In what WIRED magazine called an "epic exodus from Amazon's cloud empire," posted on the anniversary of 10 years of Amazon Web Services itself, we began to see how storage-heavy companies are migrating toward something called the "private cloud." This is a technical term that can range from simply having your own storage, run from your own data-center, or can be a dedicated, complex environment that many businesses will find confusing. For those new to the idea, web hosting can change dramatically depending on what you do; some just need a website, some need a full infrastructure to run data-intensive jobs and others, such as Dropbox, who provide storage to consumers and businesses, require all of the above.
This is why companies like Bitcasa, a formerly unlimited data storage company, had to change their business model, which stores data on Amazon's data cloud. When you're paying by the gigabyte (or terabyte, or petabyte) or data, you're paying an extra expense based on the variables of a customer. This is why Dropbox, to support millions of users, had their own infrastructure built to store their own data, that they'd manage, versus paying Amazon for the ability to store their data. Amazon's multi-billion dollar cloud storage wing is attractive to many businesses, as are competitors like Google Cloud, which recently won music-streaming service Spotify's business. You trade the extra cost for a managed server environment that is pay-as-you-go.
Another storage company, offering both a consumer backup plan and an enterprise-level competitor called B2, is Backblaze. Based in San Mateo, they were founded on the exact style that Dropbox is now adapting to, building their own open-source storage pods, a fancier way of saying a server that has software and hardware purpose-built for cloud storage that, in turn, allows them to offer it at $0.0005 a gigabyte a month, over Amazon S3's $0.022 a gigabyte a month. They report that this is based on building their storage pods for a specific task (using consumer hardware versus expensive "enterprise grade" hardware that does exactly the same job, among other things), profitably, to provide lower-cost storage. Their offering, 77% cheaper than Amazon's, is an option that many cloud storage companies may take up rather than build their own infrastructure; after all, Dropbox is a multi-billion dollar company. While nobody can know exactly the system that AWS was created on, Backblaze's servers for both consumer and enterprise functions were apparently built with the early vision of being cloud-specific. This falls in line with Google's recent statement that they want hard drives built specifically for the cloud, less reliable but larger, taller hard drives that they are calling "data center disks."
Then there is the idea of a hybrid cloud, hosting, as Ash Asutosh writing for Betanews discusses, has also become more common. The internal elements of the business, the core functionalities that cannot go wrong or go down, are handled on-site while the overall business is kept on a "public" cloud, meaning an Amazon Web Services or Backblaze B2 platform. This strikes a balance between cost and efficiency and is becoming a common approach for some businesses. Cleversafe, a 12-year-old company, was acquired by IBM in 2015 for both their mastery of the private cloud and their (as TechCrunch's Ron Miller reports) over 350 patents in cost-efficient and secure data storage. Google also followed suit, making partnerships with several hybrid cloud security startups including Splunk.
It's an infrastructural challenge for startups and more mature businesses too, with many different choices to work with and an intense pressure to work with what everyone else. Huge companies rely on Amazon Web Services, for example, hence why an outage can create such chaos for companies such as Amazon, Netflix and Tinder, but the cost-effective nature of said move mitigates the danger for many. However, as time goes on, businesses will become more cost and stability sensitive, though even a private cloud can fail. The result is that they take the good with the bad, and potentially migrate from Amazon to other services, be they more affordable solutions, their own services, or a combination of the both. Amazon may have a decade behind them, but the next ten years may spell for a very different cloud war than they expected.