Crossing the
Chasm: Enhancing Cryptocurrency
Adoption by
Leveraging a Transaction and
Trading Layer
for the Points Economy and
Other Digital
Assets
-
Whitepaper
- Version 2.2
November 30,
2018
Al Burgio
Founder of
DigitalBits Foundation
aburgio@digitalbits.io
Abstract. Blockchain
technology is promoted as one of the great tech-nological advances of our time
and considered as a solution to many of the technical problems faced by
industries in sectors such as nance, automobile, and retail. Despite growing
attention and utilization, mass adoptions of cryptocurrencies has not happened
yet. So how do we cross the chasm from the vision phase to the actual use
phase? To do so, block-chain technology has to target a pre-existing legacy
market that already posses billions of user accounts globally in a digital
asset category; Not to create a competing tokenize digital asset, but instead
to transform these legacy digital assets into tokens on a public blockchain
with key functionality that bene ts both consumers and the enterprise-issuers
of these digital assets.
This whitepaper
lls the gap in the state of the art by presenting the Digi-talBits blockchain
powered infrastructure that builds a bridge facilitating the implementation of
new technologies to support and enhance our ev-ery day life interactions as
well as foster blockchain mass adoption. The DigitalBits blockchain allows for
easy asset-tokenization using a transac-tion and trading layer for the point
economy. We present a loyalty- and reward points focused running case, detail
the advantages of the system, outline the requirements and goals, as well as
the architecture of the Dig-italBits network and ecosystem. In addition, we
present the on-boarding process of digital assets and the asset tokenization
which is an indispens-able functionality of our platform. Furthermore, we
introduce the novel idea of the token name certi cation service (TNCS) that
prevents ma-licious network entities from issuing and distributing illegitimate
tokens of assets that they are not associated with. Finally, we present the XDB
token value proposition and the surrounding token economy ecosystem that fuels
the platform.
Keywords: Loyalty Programs,
Digital Assets, Tokenization, Payment, Remit-tance, Blockchain, Smart Contract,
Point Economy, DigitalBits, Liqui ed Assets
2
1 Introduction
The vision for
cryptocurrencies has been embraced by many people around the world. In 2017,
the world saw a surge in Initial Coin O erings (ICOs) with associated
whitepapers that has driven tremendous enthusiasm for the futurist
possibilities of blockchain technology and utility tokens [3][43]. However, as the end of 2018
is near, many people are beginning to wonder: When will many of these
cryptocurrencies and utility tokens begin to gain wide real world use and
adoption? How do we cross the chasm from the vision phase to the actual use
phase? To do so, blockchain technology has to target a pre-existing legacy
market that already posses billions of user accounts globally in a digital
asset category; Not to create a competing tokenize digital asset, but instead
to transform these legacy digital assets into tokens on a public blockchain
with key functionality that bene ts both consumers and the enterprise-issuers of
these digital assets.
Loyalty
and rewards points (LRPs) presents a unique opportunity as the rst digital
asset category to leverage the DigitalBits blockchain and help drive mass
growth of cryptocurrency adoption. Such programs are established means to im-prove
customer engagement and brand awareness. Needless to say, due to their
recognized and tremendous potential, these membership programs spread across
several industries, as for example, travel, retail, nancial services, and so
on. A successful loyalty and reward program ensures that a customer keeps
returning to a speci c brand in order to make purchases and subsequently earn
reward points thereby building loyalty and retaining customers over the years.
At the same time, companies that e ectively create, launch and run loyalty
programs underline their commitment to long term relationships with their
customers [29]. A 2014 report suggests
that 91% of companies employ some form of customer engagement or loyalty
program [18], resulting in 3.8 billion
individual loyalty program memberships just in the United States [12]. The average U.S. house-hold
participates in 29 di erent loyalty programs [11], whereas in the UK the average customer is a
member of more than 14 di erent loyalty programs [14]. According to a report from 2017, the estimated
overall corporate liability for loyalty points exceeded $100 billion, thereby
representing a new unsurpassed high and an enormous business potential [6].
However,
despite the widespread availability of these programs, only a small fraction of
their potential is used. Consumers are generally discouraged by un-necessary
barriers to accumulate and redeem their points and by changes to pro-gram rules
or rewards. Hence, users get easily frustrated and millions of points are
unredeemed. Furthermore, the lack of transferability and portability of most
LRP programs a ects their perceived value by customers, since they are unable
to transfer or trade those points at any time for other points they desire. Of
the $48 billion in total perceived value of points and miles issued in 2010 in
the U.S., about $16 billion worth of LRPs were left unused [20]. It is important to note,
that it is not just the customers putting up with the downsides of current
state of the art solutions. Loyalty program providers have high entry barriers
to set up their own reward program. Once they have created their own program,
providers su er from high costs of maintenance. Moreover, even if they wish
3
to achieve a certain level of compatibility with another membership
program, the underlying infrastructure is often incompatible. Hence, a more
holistic and interoperable solution is required.
Blockchain technology, also referred to as
distributed ledger system, is most noticeably known for providing the
foundation of the peer-to-peer (P2P) cryp-tocurrency and payment system Bitcoin
[34], but nowadays there a various dif-ferent
platforms out there, e.g., [25][39][48]. In the context of LRP
programs, blockchains have the potential to enable interoperable and holistic
platforms for the next generation of loyalty programs. Blockchain-based LRPs
can be eas-ily transferred and redeemed - even across other businesses or
services. In ad-dition, LRPs become interchangeable and can be exchanged for
other loyalty points or even at currencies. Finally, a blockchain-based
solution signi cantly reduce development, integration, reconciliation, and
security costs for the pro-gram providers.
In recent times, nearly a dozen companies
announced their intent to launch blockchain based LRP programs to encourage
customer engagement [38]. How-ever, most of them merely translate the
legacy LRP programs into a blockchain based solution with separated data silos
and a lack of interoperability. Others are built on rst generation blockchains,
such as Bitcoin and therefore, su er tech-nological downsides such as high
transaction fees, limited scripting languages or missing scalability. In
contrast, we envision a solution that is not only limited to ease the creation,
transfer and exchange of LRPs on a blockchain but also o er an easy-to-use
platform which allows the tokenization of all kinds of assets at the mere click
of a button. This, as a result, fosters the mass market adoption of blockchain
technology.
This whitepaper
addresses the detected gap by introducing the DigitalBits blockchain, thereby
answering the question of how to enable easy asset-tokenization using a
transaction and trading layer for the point economy? In order to answer this
question with a separation of concerns, we pose the following sub-questions:
What are the goals and requirements of such a system? What is the architecture
of the DigitalBits asset tokenization platform? What are the system-engagement
processes for the stakeholders? What is the long term vision of DigitalBits?
The remainder of this paper is structured as
follows: Section 2 introduces background information, a running case as well as related
work. Section 3 fo-cuses on de ning functional as well as quality goals and the
requirements of the DigitalBits system, together with the positioning of
stakeholders. Next, Section 4 analyses and outlines the resulting system
architecture that we derive from the requirements. Section 5 expands
on the system engagement processes for the stakeholders. Finally, Section 6 concludes
this work.
2 Technical Background and Running Case
4
paper. Next, the state of the art on blockchain-based LRP solutions is
presented in Section 2.3.
2.1 Loyalty and Reward Points
Programs
LRPs are a
means adopted by companies and brands to engage the consumer repeatedly,
develop a long standing relationship with the consumer and encour-age the
consumer to choose a particular set of products o ered by a brand (or a group
of brands) over other brands o ering similar products. In the current
landscape, there are two types of LRP programs. One is a single business
pro-gram in which the points are issued by the same business, such as the Advantage
program of American Airlines and Starbucks Rewards or Starbucks. The second
type is a multi-business program, in which the points issuer is a third party.
For example, LoyaltyOne and its Air Miles program or Payback in Germany.
A variety of companies and brands have applied
these LRP programs. Studies show that roughly 80% of Americans belong to some
type of reward program [47]. The average US household belongs to 29 loyalty programs [5] and 71%
of loyalty program members say that they are open to joining more programs [42]. There
were 3.8 billion individual loyalty memberships in the US in 2017 and 175
million loyalty memberships in Canada in 2016 [19]. These
gures re ect a 15% growth in the U.S. and 35% growth in Canada since 2014.
2.2 Running Case
HotelBrand is a ctional ourishing chain of hotels with a global
presence. Sim-ilar to other hotel chains, HotelBrand has an ongoing loyalty
program with roughly 30 million customers and approximately 40 billion points
in circulation. However, owing to the ercely competitive hospitality market,
HotelBrand is currently languishing at the 10th place. Moreover, it is facing
heavy competition from a variety of innovative startups that are primarily
focused on a sharing economy. In order to enhance their market share in the
upcoming years, Hotel-Brand has identi ed that a key solution to strive in the
market should be the overhauling of their LRP program with the objective to
make it more customer-centric. A customer-centric LRP program has features such
as ease of sign-up and use, and more importantly, facilitate high liquidity,
i.e., allow the customer to easily collect, monitor, share and utilize the
accumulated points. High liq-uidity results in higher customer engagement and
an increase of the perceived value of the LRP program. This encourages new
customers, in addition to their current clientele, to avail their services over
their competitors. HotelBrand is naturally looking to expand and increase their
worldwide reach and aspires to be the market leader. Therefore, they are
interested in creating a platform that can dynamically and easily scale up with
their ever ambitious goals. However, HotelBrand realizes that the following
factors could the revamp of their LRP program:
{ Infrastructure (including maintenance, scale
& security): High liquidity re-sults in a signi cant increase in the number
of transactions that need to
5
be performed on a daily basis. They have to further take into account
that there could be spikes in the amount of transactions during the day, based
on consumer behavior patterns. A large part of the transactions are expected to
be performed online. Their infrastructure has to be upgraded in order to cater
these demands. A cloud-based system is appealing, however, in addition to the
huge cost, this requires in-house expertise in maintenance, dynamic scaling,
and security.
{ Consumer
experience: In order to accommodate the current trends to operate via mobile
Apps, an App rich in features and appeal, has to be developed. The major
concern though, is of consumer engagement. Although Apps are easy to install,
the large number of them on a consumer's phone tends to su er the same issues
as that of multiple loyalty cards with a consumer { a lack of enticement for
the customer. The App might be yet another App on the consumers' phone and thus
not used frequently.
{ Costs & Time: Regardless of prioritizing and choosing an in-house
or a cloud-based version, the costs of creating and then maintaining such a
program would be huge. Moreover, the launch of an overhauled LRP program also
requires a lot of development time.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5. Buy Product
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1. Obtain XDB
|
|
|
|
|
|
|
|
|
9. Trade XDB/
|
||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
HotelBrand-tokens
|
||||
|
|
|
|
|
|
|
|
(for
minimum
|
|
|
|
|
|
6. Issue
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
balance)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
HotelBrand-tokens
|
|
|
|
|
|
|
|
|
|
|
DigitalBits
|
|
|
|
2. Request
|
|
|
h
|
|
|
7. Give
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
HotelBrand-tokens
|
|
|
|
|
|
|
|
|
|
|
|
||
|
|
|
Network
|
|
|
|
|
|
|
|
|
HotelBrand-tokens
|
|
|
|
|
|
|
|
||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4. Issue
|
HotelBrand
|
8. Get
|
|
Consumers
|
|
||||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
||
3. Create HotelBrand-tokens
|
HotelBrand-tokens
|
|
|
|
|
|
|
|
|
|
|
|
|
||||||||
|
|
|
|
|
|
upgrades /
gifts
|
|
|
|
10. Sell for Fiat
|
|||||||||||
(after checking min-balance
|
|
|
|
|
|
|
|
|
|
|
|||||||||||
|
|
requirements)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
||||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Fig. 1: Use-case
of HotelBrand creating and using HotelBrand-Tokens on the DigitalBits network:
1.) HotelBrand obtains the minimum XDB-tokens required for participation in the
DigitalBits network; 2.) Create HotelBrand-Tokens; 3.) HotelBrand-Tokens being
created by DigitalBits network; 4.) HotelBrand-Tokens issued to HotelBrand; 5.)
Consumer-Tom booking a room o ered by HotelBrand; 6.) HotelBrand issuing the
corresponding HotelBrand-Tokens; 7.-8.) Consumer-Tom choosing to trade the
tokens for a free room at HotelBrand; 9.) Consumer-Tom choosing to trade some
of her HotelBrand-Tokens or XDB-tokens with other consumers; and 10.)
Consumer-Tom choosing to sell her HotelBrand-Tokens or XDB-tokens for at
currency or other cryptocurrency.
6
HotelBrand
is monitoring the hype behind blockchain, championed as the solution to many
prevalent technological problems. Nonetheless, early prototypes investigated by
their teams revealed that blockchain based solutions have pitfalls such as huge
transaction fees and time. Recently, HotelBrand is intrigued by the blockchain
based solution o ered by DigitalBits, especially the manner in which it
overcomes many of the limitations of the earlier generations of blockchain.
A key feature provided by DigitalBits is that
the o ered tokens have high liquidity, in alignment with HotelBrand's
customer-centric focus. They appre-ciate that DigitalBits allows them to create
a new LRP program at virtually no cost1. They also realize that the system can
dynamically scale in order to handle their global consumers. Moreover, they
realize that they can start o by using the basic app provided by DigitalBits
and later invest on developing a brand-speci c app on top of the basic app by
making use of its APIs and SDK. The only cost they incur is a very low
transaction fee2, a cost that is required to ensure the
integrity of the transaction. The fee is used to sustain the ecosystem.
HotelBrand, therefore, uses the money saved to pass on more bene ts to the
con-sumers and also increase their marketing and consumer engagement
activities, and more importantly, focuses on their core product.
As depicted in Figure 1, HotelBrand
obtains the minimum XDB-tokens . They then proceed to create a new token,
HotelBrand-Tokens for their brand and specify the maximum tokens that they
would like to have. When Consumer-Tom books a room at one of their hotels, they
transfer some tokens to Consumer-Tom based on their own internal algorithm.
Consumer-Tom is immediately able to see the increase in HotelBrand-Tokens in
her DigitalBits-enabled App which she regularly uses with other brands.
Moreover, she is glad not having to install yet another app or ll up a form for
another loyalty card. At a later point in time, Consumer-Tom would like to use
some of her HotelBrand-Tokens to avail a free room from HotelBrand and also
convert the rest into tokens to avail a discount on her next ight. Consumer-Tom
also periodically chooses to convert some of her tokens into XDB-tokens to bene
t from the growing value of the DigitalBits network.
2.3 Related Work
Currently, most
LRP programs exist in silos, i.e., belong to a certain brand (or group of
brands) and cannot be easily transferred or traded with. Since each of them
exists in custom-built infrastructure, it is practically impossible to merge or
facilitate real-time interworking among two di erent LRP program, even when the
brands wish to do so. Moreover, a peer-to-peer based platform that facilitates
the transfer and trade between di erent LRP programs has not historically
existed. This problem does not merely arise from businesses desiring
1 they will have to obtain
the minimum XDB-tokens (DigitalBits native token) re-quired to maintain the
account, which is currently set at 10 XDB-tokens
2 currently set at 1
XDB-tokens for 100,000 transactions and can be changed only with consensus
7
to hold on to their customers, but also that the barrier to facilitate
such a system is high. Businesses that are keen on making their platform more
open face hurdles of scale, cost and time. As a result, customers are forced to
carry di erent cards for each of the loyalty programs or install a new app for
each of them on their phone. Apart from the issue of overloading a
wallet/phone, one also has to weigh in the chances of a customer remembering if
he or she is a member of the loyalty program or not. The entry barrier for new
players is also high when we consider the high costs, time to launch,
maintenance/scaling e orts, security and the need for skilled personnel.
|
DigitalBits
|
Stellar
|
Ethereum
|
|
|
|
|
|
|
Blocktime (w/
|
2-5s
|
2-5s
|
5m to 1h+
|
|
Confirmations)
|
||||
|
|
|
||
# of Confirmations
|
1
|
1
|
30
|
|
Processing Method
|
Validation
|
Validation
|
Confirmations via
|
|
Mining / PoW
|
||||
|
|
|
||
Transaction Costs
|
Very Low
|
Very Low
|
High
|
|
Multi-Asset
|
Built-in
|
Built-in
|
Custom App via
|
|
Smart Contracts
|
||||
|
|
|
||
Distributed Exchange
|
Built-in
|
Built-in
|
Custom App via
|
|
Smart Contracts
|
||||
|
|
|
||
Compliance Mechanism
|
Built-in via
|
Built-in
|
No
|
|
compliance
|
||||
|
server
|
|
|
|
Inflation
|
No
|
Yes
|
Yes
|
|
Mining
|
Pre-mining
|
Pre-mining
|
PoW & PoS
|
|
Certified Token Issuer
|
Yes
|
No
|
No
|
|
TNCS
|
Yes
|
No
|
No
|
|
Automatic Algorithmic
|
Yes
|
No
|
Yes
|
|
Native Token Distribution
|
||||
|
|
|
||
|
|
|
|
Fig. 2: Platform Comparison of DigitalBits,
Stellar and Ethereum.
Blockchain
has the potential to ease many of the aforementioned problems. However, many of
the blockchain based solutions are in fact Decentralized Apps (Dapps) built on
Ethereum [48]. This subjects them to
issues such as high trans-
8
action costs,
veri cation time and low transaction processing rate. Moreover, since ETH3 continue to be created every day, they are also subjected to in a-tion.
Solutions such as Qiibee [40], have
a huge entry cost by binding the value of tokens to their native currency,
similar to how certain monetary systems used the gold standard [17]. A key issue with Dapps is
that they can support only one native token, similar to many of the legacy LRP
programs. Therefore, all participating brands have the same native coin with
the same features and con-nement to a silo. Alternatively, blockchain platforms
such as Ethereum4 [48], Waves5 [23] and
Stellar6 [31][45] allow
users to create their own tokens. Digi-talBits, a fork of Stellar, shares many
of the bene ts of Stellar, but di ers in key aspects such as not subjecting its
tokens to in ation, and is developing a token name certi cation service and an
automatic algorithmic token distribu-tion. Figure 2 presents a high level comparison between DigitalBits, Stellar and Ethereum.
Waves, which relies on Proof-of-Stake (PoS), provides similar features like
Ethereum and is comparable in performance to Ethereum. DigitalBits,
out-performs both Ethereum and Waves in terms of the number of transactions per
second (10,000 transactions per second [7]) and the time to verify transactions
(approximately 3-5 seconds [45]).
Additionally, it provides features such as legal compliance, multi-asset
support and a distributed exchange.
3 Goal Modeling and Requirement Engineering
The previous
section introduced the DigitalBits running case, that we now use to deduce the
functional- and non-functional requirements that our system must adhere to, in
order to address the drawbacks and limitations of current LRP programs. To
systematically capture the necessary requirements, we use one part an
Agent-Oriented Modelling (AOM) methodology [46], i.e., goal models. In the context of system
development and software engineering, good requirements follow certain
characteristics, e.g., requirements address one issue only and are completely
speci ed without missing information and they must be atomic as well as without
conjunctions [13][22][36]. In addition, the
requirements have to be consistent and do not contradict itself, or in
correlation with other requirements.
The
AOM methodology [46] is
a socio-technical requirements-engineering ap-proach. It is used to model
complex systems that consist of humans, hardware agents, and software agents in
a changing environment (also referred to as dis-tributed socio-technical
system). Essentially, an AOM goal model presents the software and the tasks it
can perform from an agent-oriented view. It therefore enables both, technical
and non-technical stakeholders, to capture and under-stand the functional and
non-functional requirements of a complex system. As depicted in Figure 3, Role, Goal and Quality Goal
are the three key elements of an AOM goal model. Roles of involved entities,
i.e., stakeholders, are represented
3 Ethereums native token
9
in the form of
sticky men. Sticky men represent both human entities as well as entities such
as apps on their phones that act on behalf of the human entity. The functional requirements
are depicted as parallelograms and refer to goals of the modeled software
system. The non-functional requirements are depicted as clouds and refer to quality
goals of the modeled software system. The AOM goal model follows a tree-like
hierarchy. The root value proposition, i.e., the main goal, of the modeled
system is at the top. This root goal is decomposed into sub-goals where each
sub-goal represents an aspect for achieving its parent goal
[30]. These sub-goals are further
decomposed into multi-layered sub-goals until the lowest atomic level is
reached. Roles and quality goals may be assigned to either the key goal or the
sub-goals and are inherited by all the lower-level goals.
Role Goal Quality Goal
Fig. 3: Selection of AOM notation elements.
This section presents our AOM goal model to de
ne the requirements of the DigitalBits system thereby highlighting the
functional and non-functional goals of the platform. In Section 3.1, we rst
address the top-level goal model comprising the value proposition and the
directly linked functional and quality goals along with the relevant
stakeholders. Later, Section 3.2 focuses on the lower-level re nements of the
model.
3.1 Top-Level Goal Model
Figure 4 presents the top-level AOM goal model of the system using the mod-eling
method described above. The main value proposition is to provide easy
asset-tokenization using a transaction and trading layer for the points economy
and other digital assets, thereby representing the root of the goal model. The
complex main value proposition is split into four sub-goals representing the
four main functionalities of the DigitalBits system. The re ning functional
goals are enact plugins and Apps, tokenize assets, trade assets and pay and
remit. Fur-thermore, relevant stakeholders pertaining to these high-level
functional goals are also highlighted in Figure 4. These
high-level functional goals are re ned further in subsequent sections.
Besides
the four sub-goals of the top-level AOM goal model, we further iden-tify
thirteen quality goals. Nine of these quality goals belong to the main value
10
proposition
and are inherited by all its sub-goals. The remaining quality goals belong to a
sub-set of the sub-goals and are inherited by its child-goals.
|
|
Easy to use
|
Interoperable
|
|
|
|
|
||
|
Scalable
|
|
|
|
|
|
Decentralized
|
|
|
Cost efficient
|
Integrable
|
Fast
|
||
Flexible
|
|
|||
|
|
|
Secure
|
|
3rd Party
|
Provide easy asset-tokenization
|
|
||
|
using a transaction and
|
|
||
ServiceProvider
|
|
trading layer for the points
|
|
|
|
economy and other digital assets
|
Multi-hop
|
||
|
|
|
||
|
Compliant
|
|
Convenient
|
|
|
|
Automated
|
||
Enact plugins
|
Tokenize assets
|
Trade assets
|
Pay and remit
|
|
and Apps
|
||||
|
|
|
Loyalty
Program Provider
other
issuers
Consumer
Fig. 4: Value proposition and rst level re nement of the DigitalBits goal
model.
A scalable and decentralized system design is
necessary for the envisioned platform to handle the large number of LRP program,
consumers and the re-sulting transactions. The LRP program providers have a
large consumer base with a worldwide presence and therefore the platform must
be able to handle the veri cation of transactions in a fast and secure manner.
A secure service provision is crucial in terms of operational security, e.g.,
protect provider as well as consumer accounts and personal data from
unauthorized access, secure data transfer within the system between entities or
preventing data and information leaks.
The DigitalBits network is expected to support a
large variety of LRP pro-grams (and other digital assets) and must therefore be
agnostic to the speci c needs of an individual provider. The ecosystem is also
expected to include a large number of plugins and apps developed by third party
developers and therefore must be designed to be interoperable and integrable in
terms of hardware and software design. It is also crucial to interoperate at
runtime with information systems supporting other business functions.
Since the users of the platform are consumers
who are not necessarily tech-savvy, it is important for the platform to be
easy-to-use. The consumers must be able to interact with the platform, the
plugins and apps easily in order to perform actions such as register for
points, collect points, get an overview of their points and trade/sell points.
According to [36], easy usability also includes the support of
proper error avoidance in order to \anticipate and prevent common errors that
occur during a collaboration con guration".
11
A reliable enactment of all transactions and
trades is mandatory to facilitate the previous goals. The system must also
strive to provide these bene ts in a reliable and cost e cient manner. Last,
but not the least, the system must be designed in a exible manner in order to
facilitate a highly dynamic process that involves the enactment of diverse
activities, the participation of diverse partners, and the exchange of diverse
data [35].
3rd
Party
ServiceProvider
Enact plugins and Apps
(De)-register
Prepare enactment
Execute
Terminate
Interact
Send Tokens
Receive Tokens
Loyalty
Program Provider
other
issuers
Consumer
Fig. 5: Lower-level re nements of the Enact
plugins and Apps functional goal.
3.2 Re ned Goal Model
The lower-level
re nements of the four sub-goals are detailed below. First, Fig-ure 5 presents the enact plugins
and apps goal. Applications and plugins have to be registered, prepared for
enactment, executed and terminated. Moreover, they have to interact with
various entities of the ecosystem depending on the use case, e.g., receiving
and sending tokens, or other kind of data. Since nowadays most blockchains o er
Turing-complete smart contract languages, the variety of applications and
plugins in our ecosystem is quite vast. In the context of Digi-talBits, a
variety of such services is developed and o ered by 3rd party providers and
might include services such as wallets or LRP programs. While these are pretty
standard apps, more sophisticated applications will be o ered as well, e.g.,
the federation server and the token name service (see Section 4). The Digital-Bits
federation server are similar to the DNS system of the Internet but instead of
resolving website urls to IP-addresses, the federation servers resolve token
aliases to the speci c smart contract token addresses. The token name certi
ca-tion service (see Section 5) represents
a certi cation service that will map token
12
names to an entity and ensures a correct binding, e.g., certify that the
Company A-token actually belongs to Company A and was not created by a
malicious entity.
Figure
6 illustrates the re nements
of the goal tokenize assets. In order to tokenize an asset, it is bind to a
token that is created and subsequently issued on the DigitalBits platform. As
mentioned previously, the token name certi cation service being developed can
be used to certify issued tokens. Finally, the new digitalized asset has to be
managed and be able to be integrated in applications and plugins of the
platform. A key quality goal of the requirement tokenize assets is legal compliance
in terms of who can view and validate transactions.
Loyalty
Program Provider
other
issuers
Tokenize assets
|
Compliant
|
|
Create Tokens
Issue Tokens
Certify Tokens
Mangage Tokens
Manage
plugins
or
App Interaction
Fig. 6: Lower-level re nements of the Tokenize
assets functional goal.
The third functional requirement trade assets,
is shown in Figure 7. The tokenized assets of Figure 6 are
worthless if it is not possible to trade and exchange them with other parties.
Hence, this functional goal covers on- and o -chain supply and demand
administration as well as trade of tokenized assets. Entities may register o
ers or requests on-chain in order to attract business partners. For other use
cases a local supply demand management o -chain is more suitable. In today's
highly digital world, automation in trading assets is expected by the platform
users. Multi-hop, i.e., the ability to trade tokens by rst converting them to a
common token or at currency is a feature that adds value to the platform. Convenience
while making payments, remittance and trading is an expected goal to enhance
consumer experience.
13
|
Multi-hop
|
Automated
|
Convenient
|
|
Trade assets
|
Loyalty Program Provider
|
Consumer
|
other issuers
|
|
Trade on/off
|
Manage
supply/demand
|
|
|
chain
|
|
|
Register
|
Seller
Buyer
Create offer
Update offer
Remove offer
Query offers
Matching
Currency
Conversion
Fig. 7:
Lower-level re nements of the Trade assets functional goal.
Consumer
Convenient
Multi-hop
Pay
and remit
Remit
Pay
Pay
Choose currency
Buy
Sell
Fig. 8: Lower-level re nements of the Pay and
remit functional goal.
Finally, the fourth functional requirement pay
and remit (Figure 8) focuses on the payment and remittance functionalities of our solution.
The sub-goal pay is further re ned into selling and buying from, or to
merchants. In the context of this model, a merchant can also be another user o
ering a service, or good. The remit sub-goal consists of selecting a currency
and recipient for the remittance process. The nal functionality is the
transaction itself and pertains the payment as well as the remittance process.
Again, the quality goals of multi-hop and convenience apply.
14
The presented goal model is used in the
following Section 4 to derive our system architecture. We do not list all details of the
further re ned AOM goal model in this whitepaper due to space constraints and
in order to focus on the most relevant system components and features.
4 DigitalBits System Design and Architecture
The DigitalBits
network consists of entities that perform di erent but compli-mentary roles in
order to maintain the health of the network. The key role is played by the
nodes that run the DigitalBits blockchain-based software and con-nect to one
another. These nodes are well supported by nodes that provide ser-vices such as
compliance veri cation, mapping and RESTful APIs. Additionally, APIs, SDKs and
wallets provided by DigitalBits facilitate businesses and third party
developers to easily develop and deploy their custom apps and wallets. Below,
Section 4.1 presents the high level
network overview that consists of the various entities that play a vital role.
Next, Section 4.2 presents the technology
stack and nally Section 4.3 focuses
on the system architecture itself.
4.1 DigitalBits Network Overview
Fig. 9: High Level DigitalBits Architecture
Overview
Figure 9 illustrates an overview of the Digitalbits architecture. The DigitalBits
architecture consists of the key components described below.
4.1.1 Frontier
Frontier provides a RESTful API for the DigitalBits ecosystem. It acts
as the interface to applications that wish to access the DigitalBits network.
Frontier
15
facilitates actions such as the submission of transactions to the
network, check-ing the status of accounts and subscribing to event streams. It
also ingests and re-serves the data produced by the Digitalbits network in a
form that is eas-ier to consume than the performance-oriented data
representations used in the network.
Application developers interact with Frontier's
Restful API via the web browser, simple command line tools like cURL, or the
DigitalBits SDK. Digital-Bits maintains JavaScript, Java, and Go-based SDKs for
communicating with Frontier. There are also community-maintained SDKs for Ruby,
Python, and C#. The Frontier APIs and SDKs can also be used to build or enhance
custom brand speci c Apps and clients.
4.1.2 Network Backbone:
DigitalBits Core
DB Core
DB Core
|
DB Core
|
DB Core
DB
Core DB Core
DB Core
DB Core
|
DB Core
|
Fig. 10: Quorums, i.e., circles of trust formed among DigitalBits core
instances (represented as DB core) of various partner institutes and
individuals. The Dig-italBits core instances can choose to belong to one or
more quorums and utilize them in a hierarchical manner or based on the type of
transaction that needs to be veri ed. The nodes belonging to a quorum need not
be located close to one another as can be observed in the blue quorum set.
16
The DigitalBits core is the backbone of the DigitalBits network. The
Digital-Bits Core software interacts with a chosen subset of other instances of
cores in order to validate and agree on the status of every transaction through
the DigitalBits Consensus Protocol (DCP) which is based on the Stellar Consensus
Protocol (SCP) [4]. Similar to Stellar's SCP, DCP relies on a Federated Byzan-tine Agreement
(FBA). Unlike the Byzantine Algorithm [9], the FBA
does not have a single list of trusted validators, i.e., a centralized list of
trusted validators. Instead, an FBA allows for di erent quorums or sets of
validators to co-exist. The nodes can determine the composition of the quorum
in a decentralized man-ner. As shown in Figure 10, DigitalBits
core instances of di erent institutions can choose to participate in one or
more quorums subject to the existing quo-rum members agreeing to grant it
access. The quorums facilitates the compliance process based on legal
requirements. Organizations may choose one or more quo-rums that satisfy their
requirements. Nodes or organizations could also choose to have a hierarchy of
trust with the parameters that de ne the level of trust that a node accepts for
each transaction. In other words, quorum composition and aspects such as simple
majority vs. 2/3 majority are de ned for di erent classes of transactions,
similar to parliamentary systems in many parts of the world.
The bene t of hosting an own instance of
DigitalBits core as compared to just running an App or client is many-fold.
Transactions can be submitted with-out having to rely on a third-party and the
DigitalBits core can select its own instance of who to trust, i.e., the quorum.
The more organizations and partners contribute instances of DigitalBits core to
the network, the more reliable and robust the network becomes. Each
organization can choose to run one or more DigitalBits core nodes, which also
participate as validators.
4.1.3 The DigitalBits Network
The DigitalBits network itself is a collection of connected DigitalBit
cores run by various individuals and entities around the world. Instances of
DigitalBits core add reliability to the overall network. Additionally, they may
choose to have a Frontier server for communication in order to access the
DigitalBits Network. The distributed nature of the network makes it reliable
and safe. All these Dig-italBit cores within the network eventually agree on
sets of transactions. Each transaction on the network costs a small fee: 100
nibbs (0.00001 XDB). This fee helps to prevent bad actors from spamming the
network. The DigitalBits Foundation also maintains archive servers with live
backups of the current state in the network in order to facilitate new
DigitalBits cores to come in sync with the current status of the network.
4.2 Technology Stack
17
we provide the overall architecture and a owchart based description of
how transactions are performed in the DigitalBits network.
Own Application Services
|
Bridge
|
Federation
|
Compliance
|
|
(e.g., Points Program, Wallets,
|
||||
Server
|
Server
|
Server
|
||
Explorer and etc.)
|
||||
|
|
|
XDB wallet
Node SDK Java SDK Ruby SDK Go SDK Community SDKs
SDK
Frontier
API
frontier
DigitaBits
Core Validator
digitalbits-core
Fig. 11: Technology Stack
4.2.1 Bridge Server
The Bridge
server is designed to support applications in easily performing trans-actions
on the DigitalBits network. The bridge server enables applications to use the
federation and compliance servers to send and receive payments. As shown later
in Figure 15, when a sender wishes to
perform a transaction, the sender's client contacts its bridge server in order
to initiate the transaction. If required, the bridge server then connects the
federation server of the receiver and its own compliance server. If all veri
cations are successfully completed, the transaction is recorded in the
DigitalBits network. The bridge server on the receiver's side periodically
monitors the DigitalBits network and spots transactions destined for its
end-point and then connects to the required federation and compliance servers
as well as nally accepts the transaction. The bridge servers then inform the
respective end-points about the result of the transaction.
18
4.2.2 Federation Server
DB-Federation-ID
|
DB-Account-ID
|
|
Federation
Server
DB-Federation-ID
<->
DB-Account-ID
DB-Account-ID
|
DB-Federation-ID
|
|
Fig. 12: Federation Server Overview: The federation server holds the
mapping between the DigitalBits Account-ID and the Federation-ID and support
lookup or reverse-lookup to map from one type of ID to another.
In order to enhance the consumer experience and ease adoption by
consumers, DigitalBits associates an account with an email-like human readable
identi - cation in addition to the standard public key based identi cation
prevalent in blockchain [34][48]. The human readable email-like address allows
consumers to easily use the Apps and clients without having to familiarize
themselves with public-key cryptography [15][41].
The
role of the federation server is therefore to provide a mapping service between
the email-like human readable address and the public key based ad-dress. Figure
12 illustrates how an entity
interacts with the federation server in order to either map a DB-Federation-ID
(the email-like address) to a DB-Account-ID (the public key based address) or
vice-versa. The DB-Federation-ID is of the form usernameyourdomain.com. For
example, a consumer named \Joe" could have a DB-Federation-ID
\Joedigitalbits.io", wherein \Joe" is his user-name and
\digitalbits.io" is the domain name. The username could also be an email
id. The domain can be any valid RFC 1035 [32] domain name. The Apps or clients automatically
perform the federation lookup (DB-Account-ID > DB-Federation-ID), or
reverse-lookup (DB-Federation-ID > DB-Account-ID)
19
and thereby allow the users to perform transactions by making use of
just the DB-Federation-ID.
4.2.3 Compliance Server
Fetch Info
|
Customer DB
|
|||
|
|
|
||
Callback
|
||||
|
||||
|
|
|
|
Compliance
|
|
Sanctions
|
Sanction DB
|
||
|
|
|
|
|
|
Server
|
|
Callback
|
|||
|
|
||||
|
|
|
|
|
|
Ask User
|
Compliance DB
|
|
Callback
|
||
|
Fig. 13: Callbacks performed by the compliance server to the various
databases at the sender as well as receiver side in order to obtain the
necessary information to perform a compliance veri cation.
It is the
responsibility of the anchor to handle regulatory compliance, especially to
adhere to Anti-Money Laundering (AML) regulations. Complying with AML laws
requires certaion enterprises and institutions (EIs) to know not only who their
customers are sending value to, but also who their customers are receiv-ing
value from. However, in some jurisdictions certain EIs are able to trust the
AML procedures of other EIs. In other jurisdictions each EIs must do its own
background check of both the sender and the receiver. The DigitalBits
compli-ance protocol supports the exchange of compliance information to
pre-approve a transaction with another EI. The customer information that is
exchanged be-tween EIs via the compliance protocol is quite exible and
typically consists of the full name, date of birth and physical address.
Figure 13 illustrates
the callbacks that are used to obtain information at the sender side and also
to verify compliance in case a sender's compliance server contacts a receiver's
compliance server to verify compliance before clearing a transaction. The
following callbacks are performed:
{ Fetch info callback: This callback returns all the information about a
particular consumer on receiving the consumer's federation address.
{ sanctions callback: This callback is used to
identify if any sanctions exist in order to receive money from the sender. The
HTTP response code it
20
produces
indicates whether the payment is accepted (status 200), denied (status 403), or
if additional time for processing is required (status 202).
{ Ask user callback: This callback is called in case a sender has
requested information about the receiver. The HTTP response code it produces
indi-cates whether the information can be sent. If the callback is a success, the
fetch info callback is called to obtain the information.
4.2.4 Wallets and Apps
Businesses and third-party developers could easily develop custom Apps
by lever-aging the Frontier API and DigitalBits SDK. DigitalBits also provides
a native XDB wallet source code that can be directly used or easily adapted to
create a brand speci c wallet. The bridge server facilitates easy access for
the end points to the federation and compliance server.
4.3 System Architecture
Figure 14a presents a owchart depicting the decision making process of a bridge
server when it receives a request for a new transfer. The bridge server would
con-tact the federation server of the receiver in case a mapping from a
DB-Account-Id to a DB-Federation-Id for the receiver is required. It then veri
es whether com-pliance is required for the transaction to be validated. If
compliance is required, the bridge server contacts the compliance server to
initiate the compliance pro-tocol. The bridge server then places the
transaction as a pending transaction in the DigitalBits network. Once the
receiver accepts the transaction (see Fig-ure 14b), the
bridge server informs its endpoint. The endpoint then modi es its account
balances accordingly.
Figure 14b presents
a owchart depicting the decision making process of a bridge server on the
receiver side. The bridge server continuously monitors the DigitalBits network
for any new pending transactions destined towards it. On observing such a
transaction, the bridge server contacts the federation server of the sender in
case a mapping from a DB-Account-Id to a DB-Federation-Id for the sender is
required. It then veri es whether compliance is required for the transaction to
be validated. If compliance is required, the bridge server contacts the
compliance server to verify if compliance was already obtained. Accordingly,
the bridge server either accepts or rejects the transaction in the DigitalBits
network. If the transaction was positively veri ed and accepted, the receiver
endpoint modify its account balances accordingly.
(a).pdf
21
(b).pdf
Start
|
|
|
Yes
|
Federation
|
Federation
|
||||
Required
|
|
Server
|
||
|
|
|
|
|
No
|
|
|
||
|
|
|
Yes
|
Compliance
|
Compliance
|
||||
Required
|
|
Server
|
No
|
|
|
|
|
|
|
|
|
|
||
|
|
Yes
|
Compliant
|
||
|
|
||||
Insert
|
|
||||
|
|
|
|
||
Transaction in
|
|
|
|
|
|
DigitalBits
|
|
|
|
|
|
Network
|
|
|
|
|
|
|
|
|
No
|
No
Transaction
Accepted
Yes
Inform
Endpoint
End
Start
Pending
Transactions
in
DigitalBits
Network
Yes
Federation Yes
Required
No
Compliance Yes
Required
No
Yes
Accept
Transaction
in
DigitalBits
Network
Yes
Inform
Endpoint
Federation
Server
Compliance
Protocol
Compliant
No
Reject
Transaction
in
DigitalBits
Network
End
(a) Bridge
Server at the sender side (b) Bridge Server at the receiver side
Fig. 14:
Flowcharts depicting the decision making process of a bridge server on
receiving a request for sending or receiving tokens.
22
|
|
|
|
Sender: Tom
|
|
|
|
DigitalBits
|
|
|
|
|
|
Receiver: Jane
|
|
|
||||||||||
|
|
|
|
|
|
|
Network
|
|
|
|
|
|
|
|
||||||||||||
|
|
|
|
Enterprise A
|
|
|
|
|
|
|
|
|
|
|
|
Enterprise B
|
|
|
||||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
App /
|
|
Bridge
|
|
Federation
|
Compliance
|
Compliance
|
Federation
|
Bridge
|
|
App /
|
||||||||||||||||
Client
|
|
Server
|
|
Server
|
Server
|
Server
|
Server
|
Server
|
|
Client
|
||||||||||||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Send X Tokens
|
|
|
|
|
|
|
Get Jane
|
|
|
|
|
|
|
|
|
|
|
|
|
||||||
|
to Jane
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||
|
DB_Federation_ID
|
|
|
|
|
|
|
|
DB_Account_ID
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||
|
|
|
|
|
|
|
|
|
|
|
Jane
|
|
|
|
|
|
|
|
|
|
|
|
|
|||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|||
|
|
|
|
|
|
|
|
|
|
|
DB_Account_ID
|
|
|
|
|
|
|
|
|
|
|
|
|
|||
|
|
|
|
Request
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
||||
|
|
|
|
Authorization
|
|
|
|
Callbacks to
|
|
|
|
|
|
|
|
|
|
|
|
|
||||||
|
|
|
|
|
|
|
|
|
|
|
|
fetch Tom’s
|
|
|
|
|
|
|
|
|
|
|
|
|
||
|
|
|
|
|
|
|
|
|
|
|
|
info
|
|
|
|
Callbacks to
|
|
|
|
|
|
|||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
verify Jane’s
|
|
|
|
|
|
|||
|
|
|
|
Compliance
|
Verified
|
|
|
|
|
|
|
|
|
|
compliance
|
|
|
|
|
|
||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
parameters
|
|
|
|
|
|
||||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Send X Tokens
|
to Jane
|
|
|
|
|
|
Event stream or
|
polling for pending
|
transactions
|
|
|
|
|
|||||||||
|
|
|
|
DB_Ferderation_ID
|
|
|
|
|
|
|
|
|
||||||||||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Send X Tokens
|
to Jane
|
|
|
|
|
|
|
|
|
||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
DB_Ferderation_ID
|
|
|
|
|
|
|
|
|
|
||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Is Compliance
|
Verified?
|
|
|
|||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Compliance
|
|
Verified
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Accepted
|
Transaction
|
|
|
|
|
Inform
|
|
|
|||
|
|
|
|
Event stream or polling for
|
pending transactions
|
|
|
|
|
|
|
|
|
|
Endpoint
|
|
|
|||||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||
|
|
|
|
|
Transaction
|
Accepted
|
|
|
|
|
|
|
|
|
|
|
|
|
||||||||
|
Inform
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
||
|
Endpoint
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Fig. 15: A sequence diagram to describe how a sender (Tom) sends tokens
to a receiver (Jane).
Figure
15 presents an overview of the
processes at both the sender and the receiver side to perform a transfer. Let
us assume that Tom wants to transfer tokens to Jane. Tom's App contacts its
bridge server in order to initiate the transfer. The bridge server then contact
the receiver's federation server to ob-tain Jane's DB-Account-ID. Afterwards it
requests authorization from its com-pliance server. The compliance server makes
use of the DigitalBits compliance protocol to clear the transaction with the
recipients compliance server and ac-cordingly informs its bridge server. The
bridge server places this transaction in the DigitalBits network and wait for
the transaction to be accepted. Meanwhile, the bridge server on the receiver
side observes this pending transfer via a peri-odic polling mechanism. Jane's
bridge server veri es with its compliance server whether compliance veri cation
has been performed for the transaction. Accord-ingly, the bridge server accepts
or rejects the pending transfer on the DigitalBits network. Both Tom's and
Jane's bridge servers inform their respective endpoints of the validity of the
transaction.
23
5 Stakeholders Engagement and Services Interaction
DigitalBits
blockchain powered infrastructure builds a bridge that facilitates the
implementation of new technologies to support and enhance our every day life
interactions as well as foster blockchain mass adoption. Hence, stakeholder
engagements and service interactions are essential to the DigitalBits platform.
The underlying processes that enable engagement and interaction are the result
of collaborating tasks and sub-processes. Based on the chosen LRP program
running case from Section 2, we
outline the exemplary detailed processes and bene ts of involved stakeholder,
e.g, for tokenizing assets, trading assets as well as validating and
authenticating tokenized assets via certi cation service.
Consequently, Section 5.1 details
the on-boarding process of digital assets, followed by Section 5.2 that
details the digital asset validation and authentica-tion service. Afterwards,
Section 5.3 outlines the built-in decentralized multi-hop exchange of DigitalBits,
followed by Section 5.4 that covers the token value proposition. Finally, Section 5.5 focuses
on incentives that foster the DigitalBits community.
5.1 On-Boarding Process of
Digital Assets
In previous
sections we brie y outlined the general work ow of tokenizing an asset on the
DigitalBits blockchain. Even though this paper focuses speci cally on the tokenization
of LRP programs, the underlying concept is more diverse and can be applied to
commodities such as diamonds [37] and
gold [2], or securities
[24]. In
general, tokenization is a method that converts the rights to an asset into a
digital token that resides on a blockchain [8][21] and
DigitalBits supports the tokenization of all types of assets.
In
the context of our LRP program running case, we may face two di erent scenarios
when it comes to tokenizing those programs. First, the asset provider (or
issuer) wants to set up a brand new program without any legacy dependen-cies.
Second, the assets provider already has an existing legacy (non-blockchain) LRP
program that has to be migrated into the DigitalBits system. While the rst case
is rather straight forward, the second case is more complex.
In
either case, the asset provider (or issuer) is to choose an identi cation code
for the new asset - a combination of up to 12 letters or numbers that identify
the asset in a human-readable form. Afterwards, the asset is ready to be used
on the network. However, before other users are able to receive the loyalty
tokens of ctional HotelBrand company from the running case of Section 2.2, the users have to choose to
trust it (or the ctional company may do this in batch for all its users, if it
is electing to initially hold custody of its members' accounts), since a
DigitalBits asset is actually a credit. Therefore, users have to trust that
HotelBrand can redeem that credit if necessary later on. In the context of a
LRP program, the user usually trusts the asset provider, e.g., a retail
company. Each account can create a trustline, or a declaration that it trusts a
particular asset. In addition, users can also limit the trust to a particular
amount of tokens. This security feature ensures that small asset providers with
limited credibility do not
24
hand out
excessive amounts of tokens. In the context of our running case, such details
will be dealt with when a new user decided to join the loyalty program. By
default, users have to trust the issuer to participate in the program.
The
loyalty and reward industry uses so called reward engines to manage the mapping
between purchases and allocated reward points (e.g., staying in a hotel for ve
nights results in a reward of 25 loyalty points). Those mappings as well as the
terms and conditions that apply when redeeming the points are managed by the
reward engine and the assets' providers IT system. As mentioned previously in
Section 4, the asset providers bridge
server facilitates the interactions between the DigitalBits network and the
application-services on the asset provider side.
Finally,
going back to the second scenario - that we outlined in the begin-ning of this
section - assuming that HotelBrand wishes to migrate its existing program to
the DigitalBits infrastructure. The work ow is mainly the same; a new asset is
initially created on the blockchain and then the respective trust settings are
performed. However, further application logic on the provider side is necessary
to manage the migration process from the legacy platform to the blockchain
solution. There are three approaches to that issue: First, HotelBrand may
create accounts (consisting of public and private keys) for each legacy user
and equips the corresponding accounts with a number of tokens that is equal to
the number of loyalty points in the legacy system. The issuer can then create
the keys and manage those accounts directly via the DigitalBits SDK, or the
fron-tier server. A second solution could rely on the loyalty program bridging
their existing database via their bridge server to handle any on-chain actions.
In this case, the LRP program continues to rely on the existing accounting
database of the legacy system while handling blockchain events via the bridge
server, e.g., send and receive tokens. Finally, the third option is that in
order to issue a new blockchain-based loyalty and reward tokens to customers
with legacy tokens, these customers need to register with a DigitalBits account
ID (as presented earlier in Section 4). As soon as an account has been registered, the issuer trans-fers
tokens equivalent, or proportional to the points that the customer possessed in
the legacy program. The migration process may consist of di erent stages, e.g.,
early access program, proof-of-concept and cut-over. Point-holders that wish to
further use their accumulated points have to migrate before the cut-over in
order to ensure that they do not loose their points.
5.2 Token Name Certi cation
Service (TNCS) - Validation and Authentication of Asset Providers
In the previous
section we elaborated on the process of on-boarding a new asset to the
blockchain and tokenizing the asset. However, whether a speci c tokenized asset
actually represents the analogue asset of the real world is unclear. Ensuring a
binding of the analogue value and its digital representation is a common
prob-lem in computer science and most thoroughly studied in the context of
digital identities, e.g., [1][33]. How
can Jane digitally prove to Tom that she is fact the owner as well as analogue
representative of a speci c digital identity? An equiv-alent problem, even
though less complex than the human identity issue, arises
25
in the context of certi cates for websites [16]. Most
common solution for website certi cates and digital identities are so called
public key infrastructures (PKIs) that are either organized in a centralized
and hierarchical manner (so called cer-ti cate authorities - CAs), or in a
decentralized manner such as the PGP Web of Trust [49]. Both
concepts have di erent advantages and disadvantages that have been partially
solved by moving the underlying ideas to a blockchain-based platform [10][27][44].
In
order to prevent malicious entities from issuing tokens that represent a brand
or company that they are not associated with, we suggest a so called Token Name
Certi cation Service (TNCS) be developed for validation and au-thentication of
asset providers. Service providers within the DigitalBits network might then o
er services similar to SSL certi cate authorities that maintain a mapping
between token smart contract addresses and the identities of token is-suers. As
a result, it becomes di cult for a malicious attacker to impersonate a
legitimate company or issuer and issue a counterfeit asset on their behalf. The
advantage for users and consumers is, that they do not have to worry about
buy-ing/trading worthless counterfeit tokens that have the same name as the
original, but a di erent token address. Hence, it fosters security and convenience
for users. A blockchain-based protocol that was designed to provide such
services is the
Asset
Provider TNCS Provider
1.
Challenge Request
2.
Challenge Response
Block 1
|
|
|
…
|
|
|
Block X
|
|
|
|
|
|||
|
|
|
|
DigitalBits
Blockchain
3.Store challenge and result in
blockchain
Fig. 16: Challenge response-based validation and authentication process
on a blockchain (Adapted from [27]).
Authcoin
protocol [26][27][28] that can easily be adapted
to enable the TNCS on the DigitalBits network. The protocol provides validation
and authentication for secure identity assurance on the blockchain via a
challenge-response mechanism. In general, validation aims to prove that an
entity has access to a certain resource (in our case access to a DigitalBits
account that issued a speci c token), while authentication continues the
validation procedure by verifying the identity of the issuer and aims to con rm
that the entity with access to the account is also the actual representative of
the company. Figure 16 illustrates
the idea in the context of DigitalBits TNCS. An asset provider (token issuer)
wants the TNCS provider to approve the authenticity of their digital assets. To
do so, the issuer requests
26
a challenge
that is then created by the TNCS provider. Afterwards, the issuer ful ls the
challenge and sends the response back. Challenge as well as response are
transparently stored in the blockchain. The chosen challenge depends on the
use-case scenario, the required level of security and the given threat level of
the involved entities. In the context of a LRP program provider the challenge
might ask the issuer to provide a company statement that proves that they
issued the token and that they are in fact the company they claim to be.
5.3 Decentralized Multi-Hop
Exchange
Besides the tokenization and transfer of digital assets such as LRPs, it
also enables seamless transfer or trade of those assets on-chain with others
entities interested in receiving those assets. This helps to solve many of the
existing issues of non-tokenized assets that often su er from missing market
liquidity. In some cases, it is possible to seamlessly trade these assets even
if no direct market exists between two assets. As a Stellar [31] fork,
this decentralized multi-hop exchange works similar to the original.
5.3.1 Matchmaking
Users interested in trading LRPs or any other asset can create an o er
to buy or sell an asset. In order to make an o er, the account must hold the
asset it wants to sell. Moreover, the user is required to trust the issuer of
the asset it is trying to purchase. When an account creates an o er, the o er
is checked against the existing orderbook for that asset pair. If the o er
crosses an existing o er, it is lled at the price of the existing o er. If not,
the o er is saved in the orderbook until it is either taken by another o er,
taken by a payment, canceled by the account that created the o er, or
invalidated because the account making the o er no longer has the asset for
sale.
5.3.2 Cross-Asset Multi-Hop
Payments
Let us assume
that Jane owns points from Company A and wants to buy an item or a service from
a merchant that only accepts points from Company B. In this scenario, Jane can
create a payment in Digitalbits that automatically converts her points from
Company A into points of Company B by querying the orderbook and converting
among the points at the best available rate. In case the orderbook does not contain
any o ers for such a conversion, it is also possible to rst convert it into
another asset, e.g., points from Company C, and afterwards converting those
points to the target asset class. The number of intermediate steps before the
nal conversion is limited to a maximum of 6 hops and atomic7.
Since
cross-asset payments and conversions are simple and seamlessly, users are not
required to hold any unwanted assets just for payment purposes. Instead,
7 It will either succeed or
fail and the payment sender will never be left holding an unwanted asset.
27
they keep their preferred LRPs of their favorite brand (or any other
asset), only converting them if necessary. As a result, the DigitalBits system
can o er and ensure liquidity in previously illiquid markets, thereby enabling
a world where users never have to exchange currency except at the point of
sale. Moreover, users could choose to keep all their assets in, for example
stocks, only cashing out small amounts as they need to pay for things.
5.4 Token Value Proposition
While previous
sections focused on the architecture, services and processes of the DigitalBits
blockchain and ecosystem, the following section details the value proposition
of its native token - the XDB token. XDB serves three main objec-tives:
Firstly, as a protective security feature. Each account on the DigitalBits
blockchain is required to stake a minimum of 10 XDB tokens to ensure an
ac-count is authentic and for the send-function to be enabled on the network.
For example, if Jane wants to send 20 Tokens to Tom, her account must have a
minimum of 30 Tokens to do so. Moreover, each transaction results in a minor
transaction fee of 0.00001 XDB. Both requirements serve as protective security
features and prevent users with malicious intentions from ooding the network.
Secondly, XDB enables transaction among non-native tokens, by acting as a
bridge to facilitate trades between pairs of other digital assets, which may
not have a large direct market. Finally, DigitalBits XDB token can also be used
for fast and low-cost micro-payments and remittances.
5.5 Community Engagement
The DigitalBits
platform and the surrounding ecosystem aim to support mass adoption of
blockchain technology. Hence, a community of various partners and users is
crucial for the further development of DigitalBits. The following section
introduces plans for token allocations, scholarships, grants and donations that
will facilitate the community engagement of the DigitalBits ecosystem.
As illustrated in Figure 17, at the
launch of the network, 32% of the XDB tokens will be reserved for an initial
token sale and any that are unsold we will reserved for future use. 53% of the
XDB tokens will be reserved for rewards to users, grants, airdrops and
donations to charity. Lastly, another 15% are held by Fusechain Group Ltd. and
its subsidiary, the founding contributor to the DigitalBits network, reserved
for the Team and advisors. The initial allocation of 40.0% designated for rewards
to users as well as donations to charities is separated into two pools: The
DigitalBits Algorithmic Pool (39.60%) and the Charity Pool (0.40%). The XDB
reserved for the DigitalBits Algorithmic Pool (DBAP) is designated for give
away based on DigitalBits blockchain use and determined by DBAP's algorithms
and system currently under development. This DBAP pool will be restricted and
remain in reserve until this development is complete. Transaction fees are also
used to re ll the pool. XDB within the DBAP is released directly to users via
their respective accounts on the blockchain when the rewards are earned.
28
40.0%
|
Algorithmic Pool (39.6%) and
Charity
|
|
Pool (0.40%). Restricted until
algo
|
||
|
launch within 24months.
|
|
5.0%
|
Distributed via partnership
development
|
|
|
program, specifically driving
benefit to
|
|
|
DigitalBits;
|
|
5.0%
|
Distributed via R&D grants;
|
|
3.0%
|
Distributed via airdrops and
bounties;
|
|
15.0%
|
12.5% distributed to founders,
team, and
|
|
contractors and 2.5% to advisors
(“Team”)
|
||
|
||
32.0%
|
Up to 32% to be sold in a token
sale and
|
|
the balance to be held for
future use.
|
||
|
||
|
|
Fig. 17: XDB Token distribution at the launch of
the DigitalBits network.
Donations
from the Charity Pool are given to designated charities at a ratio of 1-to-99
with the DBAP, until the Charity Pool supply is exhausted. For example, if 990
XDB are released by the DBAP, then 10 XDB are released from the Charity Pool.
The Foundation plans to identify, designate and monitor charities to be
recipients of the XDB released from the Charity Pool. The Charity Pool will
restricted and remain in reserve until the DBAP system is developed.
Finally, we plan to give away a 5% allocation
through the DigitalBits Re-search and Development Grant Program and 5% Partner
Ecosystem Develop-ment Program from time to time. The Partner Grant Program and
Partner Ecosystem Development Program are focused on encouraging prospective
part-ners to develop and operate products, services, or solutions that are
important to the DigitalBits network.
6 Conclusion
This whitepaper presents the DigitalBits blockchain powered
infrastructure that facilitates the implementation of new technologies to
support and enhance our every day life interactions as well as foster the mass
adoption of blockchain. Moreover, our solution facilitates easy
asset-tokenization using a transaction and trading layer for the point-to-point
economy.
Based
on the use cases and scenarios, we identify the requirements and crite-ria that
the DigitalBits blockchain infrastructure must satisfy. With respect to
functional and non-functional requirements, the DigitalBits network has to be
easy to use, exible and convenient in order to reach the goal of mass-adoption.
Furthermore, interoperability and scalability as well as security and compliance
are further mandatory properties of our solution. Moreover, easy integration
29
into other services and a plugin interface for external applications are
essential to DigitalBits.
Subsequently, we derive and outline the
DigitalBits network and system ar-chitecture based on the identi ed
requirements and goals. We present the overall architecture and detail on the
key network components that enable the Digital-Bits platform. Moreover, we
outline the communication interfaces, the network topology and describe the
interplay of services and entities within the network. In order to ensure
widespread adoption, DigitalBits o ers a variety of SDKs and easy to use APIs
for developers, partners and the community.
Next,
we deal with the exemplary stakeholder engagement and service in-teraction
processes that are essential to the DigitalBits platform. First, the
on-boarding process of digital assets and the asset tokenization which is an
indis-pensable functionality of our platform. Afterwards, present the novel
idea of the token name certi cation service (TNCS) that prevents malicious
network enti-ties from issuing and distributing illegitimate tokens of assets
that they are not associated with. In order to trade and exchange tokenized
assets in a proper manner, DigitalBits provides a built-in decentralized
multi-hop exchange for cross-asset payments. Finally, we present the XDB token
value proposition and the surrounding token economy ecosystem that fuels the
DigitalBits platform.
References
1. Alsarkal, Y., Zhang, N.,
Zhou, Y.: Linking Virtual and Real-World Identities. In: Intelligence and
Security Informatics (ISI), 2015 IEEE International Conference on. pp. 49{54.
IEEE (2015)
2. Aurus.io: Aurus: Tokenized
Physical Assets - Whitepaper. URL: https: //aurus.io/wp-content/uploads/2018/08/Aurus-Whitepaper-V2.1.pdf
(2018),
(Accessed August 29, 2018)
3. Barnett, C.: Inside the
Meteoric Rise of ICOs. URL: https://www.forbes. com/sites/chancebarnett/2017/09/23/inside-the-meteoric-rise-of-icos/#f5de0bf5670c (2017), (Accessed September 01, 2018)
4. Barry, N., Losa, G.,
Mazieres, D., McCaleb, J., Polu, S.: The Stellar Consensus Protocol (SCP)
(2015), (Accessed September 02, 2018)
5. Berry, J.: The 2015 COLLOQUY Loyalty Census: Big
Numbers, Big hur-
dles. URL: https://www.colloquy.com/latest-news/2015-colloquy-loyalty-census/ (2015), (Accessed August 17, 2018)
6. Bond Brand Loyalty: The Loyalty Report 2017.
URL: http://info. bondbrandloyalty.com/2017-loyalty-report (2017), (Accessed August 15, 2018)
7. Brian Patrick Eha, American Banker: How Barclays
Aims to Bring a Billion
Unbanked into
the Fold. URL: https://www.americanbanker.com/news/how-barclays-aims-to-bring-a-billion-unbanked-into-the-fold
(2016), (Ac-cessed
August 17, 2018)
8. Cameron-Hu , A.: How Tokenization
Is Putting Real-World Assets on Blockchains (2017), (Accessed August 29, 2018)
9. Castro, M., Liskov, B., et
al.: Practical byzantine fault tolerance. In: OSDI. vol. 99, pp. 173{186 (1999)
30
10. Civic Technologies, Inc.:
CIVIC - Whitepaper. URL: https://tokensale.civic. com/CivicTokenSaleWhitePaper.pdf (2017), (Accessed August
29, 2018)
11. COLLOQUY: 2015 COLLOQUY
Loyalty Census Report. URL: https://www. colloquy.com/reports/ (2015), (Accessed August 14, 2018)
12. COLLOQUY: 2017 COLLOQUY Loyalty Census Report -
An In-depth Analysis of
Where Loyalty Is Now ...
and Where It's Headed. URL: https://www.colloquy. com/reports/ (2017), (Accessed August 14, 2018)
13. Davis, A.M.: Software
Requirements: Objects, Functions, and States. Prentice-Hall, Inc. (1993)
14. Deloitte: The Deloitte Consumer Review -
Customer Loyalty: A Relationship
not just a Scheme. URL: https://www2.deloitte.com/content/dam/Deloitte/ uk/Documents/consumer-business/deloitte-uk-consumer-review-customer-loyalty.pdf (2017), (Accessed August 15, 2018)
15. ElGamal, T.: A public key
cryptosystem and a signature scheme based on discrete logarithms. IEEE
transactions on information theory 31(4), 469{472 (1985)
16. Ellison, C., Schneier, B.:
Ten Risks of PKI: What you're not Being Told About Public Key Infrastructure.
Comput Security Journal 16(1), 1{7 (2000)
17. Encyclopedia.com: Gold
Standard. URL: https://www.encyclopedia.com/ social-sciences-and-law/economics-business-and-labor/money-banking-and-investment/gold-standard, (Accessed August 17,
2018)
18. Experian: Driving customer loyalty - Maximize
loyalty program data collec-
tion to drive insight and revenue. URL: http://cdn.qas.com/us-marketing/ whitepapers/driving-customer-loyalty.pdf (2014), (Accessed August 15, 2018)
19. Fruend, M.: The 2017
COLLOQUY Loyalty Census. URL: https://www. colloquy.com/latest-news/u-s-customer-loyalty-program-memberships-reach-double-digit-growth-at-3-8-billion-2017-colloquy-loyalty-census-reports/ (2017), (Accessed August 17, 2018)
20. Gordon, N., Hlavinka, K.: The 2011 Forecast of
U.S. Consumer Loyalty Program
Points Value.
URL: http://www.swiftexchange.com/Content/Documents/2011-COLLOQUY-Liability-Talk-White-Paper.pdf (2011), (Accessed August
14, 2018)
21. Hargrave, J., Sahdev, N.K.,
Feldmeier, O.: How Value is Created in Tokenized Assets (2018), (Accessed August
29, 2018)
22. IEEE Computer Society.
Software Engineering Technology Committee and Insti-tute of Electrical and
Electronics Engineers: IEEE Recommended Practice for Soft-ware Requirements
Speci cations. IEEE Std, Institute of Electrical and Electronics Engineers
(1994)
23. Ivanov, S.: WAVES Platform
- Whitepaper. URL: https://wavesplatform.com/ files/images/whitepaper_v0.pdf, (Accessed August 26,
2018)
24. Koverko, T., Housser, C.: Polymath: The
Securities Token Platform - Whitepaper.
URL: https://polymath.network/whitepaper.html (2018), (Accessed August
29, 2018)
25. L. M. Goodman: Tezos - A
Self-Amending Crypto-Ledger (White paper). URL: https://www.tezos.com/static/papers/white_paper.pdf (2014), (Accessed Au-gust
20, 2018)
26. Leiding, B.: Securing the Authcoin Protocol
Using Security Risk-oriented Patterns
27. Leiding, B., Cap, C.H., Mundt, T.,
Rashidibajgan, S.: Authcoin: Validation and
Authentication in
Decentralized Networks. In: The 10th Mediterranean Conference on Information
Systems - MCIS 2016. Cyprus, CY (September 2016)
31
28. Leiding, B., Norta, A.:
Mapping requirements speci cations into a formalized blockchain-enabled
authentication protocol for secured personal identity assurance. In:
International Conference on Future Data and Security Engineering. pp. 181{ 196.
Springer (2017)
29. Liu, Y.: The Long-Term
Impact of Loyalty Programs on Consumer Purchase Be-havior and Loyalty. Journal
of marketing 71(4), 19{35 (2007)
30. Marshall, J.: Agent-Based
Modelling of Emotional Goals in Digital Media Design Projects. International
Journal of People-Oriented Programming (IJPOP) 3(1), 44{59 (2014)
31. Mazieres, D.: The Stellar Consensus Protocol: A
Federated Model for Internet-level
Consensus - Whitepaper.
URL: https://www.stellar.org/papers/stellar-consensus-protocol.pdf (2016), (Accessed August
26, 2018)
32. Mockapetris, P.: Domain
names - implementation and speci cation. URL: http://
tools.ietf.org/rfc/rfc1035.txt (November 1987), (Accessed August 29, 2018)
33. Mordini, E., Massari, S.:
Body, Biometrics and Identity. Bioethics 22(9), 488{498 (2008)
34. Nakamoto, S.: Bitcoin: A
Peer-to-Peer Electronic Cash System. URL: https:// bitcoin.org/bitcoin.pdf
(2008), (Accessed
August 01, 2018)
35. Norta, A.: Exploring Dynamic
Inter-Organizational BusinessProcess Collabora-
tion: Privacy Protecting
Concepts for ChoreographingeSourcing in B2B with Service-Oriented Computing.
VDM Verlag (2008)
36. Norta, A., Grefen, P.,
Narendra, N.C.: A Reference Architecture for Managing Dynamic
Inter-Organizational Business Processes. Data & Knowledge Engineering 91,
52{89 (2014)
37. Norta, A., Levi, S., Priewer, R., Kif, E.:
CEDEX: Certi ed Blockchain Based Di-
amond Exchange -
Whitepaper. URL: https://cedex.com/img/Whitepaper.pdf (2017), (Accessed August
29, 2018)
38. Owyang,
J., Groopman, J.,
Szymanski, J., Rebecca,
L.: Analysis: Should
Blockchain
Power Your Customer Loyalty Program? URL: http://www.web-strategist.com/blog/2018/03/09/analysis-should-blockchain-power-your-customer-loyalty-program/ (2018), (Accessed August 13,
2018)
39. Popov, S.: The Tangle -
Version 1.4.2. URL: https://iota.org/IOTA_Whitepaper. pdf (2018),
(Accessed August 22, 2018)
40. Qiibee Foundation: Qiibee:
Loyalty on The Blockchain - Whitepaper 2.4. URL: https://static.qiibee.com/qiibee-White-Paper.pdf (2018), (Accessed Au-gust
17, 2018)
41. Rivest, R.L., Shamir, A.,
Adleman, L.: A method for obtaining digital signatures and public-key
cryptosystems. Communications of the ACM 21(2), 120{126 (1978)
42. Robinson, S.: The 2013
Maritz Loyalty Report. URL: https://nrf.com/ resources/retail-library/the-2013-maritz-loyalty-report-benchmarks-trends-and-lessons-brand-loyalty (2013), (Accessed August
17, 2018)
43. Russolillo,
S.: Initial Coin O
erings Surge Past
$4 Billionand Regulators
Are Worried.
URL: https://www.wsj.com/articles/initial-coin-offerings-surge-past-4-billionand-regulators-are-worried-1513235196
(2017), (Ac-cessed
September 01, 2018)
44. SelfKey Foundation: SelfKey
- Whitepaper. URL: https://selfkey.org/ whitepaper/ (2017), (Accessed August
30, 2018)
45. Stellar Development
Foundation: Stellar Basics - Website. URL: https://www. stellar.org/how-it-works/stellar-basics/, (Accessed August 17,
2018)
46. Sterling, L., Taveter, K.: The Art of
Agent-Oriented Modeling. MIT Press (2009)
32
47. Synchrony Financial and Reuters Content
Solutions: Much Love for Loyalty
Programs. URL: https://www.reuters.com/brandfeatures/synchrony/inside-retail/much-love-for-loyalty-programs (2015), (Accessed August
17, 2018)
48. Wood, G.: Ethereum: A Secure Decrentralized
Generalised Transaction Ledger.
URL: http://gavwood.com/paper.pdf
(2014), (Accessed August
01, 2018)
49. Zimmerman, P.: PGP 2.X
Manual. URL: https://web.pa.msu.edu/reference/ pgpdoc1.html (1994), (Accessed August 30, 2018)
33
Disclaimer
It is forecasted that it will take over 10 years
to exhaust the reserves within the DigitalBits Algorithmic Pool and Charity
Pool. However, this is subject to several factors, including demand and the
algorithms that will be integrated into the DigitalBits Algorithmic Pool.
Therefore, this timeline cannot be guaranteed and forecasts may uctuate from
time to time.
This
whitepaper is for information purposes only. DigitalBits Foundation does not
guarantee the accuracy of or the conclusions reached in this whitepaper, and
this whitepaper is provided as is. DigitalBits Foundation does not make and
expressly disclaims all representations and warranties, express, implied,
statu-tory or otherwise, whatsoever, including, but not limited to: (i)
warranties of merchantability, tness for a particular purpose, suitability,
usage, title or non-infringement; (ii) that the contents of this whitepaper are
free from error; and
(iii)
that such contents will not infringe third-party
rights. DigitalBits Founda-tion and its a liates shall have no liability for
damages of any kind arising out of the use, reference to, or reliance on this
whitepaper or any of the content contained herein, even if advised of the
possibility of such damages. In no event will DigitalBits Foundation or its a
liates be liable to any person or entity for any damages, losses, liabilities,
costs or expenses of any kind, whether direct or indirect, consequential,
compensatory, incidental, actual, exemplary, punitive or special for the use
of, reference to, or reliance on this whitepaper or any of the content
contained herein, including, without limitation, any loss of business,
revenues, pro ts, data, use, goodwill or other intangible losses.
Không có nhận xét nào:
Đăng nhận xét
Cảm ơn bạn đã nhận xét