![](bg2.png)
a problem arising from an increasing number of trans-
actions but a limited throughput in major blockchain
platforms.
2.1 Scalability
Scalability is one of the most important aspects of
many distributed systems such as blockchains. Here,
it refers to the speed at which participants of a peer-to-
peer network can reach consensus on the state of the
blockchain (Hafid et al., 2020). Mathematically it can
be represented as the maximum block size divided by
the block interval (Croman et al., 2016). Following
this, solving the scalability issue can be done by ei-
ther increasing the block size or decreasing the block
interval. However, external factors such as computing
power, bandwidth, and storage space (Buterin, 2021)
call for an internal solution to the problem. This is
where the blockchain trilemma arises, existing solu-
tions such as the Proof-of-Stake (PoS) consensus pro-
tocol trade in decentralization in favor of scalability.
Using only a limited number of validators, a partic-
ular type of node allowed to create and confirm new
blocks, PoS protocols can decrease required network
communication and increase its scalability. Proof-of-
work (PoW) protocols do not differentiate between
different types of nodes, and everyone has the same
rights.
Low blockchain transaction rates lead to a prob-
lem where transactions can no longer be processed
immediately. Therefore, in the context of blockchain,
scalability refers to the ability to support high trans-
actional throughput while maintaining performance.
Croman et al. (Croman et al., 2016) identified key
metrics to measure scalability of blockchain plat-
forms: maximum throughput, latency, bootstrap
time, and cost per confirmed transaction, where the
first two measurements are the most important for a
user who intends to use a blockchain without being a
miner or a validator.
Maximum throughput refers to the above-
explained concept of transactions per second. La-
tency is the time it takes for a blockchain to create
a new block, append it to the blockchain, and regard
it as confirmed. It can be divided into two parts which
are the block time and the time to finality. The for-
mer refers to the time needed to create a block and
add it to the blockchain. In contrast, time to final-
ity can be once again subdivided into deterministic
and probabilistic. Deterministic means that a block is
considered final once it is appended to the blockchain.
In other words, the block is no longer changeable
once it has been published. Probabilistic refers to the
blockchains in which a block is still subject to change
once it has been added to the blockchain, i.e., due to
the network not having reached consensus on the fu-
ture state of the blockchain. Bootstrap time refers to
the time it takes to download a blockchain and con-
firm all the blocks and transactions. Costs per transac-
tion are external factors such as setup cost, hardware
cost, storage cost, and power usage.
2.2 Decentralization
Decentralisation is the central ethos and given nature
of the blockchain technology, but also a massive bot-
tleneck regarding scalability and security. It describes
the transfer of control and decision-making rights
from a central authority to a distributed network. A
characteristic of decentralisation in blockchains is the
distrust between its participants, which is desired and
required for it to work correctly.
Measuring a network’s decentralisation depends
on the type of blockchain. Two types of blockchains
exist or rather two types on how decentralisation must
be measured. One type uses the Proof-of-Work con-
sensus protocol, while the other type uses Proof-of-
Stake or a similar consensus protocol where the rights
to create a new block are given to a node based on
staked capital. The decentralisation (and security)
of a Proof-of-Work blockchain depends on the net-
work’s hash rate and how distributed it is. A net-
work’s hash rate is the cumulative hash rate of all
the (mining) nodes participating in the block creation
competition. Therefore, the higher the network’s hash
rate, the harder it is to disrupt it.Decentralisation of
a Proof-of-Stake or similar blockchain can be mea-
sured in the number of validators, the distribution of
staked capital among the validators, and the percent-
age of token supply that has been staked. Another
metric to measure decentralisation is the Initial To-
ken Allocation. It can create unfair advantages for
a group that receives many tokens and determine the
next block and chain governance. For both Proof-of-
Work and Proof-of-Stake (or similar blockchains), it
is important to measure how many nodes or pools (a
pool is a group of miners or validators which join to-
gether to increase their chance of creating the next
block) control the majority of the network. his met-
ric is also called Superminority or Nakamoto Coef-
ficient. The Nakamoto Coefficient is defined as the
minimum number of nodes required to get 51% of the
total capacity (either in computing power or staked
capital) (Srinivasan, 2017). However, for networks
with a lower Byzantine Fault Tolerance, it is only re-
quired to control one-third of the network’s comput-
ing power or staked capital.
A Platform Selection Framework for Blockchain-Based Software Systems Based on the Blockchain Trilemma
363