The spread of the cloud has been one of the biggest changes in the way that people share information over the past decade. With internet speeds increasing and the cost of storing data online decreasing, it certainly hasn't been a surprise. It has opened the door to huge changes in how people work together and a great deal of the world's most popular enterprise software is now exclusively available through the cloud. For instance, programs like Adobe and Microsoft Office are now almost exclusively accessed through the cloud. Companies like Dropbox and Box have reached billion dollar valuations through giving people the ability to easily share files from wherever they are working in the world.
However, this is only a drop in the ocean as more and more companies are looking at big data as a service (BDaaS) offerings with Forrester forecasting a 39.5% annual growth in the market through to 2021. In fact, according to Forrester, over 50% of enterprises will embrace a public-cloud-first policy for data and analytics in 2018, and by 2021 the BDaaS market is predicted to be worth over $30 billion. But what exactly are the advantages of this over more traditional storage and analytics systems?
For a long time there was a general consensus that holding your data in the cloud was somehow inviting people to steal it. You were essentially just giving hackers free access to your data because you didn't have it locked down in a server behind a closed door in your office. However, common sense has since prevailed and for the vast majority of use cases it has become clear that storing data on the cloud is safer than bringing this storage in-house.
There are several reasons for this, the most simple one being that the entire reputation of the cloud storage or computing company is reliant on security, so rather than it simply being another business function within the company, it is front and centre. Cloud companies therefore spend millions of their security, something that the vast majority of companies simply cannot do.
However, beyond this basic fact, according to John L. Miller who worked on several distributed and cloud storage services and has a PhD in distributed systems, cloud services 'typically [store] at least three copies of each piece of data, all in different places. You need to have all three copies disappear at exactly the same moment (e.g. from three separate hard drive failures) to lose that data. After working on Exabyte-scale storage services, I have only seen this happen with a relatively small (e.g. 10^-9) fraction of data. And even that can usually be recovered, it just takes a couple days.' Where most companies may do a daily backup, cloud based services will constantly backup in multiple locations.
In fact, the only element that is vulnerable on cloud computing is the exact same vulnerability that every single system has, and that's human error. According to the UK's Information Commissioner's Office 62% of all incidents of data loss reported to them was a result of this, compared to just 9% for insecure webpages and hacking combined. The main cause of this is people simply giving away their passwords or choosing easily guessable passwords, making it simple for hackers to access systems. In this case it doesn't matter if data is on a personal computer or in the cloud, people could get access.
The cost of using cloud-based storage or computing solutions is something that is not often discussed, but is considerably cheaper than any kind of on-site system. For instance, if you consider the price of equipment for an on-site server, then the housing of it, then the security of it, suddenly a monthly subscription makes considerably more sense.
In the future this is likely to be even more relevant as more powerful and expensive technologies come to the market, with quantum computing being a prime example of this. Google and IBM, the two companies already leading in the BDaaS industry, are rumored to be creating quantum computers that can be accessed through the cloud. Given that each quantum computer currents costs over $1 million and needs to be housed in a highly controlled environment. This makes it almost impossible for most companies to use it for one off projects where it could make the most difference and also means that the potential will be concentrated in only the most rich and powerful companies, consolidating their power.
Success breeds success, especially when it comes to data, but this often leads to issues with scalability. Traditional on-site data and analytics functions require different levels of hardware dependent on their needs. For instance, as the amount of data that needs to be analyzed increases, the power to analyze it also needs to increase, as the success of data analysis increases, the speed at which it's needed increases, etc.
This leads to the constant need to upgrade equipment, which for on-site equipment can mean completely replacing systems every few months at great expense and also causing significant disruption. At the same time there are periods where this kind of technology is completely necessary because of increased workload, and others where it won't be necessary. With on-site systems companies are forced to use more expensive or complex hardware regardless of whether it is always needed, but with a BDaaS model it is much simpler to scale up or down when required without the huge expense of needing to buy an entirely new system.
One of the best known business trends of the past decade has been mobile working, whether that's working from home a couple of days every week or working from the other side of the world. Initially, this created some big problems in terms of access to technology and data because servers were in office and often the most sensitive data and powerful computing power could only be accessed in-office. However, with BDaaS this removes this issue completely as data and analytics mechanisms are available wherever somebody is in the world.
It is, of course, possible for analysis to be done locally wherever somebody is working as most modern computers have the capability to perform data-heavy tasks with relative ease, but the issue with this comes when there are multiple people working off a single data set, especially when this data can be relatively dynamic or needs to be analyzed in real-time. To get around this issue BDaaS allows companies to both store and analyze data in a single place, accessible and editable from anywhere.