The Future of Data Storage: Cloud?

Data storage, as we know it, is primarily in the form of tapes, flash drives, hard disk drives and solid state storage. With data storage being the fastest growing semiconductor technology, the world is doubling storage capacity every 12-18 months [1]. Data storage firms are increasingly trying to push the boundaries of physical storage, with Seagate announcing its 60TB SSD in August 2016 – the world’s largest SSD [2]. However, with improvements in internet bandwidth and the ever-decreasing costs of storage, it has become more economical for firms and individuals to store their data in a centralised location: the cloud. By outsourcing data to data storage firms, it removes the need to maintain and replace hardware, whilst also making it accessible from anywhere and everywhere. By 2020, it is estimated that one-third of all data will pass through the cloud, and businesses will spend $191 billion on cloud services (up from $72 billion in 2014) [3].

 

What Exactly Is Cloud Computing?

 

To summarise the key differences, most people are aware of local storage. This is when the data is physically close to you, making it fast and easy to access from your hard drive or local network. Cloud computing differentiates itself by having the data accessible over the internet, or at least synced with other information from the web [4]. Many of us are actually using cloud computing without even realising. A lot of services we regularly use are likely to be powered by cloud computing. For example: sending emails, watching Netflix, listening to music through Spotify/Apple Music, playing online games and much more. In fact, due to this unfamiliarity with cloud computing, as President of UCL FinTech Society (Europe’s largest Financial Technology society), we hosted an event at Google’s London Headquarters – to raise awareness of Google Cloud Next ’17 among UCL students.

 

Cloud computing is a big shift from the traditional norm, but this transformation also comes with many benefits. By simple economies of scale, the capital expense of buying hardware and software is effectively negligible. Organisations no longer need to run their own data centres, with racks of servers and cooling. Using the cloud is on demand, and therefore, businesses have flexibility in choosing how much capacity they need at any given point – advanced planning is no longer needed. The cloud also boosts productivity, by providing the right amount of power/storage/bandwidth, exactly when and where it is needed [5].

 

Different Types of Cloud Services

 

There are 3 main types of cloud services. The foundation of cloud is Infrastructure as a Service (IaaS), which is renting IT infrastructure on a pay-as-you-go basis. In other words, this is a means for selling cloud computing capacity, and is also the most flexible computing model. Platform as a Service (PaaS) is one step above, and provides an on-demand environment for developing, testing and delivering mobile/web applications – without the need to manage the infrastructure behind the scenes. Some examples of this are Google App Engine, Windows Azure, Cisco WebEx Connect and Red Hat’s OpenShift. Finally, Software as a Service (SaaS) is the largest player in the cloud market and is often available on a subscription model. SaaS uses the web to deliver applications managed by a third-party – this removes the need for technical knowledge and is accessible to the mass market (e.g. Dropbox) [6].

 

Cloud computing can also be segregated by how it is deployed, and there are 3 ways of doing so. A public cloud is owned and operated by a third-party, who often provides its service using a web browser. A private cloud is internal and exclusively used by a particular organisation. Even in this case, the organisation can pay a third-party to host their cloud, but the services and infrastructure are maintained on a private network (thus adding an additional layer of security). The third method is a fusion of the aforementioned 2, resulting in a hybrid cloud, which allows data/applications to be shared between the public and private realm – giving businesses more flexibility/options [5].

 

Real Life Example: State of California

 

Deputy Director & Deputy State CIO, Mr. Chris Cruz talked about the State of California’s Open Data Portal (https://data.ca.gov), which is a new technological initiative to unlock government data and have public datasets. From the current 325 datasets available to the public, they aim to make 425 datasets available to Californians – who will hopefully assist in the development of California.

 

In regards to data storage, the State of California currently has over 133 data centres. However, they have already started the shift to cloud computing, and plan to centralise everything into 3 main data centres – thus reducing costs (from the current $3-5bn per annum), improving cybersecurity and unifying the various departments. Mr. Cruz also mentioned that the State of California is releasing its State Innovation Lab (http://innovate.ca.gov/), which will allow the development and delivery of more innovative government services (CalCloud).

 

Looking Even Further Ahead

 

IBM researchers have managed to store 1 bit of data on 1 atom, by using the magnetic pole orientation to define the bit values of 1 and 0. A storage device built with this method would be 10,000 times denser than the hard drive disks and SSD’s available today. Although, the data only remains in the atom for a few hours. This method is seen as the ultimate storage feat, since it does not really get any smaller than a single atom [9].

 

To counteract this longevity problem, DNA can also be used to store data, since it does not degrade over time, it is very compact and remains future proof (since DNA has been around for approximately 3 billion years, and will not be going away anytime soon). It is estimated that only 4 grams of DNA are needed to hold one year’s worth of information produced by mankind. This method has already been tested by Harvard researchers in 2012, but is too expensive to become mainstream. According to one study, it currently costs $1300 to encode 83kb of data [10]. Therefore, it is safe to say that cloud computing will dominate the storage industry for the coming years, but new methods are being tested everyday.

 

Cloud computing is a very interesting topic, and only time will tell what the future of data storage will transform into. Many research studies have predicted that there could be a point in time when even the cloud will not be as important, since all smart devices will be interconnected and regularly pinging data between one another. Thus resulting in a global network of devices that continuously transfer data, without the need for a centralised hub.

 

References

 

[1] http://www.computerworld.com/article/2473980/data-storage-solutions/data-storage-solutions-143723-storage-now-and-then.html

[2] https://arstechnica.co.uk/gadgets/2016/08/seagate-unveils-60tb-ssd-the-worlds-largest-hard-drive/

[3] https://www.backblaze.com/blog/the-future-of-cloud-backup/

[4] http://uk.pcmag.com/networking-communications-software-products/16824/feature/what-is-cloud-computing

[5] https://azure.microsoft.com/en-gb/overview/what-is-cloud-computing/

[6] http://www.activemediaacademy.com/map-of-the-clouds/

[7] http://www.infostor.com/backup-and_recovery/cloud-storage/the-future-of-cloud-storage.html

[8] https://www-03.ibm.com/press/us/en/pressrelease/44436.wss

[9] https://www.theregister.co.uk/2017/03/09/ibm_stores_1_bit_on_one_atom/

[10] https://www.backblaze.com/blog/data-storage-technologies-of-the-future/