The global cloud index is a tool developed by Cisco in an attempt to estimate global data center and cloud-based traffic growth and trends. A complementary resource to existing empirical studies around IP traffic, this report is an important measuring stick for vendors, enterprises, and anyone who likes to know where the proverbial hockey puck is traveling.
Putting aside discussions about Cisco's long-term prognosis, there is no denying that the company that has huge exposure to internet traffic via its massive customer base. Generating some insights from that information, and teasing it out into macro trends, is a useful piece of work.
So what findings does this year's global cloud index come up with? At a high level, when it comes to global traffic, annual global data center IP traffic is projected to reach 10.4 ZB (zettabytes) by the end of 2019, up from 3.4 ZB per year in 2014. At the same time, annual global cloud traffic is projected to quadruple, reaching 8.6 ZB (719 EB per month) by the end of 2019, up from 2.1 ZB per year (176 EB per month) in 2014, and is expected to account for more than four-fifths (83%) of total data center traffic by 2019.
Annual global cloud IP traffic is expected to reach 8.6 ZB by the end of 2019, up from 2.1 ZB per year in 2014. In an interesting glimpse into how new technologies are helping drive efficiencies in spite of this massive increase in traffic, networking technologies such as SDN and NFV are expected to streamline data center traffic flows such that the traffic volumes reaching the highest tier (core) of the data center may fall below 10.4 ZB per year, and lower data center tiers could carry over 40 ZB of traffic per year.
In terms of how this traffic looks on a regional basis, perhaps unsurprisingly North America will have the highest cloud traffic volume (3.6 ZB) by 2019, followed by Asia Pacific (2.3 ZB) and Western Europe (1.5 ZB). North America will also have the highest data center traffic volume (4.5 ZB) by 2019, followed by Asia Pacific (2.7 ZB) and Western Europe (1.8 ZB).
To give an idea of what these measurements mean in terms we can understand, 8.6 zettabytes is equivalent to:
- 120 trillion hours of streaming music (Equivalent to about 21 months of continuous music streaming for the world’s population in 2019).
- 21 trillion hours of business web conferencing with a webcam (Equivalent to about 17 hours of daily web conferencing for the world’s workforce in 2019).
- 5.7 trillion of high-definition (HD) movies viewed online (Equivalent to about 2.0 daily streamed HD movies for the world’s population in 2019).
The comparisons go on, but you get the drift -- this is a serious amount of traffic. The other interesting finding is around the movement of data away from fixed devices and into the cloud. Today, 73% of data stored on client devices resides on PCs. By 2019, the majority of stored data (51%) will move to non-PC devices (e.g., smartphones, tablets, M2M modules, et al.). By 2019, 55% of the residential Internet population will use personal cloud storage (up from 42% in 2014). Globally, consumer cloud storage traffic per user will be 1.6 gigabytes per month by 2019, compared to 992 megabytes per month in 2014.
The report tried to estimate what the Internet of Things will mean for traffic generation. In particular, it looked at the long-promised notion of the smart city. The report found that a smart city of 1 million people will generate 180 million gigabytes of data per day by 2019.
Finally, the report looked at the split between public and private cloud infrastructure. While this is a topic for much debate, and will continue to be so going into the future, the report extrapolated historical trends to come up with a picture of the future for the cloud. Public cloud workloads are predicted to grow at 44% Compound Annual Growth Rate (CAGR) from 2014 to 2019, while private cloud workloads will grow at a slower pace of 16% CAGR from 2014 to 2019. What this means is that by 2019, 56% of the cloud workloads will be in public cloud data centers, up from 30% in 2014 (CAGR of 44% from 2014 to 2019). Conversely, by 2019, 44% of the cloud workloads will be in private cloud data centers, down from 70% in 2014 (CAGR of 16% from 2014 to 2019).
Obviously, these are mere predictions, and the one thing we know about predictions is they rarely prove completely accurate. But as a glimpse of the future, this is some interesting work.