Ideas for PPP Business Models

For
RailTel Corporation of India Ltd.








K Manohar Raja
GM/SBC



Contents


FOREWORD ................................................................................................................................................... 3
1. Broadband through PPP is important for RailTel ...................................................................................... 4
Why is broadband through PPP important to RailTel? ............................................................................. 4
Which Model will strengthen RailTel? ...................................................................................................... 4
2. Railtel strengths ........................................................................................................................................ 5
3. PPP Models ............................................................................................................................................... 6
3.1. RailWire .............................................................................................................................................. 6
3.2. Content Delivery Networks ................................................................................................................ 7
3.2.1 Concept ........................................................................................................................................ 7
3.2.2 Statements and Evidence............................................................................................................. 7
3.3.3 Beneficiaries of a Pan-India CDN ............................................................................................... 11
4. Business Summary .................................................................................................................................. 13











FOREWORD

In an interesting article forwarded by a RailTel colleague, it was surprising to note that Nokia
sells the highest number of cameras and not Sony, Canon or Nikon. Bharti is the biggest player
in the music business in India and not HMV or Sarigama. Apple, a computer firm suddenly
became a force in audio (I-pod) and mobile (I-phone) industry. In any industry vertical exposed
to market forces, organizations have to strive to maintain their core competence, evolve and
adapt to newer environments as others encroach into their core areas; more so in technology
companies. RailTel’s business domain of Telecom, being one of the most competitive business
verticals in the industry where market forces and technology advancements change equations
in very short time, RailTel has to secure its core competence and constantly evolve to take on
the market.















1. Broadband through PPP is important for RailTel

Why is broadband through PPP important to RailTel?
India ranks 129
th
as per the ICT Development Index (IDI) for Internet access in the ITU
(International Telecommunication Union) list of 154 Countries. Although India currently has
about 8 million broadband homes, the broadband penetration is far less when compared to
other developing economies. Studies conducted by World Bank indicate that 10% improvement
in broadband penetration would increase the GDP by 1.38% - and this impact is likely to be
higher in developing economies like India.
The Department of Telecommunication (DoT) hopes to move from a modest 8 million to
optimistic 250 million broadband connections by the end of 2012. Although difficult, the figure
could be hopefully attained by early 2015. Assuming an ARPU of Rs320, The market size is an
annual business of Rs96000 Crores. However, experts peg the ARPU at a much higher level
with anticipation of revenues from value added services. With its wide reach, if RailTel targets a
10% market share, the annual business would be Rs 9600crores or better.
Understanding the impact of broadband on economic growth, most countries have spent /
planned substantial spending to improve their penetration and quality of broadband services.
Being a PSU, we could stake a claim to the USO funds if we have viable plans.
Retail service is a huge challenge in terms of technology, investment and brand building. Given
the enormous scale of the challenge, the PPP model for broadband penetration offers an
opportunity for RailTel to leverage its strength and overcome its weaknesses by choosing
partnerships that reinforce RailTel’s commitment to achieve better broadband penetration and
the value we can bring to public investment.
Which Model will strengthen RailTel?
Opportunities for a public-private partnership lie on a continuum where the public sector can
play a micro role, a minor role or a major role. The degree of control and public good required is
essentially a political choice. However, if this was a debate in 2000, it might have been
appropriate for Railways to give away dark fiber for a percentage on the gross revenue.
Networks could have been built faster, fiber utilization would have been good and return on
public investment would have been decent.
Today, all major Telco’s have considerable fiber plant in place already, connecting all the
lucrative markets. Since a commercial entity like RailTel has been incorporated and is
performing creditably in the last 10 years, we should consider models which will strengthen
RailTel and also achieve the national goal of better broadband penetration.



en

While the PPP models can help RailTel scale the formidable challenges of investment,
organizational flexibility and complexity of technology, it will be in RailTel’s interest to look at
models which accentuate RailTel’s strengths and help RailTel move up the value chain so that
we remain a profitable company along with meeting our public goals.
2. RailTel Strengths
To determine which model is appropriate we need to have consensus on our strengths:
 RailTel can create and maintain abundant inter city band width
In the last decade RailTel has done well for shelf and solved the challenges in
transmission technologies of SDH and DWDM which has helped us to create
abundant bandwidth in our network.
While this was our USP a few years ago, it is not anymore. More networks are
getting lit and increased sharing of infrastructure among Telco’s are a major reason.
We need to note the fact that nearly 90% of our business is only intra-circle.
 4,000 PoPs across the length and breadth of the country
While other telcos have built networks of comparable size they do not have PoPs
spread out geographically like RailTel and this is one area where RailTel will retain
its strength for sometime and Telcos will be keen to participate in the PPP model
currently envisaged by CO for this reach.
 Being a PSU, RailTel can lay stake to USO funds and other Govt initiated broad
band projects if we have a credible plan.
Optical Fiber
IP MPLS
NGN Broad band Services
VAS Cloud Computing
SDH
Current
Business
position
Max
Profit
Max
Value
USA, Australia, Korea and several modern economies have seen enormous
government spends to increase broadband penetration as they have realized the
positive effect it has on the GDP and e-governance. Currently, only BSNL is
positioned to receive these funds.
 Highly recognizable brand name known for reliable service with consumers as well
as partners in India.
RailTel has not taken advantage of this attribute at all. The RailWire model attempts
to do this.
3. PPP Models
With the twin goals of increasing broad band penetration and improving the business potential
of RailTel, the following two PPP models are proposed which seek to leverage RailTel’s
strengths and move us up the value chain:
1. RailWire Retail Broadband
2. Content Delivery Networks

3.1. RailWire

The trials of retail broadband service has been on in Whitefield, Bangalore for nearly 2 years
now and presently we have about 2500 customers in a competitive environment. The proposal
has been discussed with RailTel Top Management several times and hence only the gist is
mentioned here.
RailWire seeks to leverage the world-class managed services expertise available in Indian
software companies and also leverage the 85 million cable homes which have the capacity to
turn into high performance broadband access networks by tweaking the Coax cable or using
Cat 5 or OFC cable. Even if 10% of the cable homes are wired, then broad band penetration in
India will double. By taking away the technology complexity and giving a recognizable Govt
brand name, local cable operators will be able to retain their cable customers by increasing the
bouquet of services to the customer. This simple yet highly scalable PPP model will dramatically
increase RailTel’s top and bottom line and also help broadband penetration in a rapid manner.
Subject of RailWire has been discussed several times in RailTel and analyzed and vetted by
Finance. The latest proposal vetted by Finance is embedded hereunder--
C:\Documents and
Settings\AGMRCIL\My Documents\Final railwire\ppp\A Report on RailWire Pilot29sep2010.doc

C:\Documents and
Settings\AGMRCIL\My Documents\Final railwire\ppp\RailWire_Report__(Whole, IBW).xls




3.2. Content Delivery Networks
3.2.1 Concept
Content Delivery Networks are a network of caching servers designed to optimally
deliver content by running sophisticated algorithms. For instance, when there is a 10
Gbps network backbone and 100Gbps central server capacity only 10Gbps can be
delivered, but when 10 servers are moved to 10 edge locations total capacity can be
10*10 (100 Gbps).
The idea of CDN as a multiplier for broadband penetration and the business case for
RailTel is presented in the rest of this paper.
Basic CDN Architecture


3.2.2 Why CDN makes a good business case for RailTel…
The idea is not straight forward and has been missed even in the Broadband
Consultation Paper released by TRAI in June this year. To convey the concept, the
following set of statements are made and evidence provided to validate the statements
and in the process the role of an India facing CDN as a multiplier for broadband will
become evident.
1. Content made for Indian audience is mostly hosted outside India. This „has led to
high international bandwidth cost for Indian ISPs which in turn has led to high
Internet prices for Indian customers. High Internet prices have caused low
penetration levels and thus low profitability of Indian portals.
In a survey, of the top 500 web sites served in India, only 150 are India-focused and
the remaining 350 are international. Greater than 90% of the 150 India-focused web
sites are hosted abroad. Portals have chosen to host abroad where prices are
cheap at the cost of significant performance disadvantage which leads to lower
revenue prospects. The three main issues which lead to this vicious circle in India
are—
(i) High hosting prices
Indian Data Centers charge Rs.5.50 per GB per month while their U.s.
counterparts charge Rs.0.14 per GB per month..
(ii) Less Local Delivery
As most of the content is hosted overseas, this overburdens Indian ISPs that
have to pay for high international bandwidth plus NLD cost.
(iii) High broad band prices
This has led to low Internet penetration and less demand for high bandwidth
content and application.
If the current eco system continues, increase in Internet utilization of
consumers will lead to higher international bandwidth usage leading to
continued high broad band prices. These factors will in turn lead to continued
low Internet penetration levels and increasing Forex outflow. Forex outflow in
terms of hosting and IBW costs is estimated to be greater than $1.055 billion
Saving Forex for India
 By delivering ‘Indian origin’ content locally a CDN can significantly saves Forex
outflows.
 By delivering international content for Indian audience locally, a CDN can lead
to Forex inflow.
 The net changes in FOREX flows (Indian origin + international content) can be
up to $ 5 Billion.

2 TRAI has over simplified the Internet performance issues by stipulating the
contention ratio of 1: 50. The Internet experience in India is seriously affected by “
middle mile issues”
Sometime ago, the biggest hurdle in Internet was the last mile. But today research
has indicated that the biggest bottleneck in achieving performance, reliability and
scalability for commercial grade web application is the ‘middle mile’. In Internet
parlance ‘first mile’ is the origin server infrastructure and fair amount of development
in achieving good first mile performance has been achieved as it is within the
application architect’s control. Last mile refers to the Access Network and with a
slew of technologies developed in the last decade; this problem is considered solved
at least in the western countries and even in India where the number of dial-up users
is quite insignificant and improved performance of newer DSL connections.
The middle mile refers to the nebulous and heterogeneous networks between the
origin server and access network of the end user. The middle mile injects latency
bottlenecks, throughput constraints, reliability problems to the web application
performance.

Like everything else, the Internet also is shaped by market economics. Money flows
into the Internet from the first and last miles as companies pay for hosting and end
users pay for access. Both these sections have grown tremendously in the last
decade, while on the other hand there is very little economic incentive in the middle
mile which is made up of peering and transit points where the networks trade traffic.
As we have observed, in the case of NIXI, Telco’s want to minimize traffic coming
into their networks that do not get paid for. As a result, peering points are often
overburdened causing packet loss and service degradation.
Other issues which plague the middle mile include trans-oceanic cable cuts, power
outages and DDOS attacks and challenges in managing vulnerability issues of BGP
routing and most importantly, the inherent throughput limitation in TCP Protocol. It
was noticed in some of our Railwire experiments where customers where ready to
pay for 8 Mbps bandwidth; the actual Internet performance did not match the
expectation even though we could establish a clear 8 Mbps connection up to the
gateway. The reason is most of the content lies outside India mostly in the U.S. and
TCP over 10,000 mile optic fiber cannot deliver more than about 270 Kbps
throughput(in ideal conditions) . In this scenario the media rich applications which
are popular globally will result in continued poor performance in our country
While the general problems in the middle mile is affecting Internet experience as a
whole throughout the world, India is particularly badly affected as content hosting is
mostly in the U.S. unlike that of China and compounded by the peculiar nature of
Internet gateway business in India.

A statistical comparison between India and Chain’s broadband penetration is
below

 India significantly lags China in Penetration.

 While e-commerce contributes to more than 1.5% of China’s GDP, India experiences
Net FOREX outflow caused by overseas content hosting.




© 2010 Confidential
The Indian Internet ecosystem model is unsustainable in the long run and needs
significant overhauling

3 Current market scenario is encouraging major players to benefit from this
situation and hence there is tendency to maintain status quo.
All the major private Telecom players like Bharti, Tata and Reliance own gateways
and undersea cable with only a fraction of their capacity utilized. Since they provide
overseas Internet traffic going through the gateway, no initiatives have been taken
for local peering and content delivery even though it presents an opportunity. Having
invested considerably in creating International Gateways and shared ownership in
submarine cable systems their business interest lies in consumption of international
bandwidth in India.

4 With initiatives by the govt like NKN, Sakshat Portal, NPTEL, SWAN etc India
requires an India facing CDN to enhance user experience.
NKN and other MHRD initiatives aiming to use technology to enhance learning is
resulting in a lot of local content getting created. While NKN has is creating the
network it has missed out on the CDN which presents an opportunity for RailTel.
Even the flagship NPTEL content is ironically hosted on Youtube as there is no other
content dissemination service in India.
Another reason that CDN is very important to India is the proven fact that lower the
development level or education level of a country then the higher BW applications
like video tend to be more popular.

5 Internet is going to be increasingly media rich with Video growing exponentially
on the Internet and “middle mile” issues will hamper the Internet experience in
India.
The fig below indicates the projected popularity and usage of applications on the
Internet from a developed nation perspective. Since India is generally two years
behind in adoption of technology and services on the Internet we should anticipate
an explosion of media rich applications in the next few years where CDN can play a
big role in enhancing the user experience.




6 Cloud computing is poised to take off in the next few years bringing with it
considerable economies of scale but a CDN based application acceleration is a
must for low latency applications.
Cloud computing applications are latency dependent and require high
throughput. According to Industry watchers cloud will proliferate in the next
few years and touted as the next biggest thing in IT bringing with it
economies of scale and cost efficiencies of 70 to 80%. However this market
will need web acceleration platforms for achieving the penetration levels
required for monetizing the market. These services can only be delivered
through a pan-India CDN.

3.3.3 Beneficiaries of a Pan-India CDN
There are five key stakeholders in the Internet value network: Consumer, Government, ISP,
IDCs and Content providers/Portals.
A Pan-Indian CDN will benefit all of these key stakeholders in the following manner:







For Government
 Significant Foreign Exchange saving, leading to
improved GDP
 Increase in Internet penetration across India
 Improving sustainability of Internet ecosystem
































For RailTel, CDN is a natural value-add as CDN requires mini-datacenters for which the
Railways have land with Power supply and huge bandwidth. RailTel has the required bandwidth
for the CDN which could start at 1 GB minimum and would need upwards of multiple10 GB’s as
the growth increases. If we can take the proposal of CDN to the ministry we could raise funding
from USO fund as the CDN will be a multiplier for broad band penetration as illustrated above.
For India Internet Consumer
 Cheaper Internet prices
 Excellent web experience
 Access to high bandwidth applications


For Internet Service Provider
 Reduced international bandwidth cost
 Increase in profitability
 Growth in Indian market size



For Internet Data Center
 Increase in proximity to customers
 Growth in Indian market size
 Increase in high bandwidth applications



For Content Provider / Portal
 Increase in consumer satisfaction
 Increase in revenue generation opportunity
 Increase in competitiveness to international portals



Though the CDN revenue streams and use cases are many, business model focusing only on
the two main use-cases is worked out in the embedded Excel sheet of the business model and
the summary of the business projection is depicted below.

C:\Documents and
Settings\AGMRCIL\My Documents\Final railwire\ppp\Oct11th CDN Financial model (2).xls



4. Business Summary

2011 - 2012 2012 - 2013 2013 - 2014 2014 - 2015 2015 - 2016
Cumulative
Cost (Rs. Cr.)
279 878 2,155 3,898 6,023
RailTel Cumulative
Revenue (Rs. Cr.)
359 1,183 3,146 5,982 9,552



CDN Financial Model
-
2,000
4,000
6,000
8,000
10,000
12,000
2011 -
2012
2012 -
2013
2013 -
2014
2014 -
2015
2015 -
2016
Cumulative Cost (Rs. Cr.)
RailTel Cumulative
Revenue (Rs. Cr.)

Conclusion:
Earlier in an interview to NDTV, Mr Sam Pitroda has said that content should be brought closer
to the users for a richer internet experience. CDN is the best way to do it and support future
media rich applications as well. Since the technology for CDN is complex not a commodity, it is
an ideal case for PPP. If RailTel is first off the block then we can garner all the govt business
and prevent other CDN’s being built on Govt expense. For RailTel, it is presents a clear and
present business opportunity as our traditional business is shrinking.
Even though it was not planned as such, RailWire delivering CDN services will have a huge
advantage over other broadband services and together RailTel will be able to transform the
broadband scenario in India.