Developed for several years thanks to the Bitcoin (BTC) protocol; the mainstream solution to
build trust on a network of untrusted individual nodes has been based on the idea that having
to possess sufficient computing power to overrule most of the other nodes would be too
expensive to achieve. This is, in very short, illustrates the concept of the Bitcoin style Proof
of Work (PoW) consensus building mechanism. Basically, for any given wannabe attacker,
the more time passes, the more work is performed, the more difficult it gets to cheat. Soon, it
becomes economically and physically unfeasible to overrule a large enough network.
Underlying the Proof of Work (PoW) approach is the “work” aspect. Work requires energy,
capital investments, human efforts, space, opportunity cost, and so on. With a growing
Proof of Work (PoW) based network, the desirability of attacking it grows as the market
capitalization grows and so; the processing power required to keep the network safe and
secure is exponential. The safety and sustainability of this approach has been questioned.
Obtaining a majority of the total computing power is not outside the range of large players,
most notably Ghash.io and certain other players that could obtain sufficient cryptographic
hashing power today.
There are other alternatives to the risky and unsustainable PoW consensus building
mechanism. Nxt(coin) has chosen a different and more secure path to obtaining consensus,
called “Proof of Stake”. The concept is not based on computing power, but rather on “being
part of the system”. Nodes will trust each other based on the amount of “stake” each other
has. The underlying idea being that the ones with the most stakes, are also the ones with
the least interest in cheating the system. This approach, at least in Next full Proof of Stake
implementation, has at least three major theoretical benefits: Transparent forging, instead
of mining, reduced long-term transaction fees and decreased likelihood of the so-called
Although it would have to be demonstrated, Proof of Stake (PoS) seems to have been first
proposed in July 2011 by a forum user called QuantumMechanic. At least two implementation
specifications are known to have been released since then, one by Meni Rosenfeld and one
by Cunicula. At least two Cryptocurrencies are known to use a Proof of Stake consensus
building mechanism; one is Peercoin (PPC) which is a combined Proof of Stake / Proof of
Work implementation released in August 2012. Nxt is the first completely Proof of Stake
driven currency implementation and was released on Nov 24th
Transparent forging, predictability and
Similarly to mining in Bitcoin terminology, preparing a block containing the transactions,
sending it officially on the network for inclusion in the block-chain and receiving a financial
compensation for this task is called forging in Next cryptocurrency ecosystem. To understand
Transparent Forging one first needs to understand the forging process itself. The likelihood
of forging a block for an actively forging account is proportional to its amount of NXT
compared to all active forging NXT in the network.
While some long-term lack of precision will avoid the possibility of specific attack schemes
such as shuffling attack, the near-future forging node should be known as precisely as
possible to allow for efficient network utilization. Transparent forging implies that any node
can predict with a decent probability which one of them will generate the next block, and when
it will be done. The target expected to forge the next block is based on the Base Target (same
for all nodes and accounts), the time since the previous block was forged (again, same for all
nodes and account) and the account effective balance, following the simple formula:
Target = BaseTarget * TimeSinceLastBlock * Balance.
If your account is the lucky target coming out of this formula, you’ll generate the next block.
As a consequence of all this, the transactions can be sent directly to the node(s) expected to
forge the following block(s), provided it’s publicly listed, saving considerable time and effort in
transactions transmission and block elaboration. The network could also detect which nodes
aren’t taking part in block generation itself and act accordingly, which is a fence against some
specific attack schemes.
Today, transaction fees are only a fraction of the total block reward. In the coming years
though, transaction fees will form the vast majority of each block reward. The Bitcoin Proof of
Work system will force miners to require a “minimum” transaction fee to be profitable. Even
with a reduction in complexity, miners must utilize physical resources in order to mine blocks.
It is likely that the mining industry will then consolidate into a few very large players, who will
drain most, if not all, the transaction fees to them, in exchange for sufficiently large computing
power. Proof of Stake does not have such a limitation. Any client connected to the network,
regardless of its computing power, can forge a new block. There will be no large investments
required from him. The result is that transaction fees will necessarily be lower than in any
Proof of Work-based system.
In the Bitcoin Proof of Work system, an attacker could obtain over half the current network
hashing capacity in order to have a chance to perform successfully the so-called “51%
attack”, allowing him to perform double spend transactions for example. While the
mathematics might prove it to be unlikely and expensive enough, surreptitious control of the
Bitcoin network remains achievable.
Proof of Stake is designed to be a more robust system than any Proof of Work system. A
consensus altering attack would require the purchase or procuration by any other means of
a large, majority stake in the system. First of all, market price discovery makes it impossible
for an attacker to do this undetected. Second, performing this attack by placing buying orders,
would likely influence the market price of the currency. In the event that one could obtain
control of such a large amount by other means (loans, borrowing, agreements), performing
the attack would become economically useless, as it would in return decrease the value of
the currency, and so decrease the total value the attacker possess by far more than what the
attack would potentially bring.
Other types of attacks are possible; think of account shuffling for example. Some are, or will
be, inherent to the Proof of Stake system. But none have the risk magnitude of the 51%
attack today, and all are fenced by the implementation in place. With broader adoption, it
is likely that new attack schemes will be discovered, but by linking the forging power to the
stake an attacker needs to have, the system ultimately favors honest and trustful behaviors by
endangering the market value of the stake the attacker puts at risk.
While the Proof of Work consensus building mechanism has obtained large community
acceptance and is today the prevalent model, it is clear that this system has several critical
flaws endangering its future development. Nxt(coin) is offering an elegant alternative, having
features no other Proof of Work based Cryptocurrencies have today. Reduced network
reliance on computing power, forging system, reduced network traffic, increased predictability,
reduced exposure to known attack schemes are some of the large improvements offered by
this new approach.
Author: frmelin Proofreader: Brian Snyder