Space Internet:

Into the Low Earth Orbit Satellites Era

Dr. Riad Hartani, Dr. Vishal Sharma

June 2019
Space has always been the final frontier for human kind. The emergence of the Internet
has arguably been one of the most disruptive and exciting developments of the last few
decades. As we enter the 2020s, space and Internet technologies are converging. Global
technology leaders that consider the Internet evolution, in terms of adoption,
affordability, performance and reach, fundamental to their continuous growth, are
pouring tens of billions into space Internet technologies at the moment. This makes for
exciting, yet risky, times ahead!

With a track-record of contributing to the development of some of the foundational

technologies of the modern Internet – covering protocol design and standardization,
algorithm design, system architecture, and large-scale system development and
deployment – we believe we can actively help the space Internet eco-system players
forge a path that is based on sound engineering and business design principles that thus
help them minimize significantly the risk of this new enterprise.

Disruptions in the fundamentals of the architecture and design of the Internet

infrastructure and its deployment do not occur often. In fact, Internet evolution over the
last two+ decades has been mostly progressive and incremental, ever since the major
shift from using circuit switching technologies to using Internet packet-switching
technologies at scale occurred in the late 1990s. This has been a long but steady
evolution over time from Time Division Multiplexing (TDM)-based networks to a family
of packet-based networks, including Frame Relay, Asynchronous Transfer Mode (ATM)
and into Multi-protocol Label Switching (MPLS) and their various instantiations. It has
also included the shifts from circuit-based voice to IP-based voice and other multimedia
services. In parallel, various iterations of wireless technologies have been deployed,
from 2G to 4G/LTE, converging to the 5G cellular networks that are in the early stages
of deployment today. This has been complemented by the very rapid growth of the
broader ecosystem supporting the Internet evolution, in the form of large-scale data
centers and cloud networking, software operating systems, and over-the-top
applications. In fact, it is primarily this evolution of Internet connectivity models and
underlying technologies that led to the growth of the Internet eco-system, as we know it
today.

Few interesting paradigms have been emerging over the last few years, with the
potential to impact the Internet infrastructure and the design and deployment of
Internet-based services. These paradigms will have significant consequences for
content delivery models, cloud networks, distributed computing and the economics of
over-the-top application rollouts. These include aspects such as blockchain and
decentralized Internet technologies, quantum communications and low earth orbit
(LEO) satellite communication networks. This paper focuses specifically on LEO
networks, and mostly addresses the challenges to overcome to ensure their potential
success. It provides a glimpse of how the technologies, protocols, standards and
mechanisms developed for terrestrial and wireless Internet networks can be leveraged

Page 2 of 9
to speed up deployments of LEO-based communication networks over the next few
years.

Simply put, LEO networks are satellite-based constellations that orbit the earth at
altitudes of 1200 miles or less. These constellations have existed for a while, and
numerous ones have been launched in the past, with the Iridium network from the late
90s being the most well known. The novelty is that these recent networks launches are
focused on enabling global-scale Internet connectivity, bringing in a new era of space-
based Internet technologies.

Indeed, all the major internet/cloud providers are working on various aspects of such
deployments, including Amazon, Google, Facebook, Microsoft as well as large-scale
space technology players such as Virgin and SpaceX and technology investors such as
Softbank. Also active are some of the existing satellite communication providers already
present in GEO (Geostationary Orbit) and MEO (Medium Earth Orbit), as well as venture
capital-backed startups, and government funded consortiums in China, Japan, Korea,
Europe and North America. Most constellations launches are being planned in the 2020-
2025 timeframe, with tens of billions of dollars being invested. At the same time, this is
still a high-risk initiative given the technical and business challenges that need to be
solved. As such, this is a high-risk high-return equation and only time will tell on how it
will impact the Internet evolution, global competitiveness, and Internet geo-politics
over the next decade.

The new LEO satellite networks being designed at the moment bring in a whole new set
of opportunities, taking advantage of the potential low latency, broad reach and high
capacity of such networks. The scale of investments going into such initiatives,
primarily from the private sector, adds a significant advantage to their potential. These
LEO space networks are being designed with the intention of leveraging the
mechanisms designed for terrestrial networks such as those for routing, switching,
Quality of Service (QoS), resources management, software-defined network (SDN)
control, virtual network functions (VNF) orchestration, and cyber-security. Yet, a lot of
these mechanisms are far from optimal given the characteristics of LEO space networks,
in terms of mobility, terrestrial-to-space wireless link management, and space-to-space
wireless link connectivity. In some cases, these mechanisms need to be highly adapted,
and, in other cases, fully redesigned. In fact, these LEO space networks are in early
stages of taking advantage of the internet/wireless networking mechanisms that have
been developed, deployed, and, in some cases, abandoned over the last 20+ years.

There is an opportunity to leverage state-of-the-art Internet designs and evolving those

optimally to enable the deployment of this new generation of space networks. In this
paper, we present a non-exhaustive review of some of the key aspects that need to be
addressed, both in the service offerings and the technology development fronts. For
each one of the dimensions considered, we list some of the aspects that require further

Page 3 of 9
work, and that could take advantage of the available Internet mechanisms and
standards, for the specific context of LEO networks.

Adapting Internet services and customer application offerings

o Adaptation of the Service Level Agreements (SLAs) and definitions of Key

Performance Indicators (KPIs) is required because the IP-based service SLAs
have primarily been defined with terrestrial networks in mind. Adapting
these SLAs to LEO satellite networks is a must, as this has a direct impact on
traffic management/engineering solutions that need to be put in place on the
satellites, on the coordination between terrestrial and satellites networks,
and on load balancing across space segments, among other things, and this is
needed on both the data and control planes.
o Various services targeted by LEO networks are focused on well-known
internet services offered by existing terrestrial/wireless networks, such as
business-centric Layer 2 and 3 services, and Virtual Private Network Services
(VPNs). There are new opportunities for services that would leverage the
new cost structure of LEO network deployments in terms of coverage,
bandwidth and latency, as well as the potential new Layer 3 routing
topologies that they bring such as global routing with a reduced number of
autonomous systems and new peering/transit models.
o The analysis of new services includes aspects that would piggyback on the
deployment of distributed mobile edge computing solutions with highly
distributed data centers and clouds, content delivery networks, and public
safety networks.
o There is an opportunity to revisit the technologies and deployment models of
peer-to-peer (P2P) based networks, and to leverage the characteristics of
LEO networks for bringing in new topology models for designing and hosting
peers’ hierarchies and topologies. It would also be interesting to analyze how
this would complement the ongoing block chain-lead initiatives for
incentivizing the use of P2P networks at scale and the evolution of
distributed file systems and associated data distribution techniques.
o The emergence of LEO networks opens up new opportunities for the
deployment of global Mobile Virtual Network Operations (MVNOs) given the
large-scale geographical nature of LEO networks and their underlying
economics.
o Multi-media service delivery, including the delivery of voice and video
services directly over LEO networks, calls for a rethink of the mechanisms
designed for LTE networks, such as those in the IP multi-media systems
(IMS), in roaming models, and in inter-connection architectures.
o The global nature of LEO networks, and the new interconnection models they
provide to connect with terrestrial wireline, wireless, submarine and cloud
networks, has the potential to significantly change the dynamics of rolling out
high-speed broadband in rural regions, particularly in the developing world.
Page 4 of 9
 It is as such, a clear opportunity for many countries to explore ways of
speeding up the implementation of their digital infrastructure strategies.

Harmonizing Internet Routing and Signaling Protocol Design

o Adaptation of the Internet Gateway Protocols (IGP) and potentially the

Border Gateways Protocol (BGP) for global routing to accommodate wireless
links with very specific characteristics (this includes for e.g., satellite-to-
satellite links, ground-to-space fixed wireless links, and mobile users-to-
space wireless links), directly impacts layer 2 and 3 topology information
dissemination, path computation, the mapping of demands to paths and load
balancing over paths.
o Incorporating wireless link characteristics into QoS metrics and using these
for traffic routing, for the earth-to-satellite links as well as satellite-to-
satellite links.
o As LEO networks get progressively deployed, and given the challenges in
addressing their specific predictability, reliability and availability
characteristics (e.g., weather and capacity limitations), there is a clear need to
build network control models that leverage the potential complementarity of
other technologies, including 4G/5G, microwave backhaul, and submarine
networks to ensure end-to-end SLAs are satisfied with the right economics.
o The handover models typically deployed in 3GPP-based 4G/5G networks
would need to be adapted for the cases of mobile and high velocity satellites,
as these call for different mechanisms to ensure data continuity with
appropriate quality-of-experience requirements. This is even more so when
dealing with dual network-element mobility scenarios, which include mobile
user terminals and mobile satellites.
o The data-path connectivity protocols, centered on the various Layer 2/3
IP/MPLS mechanisms, as well as their corresponding control planes, would
benefit from potential adaptations that would make them suitable for
carrying payloads over multi-hop space segments.

Evolving QoS and Traffic Management Mechanisms

o Data path resources management: Building on top of existing Transmission

Control Protocol (TCP) and User Datagram Protocol (UDP) (and their various
alternatives developed for existing GEO satellite networks), while taking into
account the high-latency, high-loss wireless links, and the need for
compression, QoS signaling, and so forth means that these need to be adapted
to LEO networks, as their characteristics are very different from those of
standard GEO space networks.
o The design and dimensioning of oversubscription models over LEO space
segments has to be fundamentally adapted relative to the models in use in
terrestrial networks. This is due to the specificity of traffic models in terms of

Page 5 of 9
 network capacity demands, the variability of the physical/logical space and
ground-to-space topologies, and the mechanisms available on the data and
control paths for short/mid-term traffic/resources management
o For several potential LEO-based service applications, the data-path adaptive
and reactive optimization mechanisms that are used to optimize performance
and efficiency in 4G/5G packet core networks would benefit from
adaptations that take into account the unique characteristics of multi-hop
space networks.

Reconciling NFV, SDN and Operational Systems

o LEO space networks are global; hence, there is a need to consider ways of
deploying SDN and centralized/distributed network controllers and
orchestrators in a way that satisfies the latency, QoS, and security
requirements of these networks, while optimizing the cost of deployment and
operations.
o This is also the case for the deployment of Operations Support Systems (OSS)
and Business Support Systems (BSS) data models, for data ingestion and
processing, and for effecting corresponding actions for the management of
the network and the orchestration of services.
o As terrestrial networks evolve towards NFV models, there is a clear need to
leverage these concepts for the design of LEO satellites, for some of the data
path functions (e.g. routing, QoS, and service adaptation), while considering
the constraints of satellite design (for example, need for simpler operating
systems, efficient link/data layers, low power, and easy upgradability).
o The interaction between the VNFs and the SDN controllers and orchestrators
would have to be revisited to take into account the management
requirements of satellites in which dynamic configurability over global
topologies is key.
o Interaction of the higher OSS/BSS layers with the network layer via
orchestrators across domains is another area that will need revisiting. These
interaction mechanisms were developed primarily for terrestrial networks
and would need to be adapted to LEO networks. For instance, the various
messaging / API models would need to include a different set of information
models and new messaging to map the requirements of the data and control
paths.

Tailoring State-of-the-Art Cyber-Security Mechanisms

o Cyber-security for data and control paths would require rethinking to

accommodate the characteristics of space segments given the constrained
functionalities on the satellites. E.g., the ability to process, detect and protect
their compute and network resources vis-a-vis standard routers in terrestrial
networks, with significantly more powerful capabilities. This is due to the

Page 6 of 9
 design constraints applicable to space satellites in the areas of power, space,
cost, upgradability, and support.
o Aspects that relate to data residency for all manner of network control and
management functions, including fault management, performance
management, and billing would need to be architected very differently given
the global nature of LEO networks, and the increasingly local nature of data
residency on a per country/region basis.
o Opportunity to leverage new key distribution models, including those of
quantum keys distribution (QKD) from satellites in space to enhance end-to-
end encryption.

Growing Next-Generation IoT Networks Leveraging LEO Connectivity

o The recent evolution of IoT connectivity services defined in 3GPP, Low Power
Wide Area Networks (LPWAN) and others could take advantage of LEO
connectivity characteristics for enhancing the business cases for
deployments, and for offering different types of IoT services in rural/remote
areas.
o The complementarity between terrestrial IoT networks and space-based
connectivity networks provides a new framework for global service
providers to deploy retail/wholesale IoT services at scale.
o The IoT gateways and backend architectures in use today would benefit from
interfacing with the control and management plane of LEO networks to
provide an end-to-end IoT service deployment and optimize cost and
functionality.

Fashioning Technology Standards and Regulations for LEO Networks

o Standards bodies have already been addressing the various regulations

required for the deployment of large-scale LEO networks. Various open areas
remain under consideration, however, given the global nature of LEO
networks, their impact on local regulations on a per-country basis, and the
various national/regional licensing schemes that need to be adapted.
o Standards have also been addressing aspects that relate to the management
of interference risks with GEO/MEO networks as well as the various
terrestrial networks. This is likely to be an active area of work as the
deployments progress.

o The number of operational satellites in orbit is expected to increase ten-fold

over the next decade (from ~2,000 to ~20,0000). This will, over time, lead to
the presence of many zombie satellites, used rocket stages, and random
debris fragments orbiting the planet. This will in turn call for specific
regulations and technologies to manage space debris across the globe
o The strategic importance of the Internet to the economic and political
interests of a wide range of public and private sector stakeholders could
Page 7 of 9
 make for contentious, complicated, and high-stakes debates around issues of
fair access to orbital resources. Along with this, ongoing space militarization
initiatives are likely to make this even more contentious, requiring an
evolution of the regulatory framework over time.

Major technology and financial investments are going into the deployment of LEO
networks at present. There has rarely been so much of a push to experiment, design and
launch breakthrough highly-complex Internet technologies at scale. It is a race between
leading technology players, governments, and policy makers to get various aspects of
this area ironed out that is likely to accentuate over the next few years, given how
strategic the Internet infrastructure is for the development of nations and the
competitiveness of technology corporations.

Yet, major challenges remain to be overcome. These include both technical and business
challenges. The intersection of space and Internet technologies is still in its early phases,
with lot of learnings from both sides aiming to enhance the joint value proposition.

Leveraging our own experiences developing Internet-based protocols and deploying

Internet-based services at scale over the last two decades, and building on our own
work in the design of space Internet networks, we have presented a perspective on the
issues that are crucial in emerging LEO-based satellite networks, and how they can be
addressed by taking advantage of the Internet design experience that we possess.

The next few years will likely witness a rapid evolution of these technologies, with a
possible significant impact on how Internet services evolve. A potentially high-risk high-
return equation, where there will likely be few winners and lots of losers. These are
certainly exhilarating times for Internet evolution, in a world where the Internet is, and
will continue to be, the cornerstone of the development of nations.

About the Authors

Dr. Riad Hartani

http://www.linkedin.com/in/dhartani

[email protected] Skype: riadhartani

Advancing state of the art in technology and bringing innovation to global markets has
been Riad’s ultimate passion. Riad has spent the last two decades contributing to the
development of Internet and Artificial Intelligence technologies, mostly out of the
Silicon Valley as a hub, building multiple technology startups, advising on technology
investments and rolling out innovative technologies in most regions of the world.
Having lived and worked primarily in the USA, but also in Canada, Hong Kong, Japan,
Korea and France, he has most recently founded a global technology and investments

Page 8 of 9
advisory firm based out of San Francisco, Vancouver, Tokyo and Singapore, and on the
advisory board of various technology startups and conglomerates.

Dr. Vishal Sharma

http://www.linkedin.com/in/vishalsharma

[email protected] Skype: vsharma87

Vishal has been active in the Silicon Valley eco-system for nearly two decades,
consulting to startups and Fortune 1000 companies alike in the networking and
telecommunications space, including hardware/software vendors and wireless and
wireline operators, with clients spanning four continents (US, Europe, Asia, and
Australia). He has taught and conducted research at the prestigious Indian Institute of
Technology Bombay (2004-2007), and lived and worked on both coasts of the US and in
India. In addition, Vishal advises attorneys from leading laws firms, e.g. those in the
AMLaw 100, and key investment houses on intellectual property matters. With his
activity in industry, academia, and consulting, Vishal has been a contributor to, and
worked closely with, the development and realization of the core foundational
technologies of the modern Internet – from Diffserv and MPLS to SDN/NFV, cloud
networking, and network slicing, which are key to the growth of the space-Internet.

Page 9 of 9