For the most part the world has survived the financial crisis without falling back into the Dark Age – at least for now anyways. So today I am not going to talk about banks, as the title of this blog may have led you to believe, but rather another industry we are creating that will be too big to fail.
I am not sure the financial experts out there have fully analyzed what we as a world did to ourselves during the recent bank bailouts and financial panic, but governments everywhere have reacted by putting in additional safeguards that attempt to address the issue of the financial crisis. However, one could argue that the underlying issue wasn’t addressed at all. The best approach would have been the implementation of targeted strategies for each big bank in question to avoid a world economic disaster. Instead, most of the government effort was spent on reducing the probability of banks failing. There is a big difference!
At some point the technical world realized that reducing the chance of any one piece of technology failing has its limits. These limits include increased cost, increased time to development, slowdowns in progress, etc. To avoid this list of consequences, the industry shifted their focus to building systems that are unlikely to fail, even those that are composed of less trustworthy parts. For example, it would be very expensive to create an uncrashable hard drive – which we found out after many fruitless attempts at doing so. Now most mission-critical systems have multiple hard drives with data duplicated across the drives. If any one drive breaks, the data is not compromised, and the broken link is easily (and inexpensively) replaced. The realization was that the effort and resources to build an invincible hard drive vastly outweighed the costs of simply having more redundancy.
So How is the Tech Industry Becoming Too Big to Fail?
Increased labor costs, scarcity of resources, and exponential complexity are the main reasons why organizations everywhere are moving their internal systems to the cloud. The costs are so low that now a technical leader will need to defend keeping or purchasing new equipment that will be located on site (not in the cloud).
To be clear, when I speak of the cloud I am mostly referring to today’s giants in cloud service companies: Microsoft, Google and Amazon. Other less important but noteworthy players are Salesforce, Rackspace, various VOIP providers and online financial service providers.
Moving infrastructure to the cloud initially including moving your company’s machines. Now Amazon, Microsoft and Google are creating distinct value through custom features that essentially lock applications into a single provider as a means to further strengthen their lock on your systems and data. Why wouldn’t they?
So what happens if Google ever goes down? The short answer is, they won’t. The world economy couldn’t let it happen because of Google’s unrelenting monopoly of the tech industry and influence on world markets.
Sounds a lot like the too big to fail rhetoric of around 2006, doesn’t it?
The Terrible Fat Finger
The thing is, Google doesn’t even have to fail in the grandest sense to cause a major issue. If one third of all the world’s systems are hosted through Google all it would take is a seemingly small oversight from a Google engineer to take out a millions of businesses. Maybe the system that is running an ambulance GPS, your doctor’s medical system, traffic lights… Who knows? There have already been some large profile outages in the last year, namely the Azure leap year crisis and a CloudFlare router issue.
And if you search online you will find many, many more.
To make matters worse, unlike pure commodities like those that exist in the financial space, our government officials don’t really understand tech. This makes new legislation generally ineffective when it comes to the tech industry. Think about it: what can you really do? The rate that companies move to the cloud will accelerate as the services get better and price continues to fall. There are no easy answers. A couple ideas come to mind but none sound great;
Get Redundant: Strongly encourage companies that have over a certain amount of users to redundantly store their data. Requirements could be attached to current business continuity laws such as Sarbanes-Oxley Act. Businesses hate this law so I am not sure attaching anything to such a despised law is a good idea.
Technical Insurance: Create an insurance program like the FDIC that collects a small fee from every transaction which in turn funds people and infrastructure that could be used in the cases of emergencies. For example, telephone companies have mobile versions of their cell towers they can use in case of emergencies. Not sure exactly how a system like this would work for a data center, but something is better than nothing.
Pay Up: Same as idea #2, but instead fund the employee’s salary for 90 days while customers move their data and apps out of the datacenter. No doubt there would be some crazy, tired, developers but they would figure something out.
What are your ideas?