The following originally appeared in GigaOm:
For better or worse, cloud computing has rushed toward centralization, leaving us with few options for where our data can live. It’s time to consider more private alternatives to the centralized cloud. BitTorrent’s Erik Pounds explains why.
Few trends have gained as much mind share and momentum as the rush to move all our data and computing to the cloud. The fever for the cloud runs high — and it runs largely unchecked. Some, including AppZero CEO Greg O’Connor, believe that the move to the cloud is only going to accelerate. In this vision of the future, all information will be stored in centralized services and your computing devices (laptop, phone, tablet, etc.) will all access storage in the cloud.
In the more than 15 years I’ve spent in the data protection, storage and networking spaces, I’ve witnessed firsthand the benefits many companies — and users — have reaped from this trend. In the eyes of the consumer, the internet has morphed into a group of centralized services that many of us use on a daily basis. Services such as Google Docs, iCloud, Facebook, Twitter and LinkedIn have all made the internet more usable for more than a billion people.
This all sounds great, but it bothers me on a few levels:
1. The more information becomes centralized, the less powerful we are as individuals. I’m not saying we’re on our way to being controlled by The Matrix, but we’re moving a tick in that direction. As Jon Evans argued in his article, “The Internet, We’re Doing it Wrong,” the internet is becoming increasingly consolidated into a few centralized stacks — easy to use, but ultimately putting too much control into the hands of a precious few at the expense of choice and privacy and creating a great temptation for hackers.
2. When data has to flow through a central arbiter, or must first be staged on another device (i.e. cloud-based services), it makes the internet inherently less efficient than if that data moved directly from end device to end device. This inefficiency is compounded by the massive buildings and ecologically unfriendly infrastructures that have to be created to run the cloud. And I won’t even get into the temptation for abuse created by centralizing all of our information into a few walled gardens.
The Internet was designed and built to connect end devices over large distances, allowing them to communicate. The idea was to create a communication system without a single point of failure. It established a new and sustainable way for information to be exchanged and this was great for the sciences and highly technical users (universities were some of the first to leverage this system to share knowledge), but it was accessible to very few.
These centralized services have made the internet more accessible and brought a lot of value to our daily lives, but we’re at a point where we need to evaluate which services should be centralized and which should not.
One way to do this is to offer private and more efficient alternatives. Cloud-based services undoubtedly bring a level of convenience, but if you want your files to remain private, move faster between devices and not be limited by capacity, a device-to-device solution is a better option. This would also satisfy IT managers who need to efficiently move files between two servers — whether locally or over long distances — and do not want to have high software expenses, deal with VPNs or buy expensive WAN acceleration boxes.
Even some cloud companies have good examples of technologies that allow end devices to communicate. Apple’s FaceTime first uses the cloud to find another device. But once that connection is established, audio and video data travels directly from one device to another without going through the cloud. Skype used to work this way as well.
Open Garden is another good example of the benefits of bypassing the cloud. The app enables direct device-to-device Wi-Fi, making for a faster and more efficient mobile internet.
The cloud is not just a means to hold our data or for end user services; it also runs many business applications around the world. Often, IT organizations running large corporate-owned data centers write applications to run in public cloud environments, such as Amazon Web Services, because they can not evolve quickly enough to provision internal resources. This is a great place to incubate an application, but once it gains significant traction, it may be more cost effective and secure to run the application in a private data center. For these scenarios, it’s great to see the OpenStack community and such companies as Cloudscaling, Eucalyptus and Piston Cloud develop technology so that modern apps can run in both public and private environments.
For some IT departments, regulations and compliance requirements create roadblocks for a cloud migration. The regulatory environment is not without just cause given the sensitive data in industries such as finance, legal and healthcare.
As end users and businesses, we need to have options for where our data can live. The current trend has been a rush toward centralization — far removed from the original vision of the internet. That’s had its benefits and risks. It’s now time to identify opportunities and create tools that offer more private alternatives to the centralized cloud. The public cloud will always play a role, but we need additional capabilities, while being able to maintain control over the stuff we value.