leveraging big data · 2016-01-22 · 2014-2018”, the radicati group, 2014; “google annual...
TRANSCRIPT
Looking Ahead 192
Source: “Cisco Visual Networking Index”, Cisco, 2014
Total Internet Data Traffic(In Exabytes per Month, 2013-2018)
Although it is hard to measure data accurately, it is clear that data is growing at exponential rates around the world. The sheer volume of information will require new analytical skills and more storage capacity. One type of data that could be measured is online data/mobile data
ONLINE DATA TRAFFIC VOLUME IS GROWING AT UNPRECEDENTED RATES with mobile data growing the fastest albeit from a low base
LEVERAGING BIG DATA31
60
90
120
30
0
105
135
15
45
75 Other Internet
Mobile
2018201520142013 20172016
18%
61%
CAGR(In %,
2013-2018)
0.40.81.2
2.0
3.43.44.24.3
5.0
9.6
Air TransportConsultingRetail Construction Publishing TelecomFoodProducts
Steel AutoMobile IndustrialInstruments
18%49% 39% 20% 19% 18%ProductivityIncrease 17%20% 17%36%
LA_PublicVersion(Bleed3175)_ENG_V22.indd 192 10/28/15 1:29 PM
0.40.81.2
2.0
3.43.44.24.3
5.0
9.6
Air TransportConsultingRetail Construction Publishing TelecomFoodProducts
Steel AutoMobile IndustrialInstruments
18%49% 39% 20% 19% 18%ProductivityIncrease 17%20% 17%36%
Big Data
Productivity and Sales Increase Due to Big Data Analytics(In % and US$ Billion, 2010)
VARIETY VOLUME VELOCITY
>2,500
North America
Europe
>2,000
Japan >400Rest ofAPAC >300
China >250Middle East
>200India>50
Latin America
People to Machine
Virtual Communities
Social Networks
Web logs
Archives
Digital TVs
E-Commerce
Sensors
GPS devices
Surveillance Cameras
People to People
Machine to Machine
1.2 Billionusers worldwide
196.3 Billionemail sent per day
2.095 Trilliongoogle searches in
2014
Amount of Big Data Stored across the Globe(In Petabytes, 2011)
Sources- Upper and Lower Charts: “Big data: The next frontier for innovation, competition, and productivity”, McKinsey Global Institute, 2011; “Email Statistics Report 2014-2018”, The Radicati Group, 2014; “Google Annual Search Statistics”, Statistic Brain, 2014; “Big Data”, Wipro Applying Thought, 2011
Analytics of interactions between people, between machines, and between people and machines have been helpful in identifying trends about consumers and have become a part of companies’ efforts to improve their productivity and sales. Still, to get insights from Big Data, companies have to overcome certain challenges, such as filtering noise
THE VOLUME, VELOCITY AND VARIETY OF “BIG DATA” have led to higher productivity and have enabled customization
Looking Ahead 193
LA_PublicVersion(Bleed3175)_ENG_V22.indd 193 10/28/15 1:29 PM
Insta
Frequency (per day) Large User Country Demographic Characteristics
2.5 Billion piecesof content shared
>100 Millionusers in India alone
925 Thousandnew users
500 Milliontweets
>172Thousand users
sign-up
1.5 Million usersin Thailand alone
23% of teensconsider it theirfavorite social
network
80%female
Facebook Youtube Google+ Twitter LinkedIn Instagram Pinterest
SaudiArabia
is the biggest userper capita
200 Million hoursof video watched
144 thousandhours uploaded
Sources- Upper and Lower Charts: Digital Insights Website; We Are Social; The Social Media Hat Website
The surge in the use of social media is producing a new stream of data. While social networks are mostly made up of young users, older users are joining at an even more rapid pace as smartphones prompt this age cohort to experiment with new applications. The number of active users of the largest social networks is now equivalent to the population of entire countries or continents; Facebook’s user base rivals the population of China. The frequency of use of social networks—with people accessing them, in many cases, several times a day—hints at the immense amount of information on the web
Leading Social Networks by Country/Region Equivalence in Terms of Number of Users(In Million People and In Million Active Users, Latest Available Data)
Key Statistics
UNSTRUCTURED DATA IS ALSO ON THE RISE, driven by the popularity of social networks, which now reach 28% of the global population
1,350
1,000
540
300 284187 182250
316
596
1,032
1,357
Google+LatinAmerica &Caribbean
40
China
30
YoutubeAfricaFacebook Oceania
Pinterest Snapchat
Malaysia
Indonesia
30
InstagramUS Pakistan
.
37
Looking Ahead 194
LA_PublicVersion(Bleed3175)_ENG_V22.indd 194 10/28/15 1:29 PM
Looking Ahead 195
Insta
Frequency (per day) Large User Country Demographic Characteristics
2.5 Billion piecesof content shared
>100 Millionusers in India alone
925 Thousandnew users
500 Milliontweets
>172Thousand users
sign-up
1.5 Million usersin Thailand alone
23% of teensconsider it theirfavorite social
network
80%female
Facebook Youtube Google+ Twitter LinkedIn Instagram Pinterest
SaudiArabia
is the biggest userper capita
200 Million hoursof video watched
144 thousandhours uploaded
Note: (1) PaaS refers to Platform as a Service, IaaS refers to Infrastructure as a Service, and SaaS refers to Software as a ServiceSource- Upper Chart: “Cloud Forecast and Methodology 2013-2018”, Cisco, 2014 Source- Lower Charts: “Global Mobile Data Traffic Forecast Update, 2013-2014”, Cisco, 2014
Breakdown of Cloud Traffic by Type(In %, 2013 and 2018)
Global Data Traffic over Time from Traditional and Cloud Data Centers(In Exabytes per Year, 2013-2018)
AN INCREASING AMOUNT OF DATA IS BEING STORED IN THE CLOUD, particularly from users in Asia and North America
For many organizations and people, the cloud has emerged as the answer to their data challenges – providing a scalable, flexible and automated platform that can handle both structured and unstructured data. The cloud is also a cost-effective way to support big data technologies (and data analytics) while reducing overhead costs. Risks remain, however, especially with respect to data security
7,072
1,941(27%)
2017
5,131(73%)
1,680(36%)
3,050(64%)
2014
3,828
1,551(41%)
2,277(59%)
2013
3,064
1,417(46%)
1,647(54%)
6,496(76%)
2018
CloudData Center
8,575
2,079(24%)
TraditionalData Center
2016
5,799
1,805(31%)
3,994(69%)
2015
4,730
Forecast
CAGR(In %, 2013-2018)
32
8
23
Consumer Business
2013
2018
59% 41%
66% 34% 2013
2018
22% 78%2013
2018
41% 44%
Public Private PaaS(1) Iaas(1) Saas(1)
31% 69%
15%
59% 28%13%
North AmericaAsia Pacific Western EuropeCentral and Eastern Europe Latin America Middle East and Africa
Cloud Trafficby Region(In %, 2013 and 2018)
2013 2018
LA_PublicVersion(Bleed3175)_ENG_V22.indd 195 10/28/15 1:29 PM
Looking Ahead 196
Source: Advanced Micro Devices
Computer Power Milestones(In Floating Point Operations per Second, 1960-2019)
INFORMATION MANAGEMENT HAS BEEN FACILITATED BY THE INCREASED POWER OF TRANSISTORS, which have become much faster
As processing power has increased, the estimate of how much time it will take to match the capacity of the human brain has decreased to around the year 2020. It would appear that Moore’s law--predicting the doubling of the number of transistors on an integrated circuit every two years--has been met. In fact, some experts believe we have reached “More Moore”, with transistor size and cost decreasing
1960
2009: Cray XT5-HE Goes Live
1970 1980 1990 2000 2010 2020
103
106
2008: PetaFLOPBarrier Broken
109
1012
1015
1018
1960: Univac LARC Goes Live
1976: Cray 1 Goes Live
1984: M-13 Supercomputer1993: CM-5/1024 Supercomputer
1999: ASCI Blue Pacific Goes Live
2003: Human Genome Mapped
2005: Millennium Run Simulation
2009: First World-class GPU-powered Supercomputer
2019: exaFLOPBarrier to be Reached?
Floating Point Operations perSecond (FLOP)
Equivalent computing power
Human Brain
Mouse Brain
Insect Brain
LA_PublicVersion(Bleed3175)_ENG_V22.indd 196 10/28/15 1:29 PM
Looking Ahead 197
Sources- Upper Chart: Direct Capital; AmazonSource- Lower Chart: “No Moore? A Golden Rule of Microchips Appears to be Coming to an End”, The Economist, Nov. 2013 (based on Linley Group)
97.0 100.0
219
5.5
660
7
1,049
0.0 95.5 96.594.5 98.5 99.594.0 97.593.593.0 95.0 99.096.0 98.0
1970s
1980
1984
1965COMPUTERS
CALCULATORS(Sharp)
PHOTOCOPIER(Xerox)
RAM (1 GB Data)
X Price in US$, Latest Available Data
MOBILE PHONES
Year of Reference% Decrease
Price of Sample Technologies over Time(In % Decrease, Various Years)
Transistors’ size will continue to decrease in the future but prices will rise slightly, according to the Linley Group. With price reductions as high as 100% for 1GB of RAM since 1980, and big improvements in computation capacity in personal computers, the handling, storage and analysis of huge amounts of data have become possible. Technological developments have also enabled the processing of data in the cloud where there have been additional savings
TRANSISTOR AND TECHNOLOGY-DEVICE PRICES HAVE DROPPED SUBSTANTIALLY; going forward transistors will continue to drop in size but prices may inch back up
Evolution of Transistors(In Millions of Transistors bought per US$ and In Nanometer Length, 2002-2015)
192020
16
11
7
4
2002 2004 2006 2010 2012 201520142008
3
180nm
130nm
28nm
16nm
90nm
65nm
40nm20nm
20MnNail
Comparative Sizes (In Nanometers)
Bacteria2,000 Hair
90,000
LA_PublicVersion(Bleed3175)_ENG_V22.indd 197 10/28/15 1:29 PM