For most of computingâs history, data meant âstructuredâ data or data that fits neatly into pre-defined categories and rows stored in databases or spreadsheets. But the big data movement has changed all of that with the proliferation of unstructured data analysis. Unstructured data is any data that doesnât fit into a predefined data model. It includes things like video, images, text, and all the data being logged by sensors and the myriad of digital devices. Where structured data is relatively easy to store and analyze using traditional technology, unstructured data isnât.
None-the-less, today, massive collections of unstructured data are being analyzed for altruistic purposes like combating crime and preventing disease, but also for profit motivated goals like spotting business trends. And, as weâve entered an era of pervasive surveillance â including aerial surveillance by drones and low earth orbit satellites capable of delivering 50 cm resolution imagery â media content (photos, videos and audio) are more relevant to big data analytics than ever before.
Unstructured data tends to be vastly larger than structured data, and is mostly responsible for our crossing the threshold from regular old data to âbig data.â That threshold is not defined by a specific number of terabytes or even petabytes, but by what happens when data accumulates to an amount so large that innovative techniques are required to store, analyze and move it. Public cloud computing technology is one of these innovations thatâs being applied to big data analytics because it offers a virtually unlimited elastic supply of compute power, networking and storage with a pay-for-use pricing model (all of which opens up new possibilities for analyzing both unstructured and structured big data).
Before their recent and unfortunate shutdown, the respected tech news and research site GigaOM released a survey on enterprise big data. In it over 90% of participants said they planned to move more than a terabyte of data into the cloud, and 20% planned to move more than 100 TB. Cloud storage is a compelling solution as both an elastic repository for this overflowing data and a location readily accessible to cloud-based analysis.
However, one of the challenges that come with using public cloud computing and cloud storage is getting the data into the cloud in the first place. Moving large files and bulk data sets over the Internet can be very inefficient with traditional protocols like FTP and HTTP (the most common way organizations move large files, and the foundation for most options cloud storage providers offer to get your data to them besides shipping hard drives).
In that same GigaOm survey, 24% expressed concern about whether their available bandwidth can accommodate pushing their large data volumes up to the cloud, and 21% worry that they donât have the expertise to carry out the data migration (read about all the options for moving data to any of the major cloud storage providers, and you too might be intimidated).
While bandwidth and expertise are very legitimate concerns, there are SaaS (Software as a Service) large file transfer solutions that can make optimal use of bandwidth, are very easy to use and integrate with Amazon S3, Microsoft Azure and Google Cloud. In fact, the foundation technology of these solutions was originally built to move very large media files throughout the production, post production and distribution of film and television.
Back in the early 2000âs, when the Media & Entertainment industry began actively transitioning from physical media including tape and hard drives to digital file-based workflows, they had a big data movement problem too. For companies like Disney and the BBC, sending digital media between their internal locations and external editing or broadcast partners was a serious issue. Compared to everything else moving over the Internet, those files were huge. (And broadcast masters are relatively small compared to the 4K raw camera footage being captured today. For example, an hour of raw camera footage often requires a terabyte or more of storage.)
During M&Eâs transition from physical media to file-based media, companies like Signiant started developing new protocols for the fast transfer of large files over public and private IP networks, with the high security that the movie industry requires for their most precious assets. The National Academy of Television Arts and Sciences even recognized Signiantâs pioneering role with a Technology and Engineering Emmy award in 2014.
Today, that technology has evolved in step with the cloud revolution, and SaaS accelerated large file transfer technology is expanding to other industries. Far faster and more reliable than older technologies like FTP and HTTP, this solution can also be delivered as a service, so users do not have to worry about provisioning hardware and software infrastructure, including scaling and balancing servers for load peaks and valleys. The âexpertiseâ many worry about needing is a non-issue because the solution is so simple to use. And itâs being used in particular to push large volumes to cloud storage for all kinds of time-sensitive projects, including big data analytics. For example, scientists are analyzing images of snow and ice cover to learn more about climate change, and (interesting though less benevolent) businesses are analyzing images of competitorsâ parking lots â counting cars by make and model â in order to understand the shopping habits and demographics of their customers.
Itâs always fascinating to see how innovation occurs. It almost never springs from nothing, but is adapted from techniques and technologies employed somewhere else to solve a different challenge. Who would have thought, at the turn of the century, that the technology developed for Media & Entertainment would be so relevant to big data scientific, government and business analytics? And that technology used to produce and delivery entertainment could be leveraged for the betterment of society?
Originally posted via “Borrowing Technology from Media & Entertainment for Big Data Analytics in the Cloud”
), as shown in the figure below.
