What’s Gone Wrong with Bitcoin and Can It Be Fixed In Time?
Can blockchain technology become a key building block in the future of computing? A look into cryptocurrency computing costs and challenges, the evolution of blockchain, and what it means for cloud computing and the Internet of Things (IoT).
There is no topic more controversial these days than the future of Bitcoin. Supporters claim it will change the future of finance and transform society. Critics, like the Indian Finance Ministry, claim it’s no better than a Ponzi scheme. Who’s correct?
It’s literally a Trillion Dollar Question, so a lot of people are interested in the answer. While we won’t really know the final answer for years, I believe an understanding of Bitcoin’s implementation of blockchain will shed light on what’s to come and the future of cryptocurrency.
What’s Gone Wrong?
With the tremendous Bitcoin run-up in 2017, and the associated increase in transaction load, the network is now showing deep signs of stress. This in turn is impacting people’s impressions of the digital currency and its potential for the future.
Morgan Stanley analyst James Faucette recently published a paper where he analyzed various methods to try and value Bitcoin and determined the value might be zero! Why? Well there are several considerations, but one of the biggest is its lack of acceptance as a method of payment. That’s easy to understand. Digital currency might be a great idea, but if no one uses it for purchases, the utility is pretty limited. Faucette goes on to say that Bitcoin has “virtually no acceptance, and shrinking.”
With all the hype, why aren’t more people using it for purchases? A major reason is that transaction fees, and transaction completion times, have skyrocketed. Even at its best, Bitcoin is not a speedy transfer mechanism. It is designed to take about 10 minutes to confirm any given transaction. However, in practice it’s gotten much worse recently.
As just one example of the problems with the Bitcoin network, my daughter Alexa thought it would be fun to give some Bitcoin to her grandfather — a retired banker. So, on Christmas Eve, we transferred about 0.002 Bitcoin (worth about $25 USD) to an account we created for him. In order to move that $25 we wound up paying $32 in transaction fees to the network and it took hours to confirm!
And, all this is going on while the Bitcoin network continues to use a huge amount of power — more than some entire countries! What’s going on? It turns out this is a direct result of the scalability issues in Bitcoin’s blockchain implementation. It should be no wonder that people aren’t adopting this digital currency as a payment method.
Understanding the Bottleneck
In one very important way, Bitcoin has been transformative to computer science. The underlying blockchain technology offered the first practical solution to what’s called the Byzantine General’s Problem and this allows Bitcoin to support a peer-to-peer, distributed ledger with no central authority. Blockchain is what makes Bitcoin different from a centralized payment network like Visa.
However, the cost of solving this problem of a true distributed ledger is tremendous. The secret sauce is in the creation of a network of miners. Miners are computer processes that verify the integrity and security of each transaction going to the distributed ledger. Anyone can add miner nodes to the network — without permission from anyone! Miners are paid out in Bitcoin in exchange for providing this service. It’s a core part of the Bitcoin ecosystem.
The big question is, if anyone can add miners, how do you keep out bad actors, which might falsify transactions or disrupt the ledger? Part of the answer is that many miners must agree on the evaluation of a transaction. However, the other, more interesting part of the answer is a concept called proof-of-work.
Without getting too mathematical, Bitcoin’s mining network requires each node that is helping to verify transactions provide proof-of-work in the form of trying to solve some extremely compute intensive puzzles on a continual basis. You’ll often see these puzzles more formally referred to as hashing. These hashing operations actually have little to do with the individual transactions flowing through the network or the creation of new coins — as is often assumed. It’s busywork!
This busy work is required to make it expensive to mine. If mining were trivial, then a bad actor could flood the network with a large number of small, cheap miners and take it over. Proof-of-work is designed to make that an economic infeasibility. It should cost more to fake a transaction than you could profit from doing so.
Taking us back to Bitcoin’s scalability problems, we now have the heart of the issue. Many miners are required to be involved throughout each transaction to ensure a single bad actor cannot falsify a transaction. Proof-of-work makes it economically infeasible to flood the network with fake miners.
This combination has worked well for most of Bitcoin’s decade-long history, but it’s reached a breaking point. Spiraling transaction fees, slow transaction completion times and a reputation as a growing factor in global warming, are the result. This isn’t sustainable.
What Can Be Done?
The cryptocurrency community isn’t blind to these challenges and there are numerous plans to attack this core issue. They range from relatively tactical changes to radical proposals that would obsolete blockchain technologies all together and move to other methods of distributed quorum.
Let’s examine some of the interesting things going on that show possible solutions to these issues.
Litecoin’s Lightweight Approach
It turns out that people have seen this coming for a long time. As far back as 2011, poor performance of the Bitcoin network was a hot topic. A group called Litecoin forked the Bitcoin community and created its own cryptocurrency. One of the core reasons for the fork was the desire to make changes to the core blockchain implementation that would make transactions faster.
Today, Litecoin has been successful and it one of the largest cryptocurrencies as measured by market capitalization (see Coinmarketcap.com). Also, its current performance is substantially better than Bitcoin’s.
As a test, I sent about $25 USD worth of Litecoin from my account to my daughter’s account. The total transaction fees were $0.06 (six cents!) vs. the $32 we paid to transfer $25 worth of Bitcoin to her grandfather. And, the Litecoin transaction completed in a manner of about ten minutes.
There is pretty dramatic difference when you compare the performance and fees of Bitcoin and Litecoin. So, does Litecoin have it right?
Upon deeper inspection, I’m inclined to say not yet. First, while a ten-minute completion time sounds great when it’s compared to the hours my similar Bitcoin transaction took, it’s still terrible when compared with a dedicated payment network like Visa. Do you want to wait 10 minutes at the grocery store or the fuel pump while they confirm your Litecoin payment? No.
Secondly, the Litecoin network is still much smaller than the Bitcoin network — with about one-tenth the overall transaction volume (measured in dollars) each day. While there are clearly improvements in the Litecoin blockchain implementation, the core change should make it about four-times faster than Bitcoin — not a hundred-times faster.
I believe that if Litecoin scaled up to the same size as Bitcoin that many of the same scaling issues would show up. It’s core technology is not that much different from Bitcoin.
Ethereum’s Proof-of-Stake Bet
Ethereum is generally thought to be the second most popular cryptocurrency — behind only Bitcoin. Ethereum has many novel features that differentiate it from Bitcoin — such as built-in smart contracts. And, it enjoys a good reputation for efficiency.
However, it has started to suffer as it scales up too. Duplicating my earlier experiments, I sent $25 worth of Ether to my daughter. It cost me $0.31 in network fees. That is one hundred-times less than the $32 from Bitcoin. Also, the Ethereum transaction took about 1 hour to confirm the transaction.
Here’s a picture showing the status of my Etherem transaction about one hour after being sent. As you can see, it’s been confirmed by 595 miners, but still hasn’t fully closed. Imagine that! Nearly 600 computers had worked to confirm my little $25 transaction while doing heavy-duty hashing gymnastics. No wonder these networks are burning through power!
The chart below (source: BitInfoCharts.com) shows how both networks have been struggling to keep up this year as the cryptocurrency rage increased. And, you see them both suffering even more as we reached the frenzy in December.
The Ethereum community has proposed an ambitious solution. They have proposed to replace Bitcoin’s original proof-of-work fraud protection with a new method called proof-of-stake. Proof-of-stake replaces the make-work mathematical puzzles at the core of Bitcoin’s scaling problem with a new idea.
Ethereum’s project to switch to this new miner consensus technique is called Casper. With Bitcoin’s proof-of-work implementation, the only punishment for attempted fraud is for that anonymous Miner node to be kicked off the network.
However, Casper changes the game. Anyone can still add a miner to the Ethereum network, but in order to take part, they must stake some currency. If a miner is caught attempting to participate in fraud they lose the money they staked to participate. But, if the miner participates as expected then they’re paid out interest on the staked amount as a reward.
Remember, proof-of-work was designed to make it economically infeasible to defraud the network by flooding it with small, corrupt miners. Proof-of-stake creates disincentives for trying to defraud the network in a much more direct way.
By moving away from proof-of-work’s endless hashing, Ethereum could make their network dramatically more efficient — and eliminate the global warming stigma we’re seeing around cryptocurrency. However, this code is still in development and won’t see wider deployment until later in 2018. We’ll need to wait and see how it works.
Lightning’s Payment Channels
Changing the underlying assumptions of blockchain under Bitcoin would be daunting and even dangerous. Rather that optimize Bitcoin’s blockchain, which is currently tracking about a quarter-of-a-trillion USD worth of Bitcoin, the Lightning Network is looking to create a high-performance payment network as an overlay to the underlying Bitcoin network. The project’s stated goal is to create, “Payment speed measured in milliseconds to seconds.” This would be more competitive with traditional networks like Visa.
The basic concept is that an off-blockchain account can be set up between parties that allows for many small payments. These payments are rolled up and committed to the blockchain much less frequently — thus avoiding the heavy transaction fees and delays.
Can the Lightning Network fix Bitcoin’s issues? It’s still very much to be seen. The network is still in very early testing. There are many critics who point to issues with Lightning that must be answered. These include issues around centralization, trust, and ironically, scalability. As Lightning rolls into production we will get a better picture for how it could help.
IOTA Rethinks Everything
IOTA is a relative newcomer in the cryptocurrency space — having been originally proposed in 2015. However, it’s grown rapidly and is now one of the top 10 largest cryptocurrencies by market cap.
IOTA’s name is actually a play on the Internet of Things (IoT). It’s designed for a world where billions of connected devices may need to exchange services and will need an ultra-high performance, zero-fee network. Think of it as a machine-to-machine micropayment service. This is ambitious, and goes beyond the scalability of even today’s centralized networks like Visa.
IOTA’s central innovation is called the tangle. The tangle uses a new and novel technology for distributed ledger instead of blockchain. In that way, it is fundamentally different from the other cryptocurrencies discussed here such as Bitcoin, Litecoin and Ethereum. IOTA’s tangle removes the need for miners all together — and mining is at the core of the scaling issues we’re seeing in the other large cryptocurrencies today.
Rather than depend on a network of anonymous miners to verify transactions on a blockchain, IOTA uses a different approach. Since there are no miners, there is no need for transaction fees to be paid to them. Instead, each party issuing a transaction must contribute the compute resources to verify two previous transactions from other parties. Everyone using the network is ensuring the security of everyone else’s transactions. If the concept proves out it will be revolutionary.
IOTA is very immature compared to the other, older currencies discussed in this article. With the small number of transactions currently flowing through the network, the tangle mathematically cannot offer enough protection from fraud. Thus, the system depends on central Coordinator process that is run by the IOTA organization to keep things moving. As IOTA grows they plan to shut down the Coordinator and let the tangle run on its own. However, until then IOTA isn’t a truly decentralized currency and we can’t really know how it will perform at scale. Time will tell.
A Comparison
Bitcoin runs the risk of losing its leadership position in cryptocurrency if it cannot fix its problems around performance, scale and cost. And, these problems are at the very core of Bitcoin’s implementation. Can it work past this?
By analogy, I’d like to look at a similar story from my own past. In the 90s, I became an early member of the team responsible for developing the Java platform. At the time, Java and its underlying Virtual Machine technology had terrible reputations for performance and scale.
However, the fundamental value proposition of portable software was undeniable. This meant that people decided it was worth the effort to make it work better — and I wound up founding the Java performance team at Sun Microsystems to help make this a reality.
An arms race ensued, with multiple companies and projects competing, and the performance of Java got better rapidly. Today, Java is still the number one programming language in the world, it’s used at the center of thousands of Enterprise business applications and its direct descendant, called Dalvik, sits at the heart of over 2 billion Android devices throughout the world.
The Verdict
The core idea of digital currency, run by a decentralized distributed ledger, is an idea that is here to stay. The world has seen that distributed ledger technology can work and can process billions of dollars per day in transactions — even at this early stage of development.
Of the ideas outlined above, some will succeed and some will fail. Many other new ideas, and entirely new currencies are still to be created. The genie is out of the bottle and we will see these architectures begin to impact cloud computing and IoT architectures — just as we are seeing them impact the financial services landscape.