CORE Metadata, citation and similar papers at core.ac.

uk
Provided by Institute of Business Management, Karachi, Pakistan: Journal Management System

Pakistan Journal of Engineering Technology and Science (PJETS)

Volume 6, No 2, December 2016

Impact of Big Data over Telecom

Industry
Muhammad Waqar Khan1, *Muhammad Asghar Khan1, Muhammad Alam1, 2,
Wajahat Ali1
1
Institute of Business and Management (IoBM), Korangi Creek, Karachi-75190,
Pakistan
2
Institute of Post Graduate Studies,Universiti Kuala Lumpur, Jalan Sultan
Ismail,10250,Kuala Lumpur, Malaysia

Abstract- During past few years, data is growing exponentially attracting researchers to work o
a popular term, the Big Data. Big Data is observed in various fields, such as information technology,
telecommunication, theoretical computing, mathematics, data mining and data warehousing. Data
science is frequently referred with Big Data as it uses methods to scale down the Big Data. Currently
more than 3.2 billion of the world population is connected to internet out of which 46% are connected
via smart phones. Over 5.5 billion people are using cell phones. As technology is rapidly shifting from
ordinary cell phones towards smart phones, therefore proportion of using internet is also growing. There
is a forecast that by 2020 around 7 billion people at the globe will be using internet out of which 52% will
be using their smart phones to connect. In year 2050 that figure will be touching 95% of world population.
Every device connect to internet generates data. As majority of the devices are using smart phones to
generate this data by using applications such as Instagram, WhatsApp, Apple, Google, Google+, Twitter,
Flickr etc., therefore this huge amount of data is becoming a big threat for telecom sector. This paper is
giving a comparison of amount of Big Data generated by telecom industry. Based on the collected data
we use forecasting tools to predict the amount of Big Data will be generated in future and also identify
threats that telecom industry will be facing from that huge amount of Big Data.

Keywords: Big Data, Data Science, Telecommunication.

I. Introduction
Big data analytics is not just a passing trend; it is becoming an important part in every aspect
of a communication service [1]. Whenever we talk about the electronic communication it means we
are talking about the production of data by using wired or wireless medium. The quick growth of
internet and the availability of technology everywhere, internet users are increasing day by day. On
the other hand, the boom in smartphone industry makes it easier for users to access network on the
move. This ease of access is becoming great threat for telecom industry. People are diverting from
wired to wireless medium especially on GSM to access the internet [2]. Mobile applications such as
Instagram, WhatsApp, Apple, Google, Google+, Twitter, Flickr etc., providing ease to connect, at
the same time are generating huge amount of data to tackle with.

*[email protected]

116
 Pakistan Journal of Engineering Technology and Science (PJETS)
Volume 6, No 2, December 2016

Figure 1 is representing the internet traffic trends from year 2015 to 2020 [3]. Figure 2 is
elaborating it more in terms of internet users against world population in percentage, growth in
using internet by using smartphones, average speed of internet and average speed of per Capita per
month [2].

Figure 1: Internet traffic from 2015 to 2020 [3]

Figure 2: Internet users against world population in percentage, growth in using internet by using smartphones, average speed
of internet and average speed of per Capita per month [2]

This paper is further divided into six sections; section 2 is presenting the results of previous
research in Big Data and its impact over telecom industry. Section 3 is addressing the problem
in the context of Big Data handling in telecom sector. Section 4 is presenting the forecasting and
prediction of Big Data usage in telecom sector in next 5, 10 and 30 years. Last section concludes
this paper with future directions.

117
 Pakistan Journal of Engineering Technology and Science (PJETS)
Volume 6, No 2, December 2016

II. Literature Review

This section is divided into two subsections, which are
A. Big Data
B. Big Data in Telecom Sector

A. Big Data
In [4], authors talk about the technologies, opportunities and challenges of Big Data. According
to them Big Data management and analysis can be done by considering five (5) common issues
which are volume, variety, velocity, value and complexity [4, 5, 6]. Volume of Big Data is normally
huge. However, huge amount of data in petabytes or zettabytes is not required. In order to cope with
volume problem online additional storage can be managed. Variety is another issue in Big Data
management. Different type of data is produced by different devices it need to be fault tolerant.
Researchers used different data mining algorithms to deal with variety. Velocity means speedup
data. Normalize data provides the velocity. Authors also suggested a life cycle for Big Data by
using terminologies and technologies used for Big Data. Stages they include for the life cycles are
collection, filtering, analysis, storage, publication, retrieval and discovery [4].
One of the biggest problems in the Big Data is the unstructured data, which is well known as
“human information”. Financial records, scientific computation and simulation, videos, still images,
geospatial images, weather records, call centers data generated by Facebook, tweeter, email and
WhatsApp etc. all fall under this category [7]. Unstructured data makes up to 80% of all of the Big
Data [7, 8]. Out of all Big Data, 84% of the data is unstructured not modeled and random. It is very
difficult to analyze this huge amount of unstructured data [7].
Volume of the Big Data is becoming an alarming threat over the last past few years. Around 5
billion people over the globe call, text, tweet, send emails and browse on mobile devices [9]. In year
2012 around 89 billion emails were received and sent daily and it is forecasted that it will increase
at the rate of 13% and by 2020, this amount will be touching 143 billion [10]. In year 2012, 34% of
all emails were sent through smart phones [10], and this number is also increasing rapidly with the
growth of and usage of smart phones [4, 11-18].

B. Big Data in Telecom Sector

Approximately 40% of the world population has internet connection today. The number of
internet users has increased dramatically from 0.4 Billion in 2000 to 3.4 Billion in 2016 [19].
According to 2016 edition of Mary Meeker’s annual internet trend report [20], China has the biggest
internet population followed by India and USA. Total internet traffic has increased enormously in

118
 Pakistan Journal of Engineering Technology and Science (PJETS)
Volume 6, No 2, December 2016

the last two decades. In 1992, internet transferred around 100 GB of data every day, whereas these
figures shoot up to 100 gigabytes per second (GBps) in 2002. Furthermore, total internet traffic
reached above 20,000 GBps in 2015 [21].
Mobile computing has changed the landscape of workplace dramatically and brought many
advantages to internet users such as wireless connectivity, increased productivity and easy mobility.
Google Consumer Barometer [22] shows positive trends towards acceptance of mobile devices both
in developed and developing countries. For example, in China and USA, the percentage of people
who use a smart phone for internet has increased from 33% and 44% in 2012 to 79% and 72% in
2016 respectively. Smart phones and tablets sales have already surpassed the Workstation/PC sale.
Internet live statistics [23] counter shows that more than 1.6 Billion smart phones sold in the first
eight months of 2016 compared to 142 Million computers.
 Applications of big data are countless and organizations of all sizes are looking to improve
their performance by using big data analytics tools [24]. Yin and Kaynak [25] identified many
challenges faced by enterprises when dealing with big data. One of the key challenges for the
IT professionals and researcher is to deal with the application that continuously generates large
quantity of data at very high speed. Enormous growth has been observed in the data production
during the last two years, in fact, 90% of the world’s current data has been generated in the last
two years alone [26]. Massive amount of data is continuously generated and transferred through
network by various internet entities, such as, video on demand, news streaming and e-commerce
websites. In the era of social networking, websites are producing data at enormously high speed.
For example, Twitter generates over 350,000 tweets per minute and 500 million tweets per day
on average [27]. Similarly, Google processes around 40,000 search queries in every second on
average [28]. Likewise, astronomy effort, such as Palomar Transient Factory in Southern California,
is looking for new phenomena in the sky and roughly captures 30 terabytes of data every night [29].
Mobile devices produce huge quantity of data at enormously high speed. Cisco has developed
different forecasting tools [30], based on their Visual Networking Index(VNI), that predict that
“Global mobile traffic will grow three times faster than Global fixed IP traffic from 2015 to 2020”.
Furthermore, “Global mobile data traffic will grow 8-fold from 2015 to 2020, a compound annual
growth rate of 53%”. This massive shift of IP traffic from fixed to wireless network will introduce
number of challenges for Network architects. Telecom companies are sitting on a gold mine, as they
have plenty of data. But what they require is a proper analysis of both structured and unstructured
data to get deeper insights into customer behavior, their service usage and interests real-time [31].
Since, half of the data is now being transferred via mobile and in 2012, according to study 90%
of the data will be transferred through mobile and tablet devices (wireless connections). It could be a
threat for telecom industry as they have already invested millions of dollars to process and store the
big data via wire line connections. Now the technology is getting changed and majority of the data

119
 Pakistan Journal of Engineering Technology and Science (PJETS)
Volume 6, No 2, December 2016

is being transferred on wireless connections [32]. But just to sustain in the market, telecom industry
needs to reinforce their strategies based on the current scenario. They also have an opportunity to
target this untapped market. The telecom industry has an advantage over other industries due to the
maximum bandwidth and depth of data it collects in the course of normal business. For example,
an operator serving 8 million prepaid mobile subscribers generates around 30 million Call Data
Records daily, equaling 11 billion records annually. If the same operator also provides postpaid and
fixed lines services, then there is even more volume and variety of data at the ready. Big Data helps
telecom operators improve their marketing effectiveness. Relevant actions can be taken based on
real-time information without the need to wait for data extraction or manual data mining. Big Data
can help gather real-time customer satisfaction information through social media listening or Voice
of the Customer analysis. It can also improve customer experience by identifying the most valuable
customers who would benefit from dedicated treatment and better services [32].
Different subscription fees can be charged to customer according to their usage on monthly
basis. Cloud based services could help telecommunication industries to manage and secure
customer’s data. Telecom Industry can create new products just to utilize the big data with wireless
connections. Top players like Google, Skype, and Netflix may be interested in behavioral data to
target consumers with specific content.

III. Problem Statement

Technology is rapidly shifting from wired to wireless networks and every node is busy to generate
data. This technology shift in data growth is predicting that we will face serious infrastructural
challenges, which include but not limited to, bandwidth problem, network maintenance and support
issues, computational bottleneck and MSC’s and BSC’s congestions. Currently organizations that
are dealing with Big Data are increasing resources for load balancing. However, increasing the
number of IT resources might solve the problem for the short run but this approach is not scalable
for the future growth. IT companies, such as Google, Facebook and BT are looking for some
improved protocols and algorithms to improve data transmission and computing. Forecasting the
future data growth/transmission in short and long run will help IT researchers, Telecom companies,
Governments and other stakeholders to understand where they currently stand and where and in
which direction they need to move to solve above mentioned problem. This research paper is using
Time Series forecasting to understand the data needs in 2020, 2030, 2040 and 2050.

120
 Pakistan Journal of Engineering Technology and Science (PJETS)
Volume 6, No 2, December 2016

IV. Proposed Methodology

We have collected data for past 16 years from year 2000 to 2015 of world population and
compare it with internet and non-internet users. Based on the available data by using regression
we predict for 2020, 2030, 2040 and year 2050 as shown in table 1. Rapid growth of internet users
is showing that the amount of data will be immense in year 2050. Trends shows that most of the
data will be generated by cellphones/smartphones, because of their handy and available every time
nature. This will overburden telecom industry because it will be dealing a huge amount of Big Data.
Current research shows that more than 52% of users are connected via smartphones as compare to
41% via desktop [36].

Table 1
Year wise distribution of world population, internet and non-internet users
Year Internet Users World Population Non Users
2000 414,794,957 6,126,622,121 5,711,827,164
2001 502,292,245 6,204,310,739 5,702,018,494
2002 665,065,014 6,282,301,767 5,617,236,753
2003 781,435,983 6,360,764,684 5,579,328,701
2004 913,327,771 6,439,842,408 5,526,514,637
2005 1,030,101,289 6,519,635,850 5,489,534,561
2006 1,162,916,818 6,600,220,247 5,437,303,429
2007 1,373,226,988 6,681,607,320 5,308,380,332
2008 1,575,067,520 6,763,732,879 5,188,665,359
2009 1,766,403,814 6,846,479,521 5,080,075,707
2010 2,023,202,974 6,929,725,043 4,906,522,069
2011 2,231,957,359 7,013,427,052 4,781,469,693
2012 2,494,736,248 7,097,500,453 4,602,764,205
2013 2,728,428,107 7,181,715,139 4,453,287,032
2014 2,956,385,569 7,265,785,946 4,309,400,377
2015 3,185,996,155 7,349,472,099 4,163,475,944
2016 3,424,971,237 7,432,663,275 4,007,692,038
2017 3,448,735,286 7,507,238,701 4,058,503,415
2018 3,640,880,317 7,589,116,161 3,948,235,844
2019 3,833,025,348 7,670,993,621 3,837,968,273
2020 4,025,170,380 7,752,871,081 3,727,700,702
2021 4,217,315,411 7,834,748,541 3,617,433,130

121
 Pakistan Journal of Engineering Technology and Science (PJETS)
Volume 6, No 2, December 2016

2022 4,409,460,443 7,916,626,001 3,507,165,558

2023 4,601,605,474 7,998,503,461 3,396,897,987
2024 4,793,750,505 8,080,380,921 3,286,630,416
2025 4,985,895,537 8,162,258,381 3,176,362,844
2026 5,178,040,568 8,244,135,841 3,066,095,273
2027 5,370,185,600 8,326,013,301 2,955,827,701
2028 5,562,330,631 8,407,890,761 2,845,560,130
2029 5,754,475,662 8,489,768,221 2,735,292,559
2030 5,946,620,694 8,571,645,681 2,625,024,987
2031 6,138,765,725 8,653,523,141 2,514,757,416
2032 6,330,910,757 8,735,400,601 2,404,489,844
2033 6,523,055,788 8,817,278,061 2,294,222,273
2034 6,715,200,820 8,899,155,521 2,183,954,701
2035 6,907,345,851 8,981,032,981 2,073,687,130
2036 7,099,490,882 9,062,910,441 1,963,419,559
2037 7,291,635,914 9,144,787,901 1,853,151,987
2038 7,483,780,945 9,226,665,361 1,742,884,416
2039 7,675,925,977 9,308,542,821 1,632,616,844
2040 7,868,071,008 9,390,420,281 1,522,349,273
2041 8,060,216,039 9,472,297,741 1,412,081,702
2042 8,252,361,071 9,554,175,201 1,301,814,130
2043 8,444,506,102 9,636,052,661 1,191,546,559
2044 8,636,651,134 9,717,930,121 1,081,278,987
2045 8,828,796,165 9,799,807,581 971,011,416
2046 9,020,941,196 9,881,685,041 860,743,845
2047 9,213,086,228 9,963,562,501 750,476,273
2048 9,405,231,259 10,045,439,961 640,208,702
2049 9,597,376,291 10,127,317,421 529,941,130
2050 9,789,521,322 10,209,194,881 419,673,559

A. Regression Analysis of time series analysis using SPSS:

Linear trend is found between time and Internet users. Time is considered as independent
variable (x) and Internet users as dependent variable. The equation describe as [33-35]:

Y (internet users) = - 3.841 x 1011 +1.92x108 x (time)

122
 Pakistan Journal of Engineering Technology and Science (PJETS)
Volume 6, No 2, December 2016

According to regression line estimated internet user would be raised 5.95 billion in 2030 and
9.79 billion in 2050; this is 69.4% and 95.5% of the total world population respectively. Following
graph describe Linear regression support trend line and highly significant. The correlation r=0.992
define highly positive relationship between time and internet users. Figure 3 and 4 are representing
trends of internet users till year 2050. Figure 5 is presenting the user inclination towards cellphones.
2015 data shows that 51% of users are using smartphones as compare to desktop users which were
only 42% and dropping [36].

Figure 3: Bar chart representation of world population, internet users and non-internet against time.

Figure 4: Trend line of world population, internet users and non-internet against time.

123
 Pakistan Journal of Engineering Technology and Science (PJETS)
Volume 6, No 2, December 2016

Figure 5: Figure by Dave Chaffey, showing the increase in use of mobile phones over desktop over the years [36]

V. Conclusion, Limitation and Future Directions

Forecasting by using regression shows that in year 2050 95% of world population will be
connected via internet and most of it will be using wireless mode to connect especially GSM or
future telecom architecture. It will generate huge amount of data for telecom industry for which the
industry is not ready to deal with. Most of this data will be redundant therefore proper data mining
tools and techniques are required to dig out for the required data and dump the redundant data.

References
[1] Big data analytic is just not a passing trend. http://www.analysysmason.com/Research/Content/Reports/Big-data-
analytics-Jul2014/ (last visited: September 16th, 2016)
[2] http://www.cisco.com/c/m/en_us/solutions/service-provider/vni-forecast-highlights.html (last visited: September 16th,
2016)
[3] http://www.cisco.com/c/dam/m/en_us/solutions/service-provider/vni-forecast-widget/forecast-widget/advanced.html
(last visited: September 16th, 2016)
[4] Khan, N., Yaqoob, I., Hashem, I.A.T., Inayat, Z., Mahmoud Ali, W.K., Alam, M., Shiraz, M. and Gani, A., 2014. Big
data: survey, technologies, opportunities, and challenges. The Scientific World Journal, 2014.
[5] S.Kaisler, F.Armour, J.A.Espinosa, W.Money, “Big Data: issues and challenges moving forward”, Proceedings of the
IEEE 46th Annual Hawaii International Conference on System Science (HICSS’13), pp.995-1004, January 2013.
[6] S.Sagiroglu, D.Sinanc, “ Big Data: A review” Proceedings of the international onference on Collaboration Technologies
and Systems (CTS’13), pp 42-47, IEEE, San Diego, Calif, USA. May 2013.
[7] K.Douglas, “Infographic: Big Data brings marketing baig numbers” 2002, http://www.marketingtechblog.com/ibm-

124
 Pakistan Journal of Engineering Technology and Science (PJETS)
Volume 6, No 2, December 2016

big-data-marketing/. (Last visited: August 20, 2016)

[8] A.Holzinger, c.Stocker, B.Ofner, G.Prohaska, A.Brabenetz and R.Hofmann-Wellenhof, “Combining HCI, natural
language processing and knowledge discovery- potential of IBM content analytics as an assistive technology in the
biomedical field”, in Human-Computer Interaction and Knowledge Discovery in Complex, Unstructured Big Data, vol.
7947 of Lecture Notes in Computer Science, pp. 13-24, Springer, Berlin, Germany, 2013.
[9] P.Russom, “Big Data analytics”, TDWI Best Practices Report, Fourth Quarter 2011.
[10] S.Radicati and Q.Hoang, “ Email Statistics Report”, 2012-2016. The Radicati Group, London, UK, 2012.
[11] Worldmeters, “Real time world statistics”, 2014 http://www.worldmeters.info/world-population/. (Last visited: August
22, 2016).
[12] M.Chen, S.Mao and Y.Liu, “Big Data: A Survey”, Mobile Networks and Applications, Vol. 19, No. 2, pp. 177-209,
2014.
[13] S.Kaisler, F.Armour, J.A.Espinosa and W.Money, “Big Data: Issues and Challenges moving forward”, in Proceedings
of the IEEE 46th Annual Hawaii international Conference on System Sciences (HICSS’13), pp.995-1004, January
2013.
[14] A Muhammad, “Power Management of Portable Devices by Using Clutter Based Information” IJCSNS, International
Journal of Computer Science and Network Security, VOL. 9 No. 4, April 2009, tel-00815919, version 1-19 Apr 2013
pp 237-244.
[15] A Muhammad, MS Mazliham, M Shahrulniza, M Amir, “Posterior Probabilities based Location Estimation (P2LE)
Algorithm for Locating a Mobile Node in a Disaster Area M” MULTICONF p 13-16, 2009.
[16] Muhammad Alam, Mazliham Muhammad Suud, Patrice Boursier, Shahrulniza Musa, Jawahir Che Mustapha Yusuf,
“Predicted and corrected location estimation of mobile nodes based on the combination of kalman filter and the bayesian
decision theory” , International Conference on Mobile Wireless Middleware, Operating Systems, and Applications,
Pages 313-325, Publisher Springer Berlin Heidelberg
[17] A Muhammad, MS Mazliham, Patrice Boursier, M Shahrulniza, Jawahir Che Mustapha Yusuf, “Terrain/clutter based
location prediction by using multi-condition Bayesian decision theory” Proceedings of the 6th International Conference
on Ubiquitous Information Management and Communication. P. 110-117, 2012.
[18] Muhammad Alam, “Error corrected location determination in an outdoor wireless environment by using estimation,
filtering, prediction and fusion techniques: A Wi-Fi application by using terrain based knowledge” , HAL, 2013 pages
313.
[19] http://www.internetlivestats.com (last visited: October 19th, 2016)
[20] http://www.kpcb.com/blog/2016-internet-trends-report (last visited: October 19th, 2016)
[21] The Zettabyte Era—Trends and Analysis, Cisco, http://www.cisco.com/c/en/us/solutions/collateral/service-provider/
visual-networking-index-vni/vni-hyperconnectivity-wp.html (last visited: September 27th, 2016)
[22] https://www.consumerbarometer.com/en/trending/?countryCode=CN&category=TRN-NOFILTER-ALL (last visited:
September 27th, 2016)
[23] http://www.internetlivestats.com (last visited: September 27th, 2016)
[24] BUILDING A SCALABLEBIG DATA INFRASTRUCTURE FOR DYNAMIC WORKFLOWS,
[25] http://www.emc.com/collateral/solution-overview/h11761-building-scalable-big-data-infra-so.pdf (last visited:
September 27th, 2016)
[26] YIN, BS, Kaynak, O (2015),Big Data for Modern Industry: Challenges and Trends, Proceedings of the IEEE, vol.
103, No. 2.
[27] Mukerjee, S and Shaw, R 2016, Big Data – Concepts, Applications, Challenges and Future Scope, International Journal
of Advanced Research in Computer and Communication Engineering, vol. 5, no. 2.
[28] http://www.internetlivestats.com/twitter-statistics/ (last visited: September 27th, 2016)
[29] http://www.internetlivestats.com/google-search-statistics/ (last visited: September 27th, 2016)
[30] http://www.nytimes.com/2015/08/01/science/research-scientists-to-use-network-much-faster-than-internet.html (last
visited: September 27th, 2016)
[31] VNI Complete Forecast Highlights Tool (2016), http://www.cisco.com/c/m/en_us/solutions/service-provider/vni-

125
 Pakistan Journal of Engineering Technology and Science (PJETS)
Volume 6, No 2, December 2016

forecast-highlights.html (last visited: September 27th, 2016)

[32] Telecom companies are sitting on a gold mine. http://bigdata-madesimple.com/11-interesting-big-data-case-studies-in-
telecom/ (last visited: September 27th, 2016)
[33] Big Data helps telecom operators improve their marketing effectiveness. http://www.teletech.com/resources/articles/
numbers-three-big-data-opportunities-telecoms#.V82-gvmqqko (last visited: September 27th, 2016)
[34] Rob J Hyndman and George Athana¬sopou¬los Journal of Forecasting J. Fort. 20, 315-328(2001) DOI: 10.1002/
for.795
[35] Joseph Nguyen by Regression Basics for Business Analysis
[36] Statistics for Management 7th Edition by Richard Levin
[37] Dave Chaffey, “Statistics on consumer mobile usage and adoption to inform your mobile marketing strategy mobile
site design and app development”, Mobile Marketing Statistics compilation: Mobile Market Analytics, April 27, 2016.

126