginable. All online stores, and most regular ones have easy access to their transactions in digital form. Video cameras keep an eye on dangerous corners and other interesting or boring views. The integrated GPS, GSM and Wi-Fi in phones is used to pinpoint the location of its user. (Radio)telescopes capture vast amounts of images. And, of course, web browsers and web servers are aware of the web pages we visit.
In addition, the extremely low cost of digital storage, compared to the technology available to previous generations, has made it possible to record this information and keep most of it around for a long time. At least for as long as it could possibly be useful.
Gathering digital information usually is not hard, nor expensive. There is often barely any effort involved at all, because the data is already generated or digitized as part of some other process – it just needs to be stored. Storing large amounts of data is not trivial, but also not unreasonably hard. This is also one of the reasons why most internet giants are reluctant to throw away data; it can come in handy later, but once it’s gone it’s completely gone or at least hard to regain.
The real challenge is in organizing and accessing the information in a useful and simple manner. This is a hard problem. Not only does the information need to be available, it needs to be available in a structured manner. Having millions of petabytes of raw data at your disposal isn’t very useful without knowing where to look or how to interpret what you find. Likewise, it isn’t possible to iterate over all available data for every single query. If you’ve used the web in its early days you will remember that search engines only provided a very simple subtext match over all pages everywhere and returned results in no particular (useful) order. They were some help in navigating your way through the web, but there was still a skill set (and a lot of patience) required to actually finding what you were looking for. This challenge – making all information everywhere available in a usable manner, and fast – seems to be what Google is all about.
Having all of this data available, and the ability to access it in a meaningful way, is absolutely awesome. It’s an information engineers’ wet dream. It helps eliminate information asymmetry. It’s what allows my phone to know where I am without having GPS enabled (it uses the location of the nearby WiFi access points instead). It’s what makes it possible for me to find the full details and contents of any book in mere seconds. It’s what makes it possible for last.fm to predict the concerts I’m interested in in my direct vicinity.
It’s also pretty fucking scary, considering the effect omni-accessible information has on our privacy.
Perhaps the current use of the centrally available data by the state is reasonable, but there is little or no monitoring of how this data is used. The mere availability enables easy widespread abuse by any (future) totalitarian state.
There are some international treaties that are supposed to protect citizens with cloud data that is physically hosted elsewhere. They’re just show, apparently. I was paranoid about storing my non-public data in the cloud before now, and this doesn’t really help.
Oh, and while I’m ranting. The Telegraaf is well known for their shitty journalism, but this article is crap, even by their standards.