Clown Computing–Entertaining and Attention Grabbing but a Flawed Thesis

Recently I attended Webstock, a conference in Wellington, New Zealand that is well known for bringing together technology, design and general brain stimulation. It was a great event, with some awesome speakers but one presentation, by Jason Scott, part of an activist preservationist group Archive Team, kind of stuck in my craw.

In his presentation, Scott held up some well known examples of discontinues service and drew the conclusion that these examples showed how risky cloud computing is. He went on to rename cloud computing as clown computing and proceeded to get attendees whipped up in a frenzy of derision about this awful thing called the cloud. He used the example of GeoCities, the once great website creation and hosting service that was acquired by Yahoo! and subsequently shuttered. The Internet Archive, the organization that Scott works for, announced a project to archive GeoCities pages and started to download as much content hosted on GeoCities as they could before it was shut down.

Now Scott is undoubtedly entertaining (hell, the guy gave his presentation dressed in a steampunk costume) but conflating the GeoCities shutdown with the proclamation that cloud computing is fundamentally flawed as a concept is, well, fundamentally flawed.

Apart from anything else, the Archive he’s built, while a worthy initiative, tells people that essentially you can’t trust a third party to store your data, only to create an archive that is stored…. with a third party. That’s kind of nonsensical. I questioned this approach and, jumping on the populist bandwagon, received a bewildering reply that seemed to take the view that storing files on one’s own server is far safer than to store with some evil voodoo-doll corporate.

lw

Putting aside for a moment the fact that the reply originated from someone who is an investor in a number of companies that provide a cloud service and, as such, would normally not be expected to  decry storing data on “some corporate’s cloud”, the very suggestion that cloud storage is inherently dangerous is flawed. Let’s look at it dispassionately for a minute here. As most readers know, I live near Christchurch, New Zealand, a city that was heavily impacted by earthquakes over the past couple of years. When the most damaging earthquake struck on February 22, 2011, there were many cases of people going into dangerous and partially collapsed buildings in order to extract their servers – precisely because they relied on traditional local storage for their critical files. there were also, sadly, many cases of people unable to reclaim their servers, and many of these people’s critical data was lost forever.

Compare this to the number of forward-thinking individuals who had their data sitting on one or another cloud vendor and who simply logged onto these cloud services off another computer and were back up and running rapidly.

Of course Scott’s answer to this would be that there are many individuals who relied on cloud services, only to have the service provider go out of business and close the service down. Well there are a couple of responses to this:

  1. It is virtually unknown (virtually, not completely) for a vendor to close down without giving users significant warning and opportunity to extract their data. Think of it as a hard drive that warns you when it’s about to die and gives you the opportunity to back it up before it does so
  2. In this modern age, there is no reason to have your data in a single location. Case in point: my images. I have thousands of pictures dating back many years – these live in a myriad of locations – physically on a storage device in my office, but also replicated on cloud file sharing services, synchronized to a couple of different laptops and, in many cases, on one of a variety of social networks

Rather than looking at this emotionally and increasing the misguided perceptions that suggest not being able to “see” one’s data storage device is inherently risky, it would be better if people on both sides of this debate added helpful dialog to the conversation, and gave consumers the information that allowed them to make rational decision. Cloud storage isn’t fundamentally dangerous, and on-premise storage clearly has its place too – it’s by finding the best solution to an individual’s needs and situation that we can best help them make the right decision. And as for that suggestion that cloud storage is a data-loss disaster waiting to happen? Here’s a non-emotional bottom line answer to that suggestion:

sh

 

5 Comments
  • As per the speech, the issue I have with “cloud computing” is not the technical aspects regarding redundancy and dependency from said redundancy. It was that the watchword for a lot of cloud computing is obfuscation, that is, removing end users from any awareness of their data’s current state, physical location, and so on. Over time, some have gotten better, but as my talk showed (in more examples that just Geocities), the lack of respect and validity regarding users’ control and knowledge of their data means that it can easily turn into a very dark cave indeed.

    • I disagree – the vast vast vast majority of cloud vendors are up front about where data is stored (“Hi, We’re Xero and our data is in Rackspace’s Dallas DC” for example) – it seems to me that your obfuscation angle lays into the hands of those who have a fairly archaic notion that if you can’t touch it, it must be dangerous. That (in my humble view anyway) kind of harks back to the dark ages and fear of anything unknown. With the utmost of respect, I still believe that, like for like, cloud storage is safer than traditional….

    • Reminds me of a oft used and paraphrased joke.

      Four years ago didn’t know how to spell unemployed, now I are one.

      Citing GeoCities as an example of a cloud flaw is like saying Larry Ellison was right when he said, “Maybe I’m an idiot, but I have no idea what anyone is talking about. What is it? It’s complete gibberish. It’s insane. When is this idiocy going to stop?” Now he is one.

  • Actually, I think the cloud is all about layering and abstraction not obfuscation. That said, object storage in the cloud is implemented on bare metal storage servers organized into clusters and zones of availability. Object storage providers offer enhanced data durability by using either replication or erasure coding or a combination of both. Object storage in the cloud is many times more durable than premises storage using file and block storage devices with RAID.

  • Seems to me that this discussion is conflating cloud storage/backup and cloud services.

    I don’t think anyone (Jason Scott included) would disagree that there are many benefits to cloud storage/backup. Anyone responsible for the safe keeping of data knows that one copy is never safe, local or remote. To say you have redundant storage/backups you need atleast a local copy and a remote copy (in the cloud or just ‘off site’). Experts always say you need atleast 3 copies to be truly redundant.

    I *think* Jason’s beef isn’t with the cloud, it’s with cloud (web) service providers who operate with an inferred promise of longevity, but then can disappear at a moments notice and take with them lots of data that users have erroneously thought would be safe and secure. I tend to agree that this sucks, and I am unlikely to trust any service that doesn’t give me a way to easily grab a copy of my data. Of course part of the problem is uneducated users, but where you have mums and dads who just don’t understand what it means to store data in a cloud service I think it is incumbent on the service provider to make it clear and give people an option to keep their data if they want to.

    I use many cloud services (Google, Twitter, Xero, AWS off the top of my head) but in no case do I trust them as the one and only place to store my data.

Leave a Reply