A venture-backed cryptocurrency with the promise to provide truly anonymous transactions is scheduled to launch in beta today, a move that will mark the latest in a detailed and expensive process to help ensure as many bugs as possible are removed before its blockchain supports real transactions.

Created from a fork of the bitcoin blockchain, Zcash is designed to cloak the addresses of both the counterparties participating in a transaction as well as the amount transacted.

If successful, the privacy-oriented, public blockchain could eventually form the foundation of an ecosystem of distributed applications built by both consumers and big banks looking for a more private means to transact.

Not only are the eyes of the cryptocurrency community closely watching this development, but so are would-be hackers looking for a high-value target.

So, to help ensure the soundness of this new protocol, the currency's creators, Zcash Electric Coin Company, are spending a quarter of their recent $1m venture capital investment to hire three separate auditing firms. The high-stakes environment where even competing cryptocurrencies might stand to benefit calls for a higher standard of due diligence, according to veteran cryptographer Zooko Wilcox, one of Zcash's founders.

In an interview with CoinDesk, Wilcox explained the difficulties of balancing the almost impossible task of perfectly audited code, with limited financial resources.

He said:

"The sad, unfortunate fact is that with a large codebase it's impossible to find all the bugs. When you’re doing this kind of security process, you have to choose a scope of what’s most dangerous."

To the address the problem, Wilcox focused the auditors' attentions on the changes his team has made to the bitcoin protocol's code.

After seven years of running without a hack, that part at least, can be set aside with at least some confidence.

In particular, Zcash narrowed its scope to six components, including the zkSNARK cryptography built on libsnark, the cryptographic construction of the "zk-SNARK circuit" and the Equihash proof-of-work algorithm.

"There isn't any way to look at a big codebase and know it is safe in general," said Wilcox. "You have to look at a big gray area that's more safe or less safe."

A moving target

The first step in performing an audit is to select the auditors. While this may seem obvious, the actual process of making the selection isn't always easy, as the collaboration requires a lot of trust and could include difficult conversations.

For Zcash’s first audit, which has been under way since August, Wilcox called in London-based NCC Group, a partner from a previous audit he conducted with his own security firm, Least Authority.

The publicly traded NCC Group's principal security consultant, Alex Balducci, was tasked to analyze third-party dependencies such a libsnark. Specifically, Balducci broke down the analysis into two categories: reviewing the implementation of the Zcash protocol and an audit of the source code.

Early conclusions of the audit resulted in multiple recommendations involving the way Zcash is developed. Specifically, he has advocated for the inclusion of tools to help identify coding issues during development.

"This process should be something that touches all aspects of a company," Balducci told CoinDesk. "Developers should have an awareness of the various security issues, policies should be set in place to enhance and adapt to changing security threats, audits should be performed and plans for worst-case scenarios formed."

Guns and Frappuccinos

Later this month, NCC Group will be joined by two other auditors in the process of helping minimize bugs and other vulnerabilities in the code.

Due in part to Argentina-based Coinspect's history of publishing "innovative" protocol designs, Zcash tasked the firm to validate specific threats, protocols and algorithms that only occur for cryptocurrencies.

The founder of the veteran security firm, which has audited implementations including Bitcoin Core, ethereum, monero, counterparty and bitcoinj says that cryptocurrencies prove an especially alluring target because some of the data at stake also has a corresponding token value.

CoinSpect's Juliano Rizzo compared the launch of Zcash with the launch of bitcoin. He said that when bitcoin launched, there were few if any people with the diverse skill sets necessary to hack a cryptocurrency — skills which he estimates include cryptography, familiarity with GPU-internals, awareness of ASIC-design, regulation, economics and social dynamics.

One strategy Rizzo said he looks for in his clients to help reduce the risk of theft is smart contracts that allow companies to store cryptocurrency in cold storage and that include reversible time-locked vaults so "illegally triggered" transactions can be reversed.

But even as defenses against hacks have become more sophisticated since the early days of bitcoin, so to have the attackers.

Rizzo said:

"To physically steal bags of cash from a bank in minutes, you need a gang with different skills, including firing guns and digging tunnels while pretending to sell butter cookies. A cryptocurrency theft can be carried on quietly by a single hacker enjoying a Frappuccino in a coffee shop."

The greater responsibility

Another auditor scheduled to begin work in September is Alexander Peslyak, better known as Solar Designer. His particular focus was on the Equihash proof-of-work algorithm.

In addition to being the founder and CTO of Openwall, Solar Designer is an advisor to the Open Source Computer Emergency Response Team that provides security support to open-source projects.

In interview with CoinDesk, Solar Designer explained the difficult task that other founders building cryptographically based startups face when trying to balance the nearly impossible task of creating a completely debugged codebase with a limited budget.

Solar Designer agreed with statements made individually by each of the other auditors that a perfectly debugged codebase of any "non-trivial" size "not only can’t be achieved — it can't even be defined."

Even with a $250,000 auditing budget, Zcash was forced to narrow the scope of its efforts to just those areas that weren't already largely debugged.

But for startups that aren’t funded or don’t have another source of capital, Solar Designer said that the level of due diligence required changes from project to project. In the end, he said it’s up to the auditors themselves to communicate the limitations of each project.

But that doesn't mean the diligence is optional.

"It is typical to adjust scope to budget, and the range can vary by an order of magnitude or more," he wrote. "Can't afford any? That's tough."

The 'half-life of doubt'

Zcash is scheduled to launch into beta today with all its features live.

But Wilcox is seeking to discourage the accumulation of any large amount of wealth on the blockchain between now and the full launch slated to occur on 28th October.

Even then, he said he hopes the growth-rate of the currency's value occurs slowly.

If a codebase is complicated enough, there truly are no guarantees. He calls it the "half-life of doubt", or the idea that every year that goes by without a hack his confidence increases – but it may never reach absolute certainty.

Even though Zcash is based on the as-yet un-hacked bitcoin code base, Wilcox said he’s not 100% confident it might not someday be compromised.

"The only thing that will make me satisfied," said Wilcox, "is that if years and years go by with more and more money."

Trial by fire

There’s two ways to ensure the security of a system.

The first is what security expert Bruce Schneier famously described as "security through obscurity" in a 2008 article. This, arguably, was the state of bitcoin's security through its early years, when few people knew it existed, and those who did are generally believed to have had the cryptocurrency's best interests in mind.

The second form of security though is what Wilcox calls "trial by fire."

Months before he hired auditors to take apart the code and look for weaknesses, the code was published on Github and the public was invited to search for bugs. As a result, multiple vulnerabilities were identified even before the formal audits began.

But the actual bugs exposed in a complex system like a cryptocurrency codebase are a factor of the actual bugs, and the exposure, or value, riding on the product, according to Wilcox.

Inviting additional outside auditors to review the cryptocurrency code accelerates that rate of exposure.

Wilcox concluded:

"You force yourself not to go through a trial by fire. And you want it to be seen by as many eyes as possible."

Disclaimer: CoinDesk is a subsidiary of Digital Currency Group, which has an ownership stake in Zcash.

Fire coals image via Shutterstock

AltcoinsPrivacySecurityzcash

To use radical new technology effectively, you have to be radical – otherwise, all you end up with is a retro-fitted version of the present system.

The benefits of the new technology are watered down or overwhelmed by the need to maintain the practices associated with the old system – many of which only exist precisely because of its inefficiencies. The new technology can even seem less efficient than the old one, simply because it isn't designed for use with processes from the past.

Blockchain and distributed ledger technology are radical.

Together, they could transform the way finance works, but people in the financial world are creatures of habit and old practices die hard.

So, if these technological advances are to transform finance, they need to be championed by key financial organizations.

It was exciting, therefore, to see a working paper from the Japan Exchange Group (JPX) that seriously examines the potential uses of blockchain and distributed ledger technology in capital markets. In conjunction with six other financial institutions, the researchers conducted two proofs-of-concept (PoC) covering securities issuance, trading, settlement & clearing and cash payment. They also looked at ownership, dividends and stock splits, data ownership and privacy.

From their PoCs, they conclude:

"DLT has the potential to transform capital market structure by encouraging new business development, improving operation efficiency, and contributing to cost reduction."

Sounds really encouraging, doesn't it?

The trouble is, anyone can say that. I did, in the second paragraph of this piece. I have produced no evidence to support my argument. And sadly, despite their PoCs, neither have the JPX researchers.

Instead, they noted that business practices have to change, then tried to shoehorn DLT into existing business practices.

Kept in the dark

Transparency is often believed to be desirable in financial markets, but it has serious implications for market structure. Lit markets are less open to fraud, but frontrunning and predatory behavior is common. Consequently, initiatives designed to improve transparency can encourage the growth of "dark markets".

This causes a problem for DLT. Eliminating the need for trusted third parties – which some would argue is the whole point of DLT – requires a blockchain on which transactions are public.

Unfortunately, rather than considering ways in which transactions could be made public without encouraging the growth of dark markets, JPX decided that a public blockchain was a non-starter. Financial market participants, they believe, would "never agree" to all transaction data being public.

So they compromised. They went for a permissioned DLT with trusted third parties they envision would be the same as the current trusted third parties – CCPs, exchanges and custodians.

I was left wondering what was so great about a distributed ledger when it simply distributes the functions of the present system over multiple computers. True, a distributed architecture would radically improve the system's resilience, unlike the present system where failure of a key participant can cause the entire network to freeze (think Lehman Brothers).

This is an important benefit of DLT, whether public or private, but I could not see how the claimed cost reduction or operational efficiency would materialize

Sacred cows

Next, JPX discovered that DLT would not work well in trading. So, they decided to concentrate mainly on using DLT for settlement and clearing, though there was also a discussion of ownership registry (which to my mind is a sensible use of DLT).

I've worked as a project manager in capital markets, and believe me, you cannot radically change settlement and clearing processes while leaving trading practices untouched. Changes have to be made end-to-end, or you end up with a reconciliation nightmare. Reconciliation differences between front- and back-office are a major source of cost and inefficiency in capital markets.

I can't see how introducing DLT in settlement and clearing alone would improve this. It could make matters a whole lot worse.

JPX restricted DLT to settlement and clearing because price discovery and order matching isn't efficient in a decentralized environment, and because traders often cancel or amend trades, which would play havoc on an immutable blockchain.

But why didn't they question these practices? Are there ways of pricing and matching orders without collecting lots of trades? Would introducing an immutable blockchain encourage better trading practices?

Traders are touchy creatures, but their working practices really can't be regarded as sacrosanct if there is to be radical change in capital market functioning.

Strange assumptions

A third fundamental problem, which the JPX researchers discuss but don’t really solve, is the trade-off between availability and capacity.

A third fundamental problem, which the JPX researchers discuss but don’t really solve, is the trade-off between availability and capacity.

Centralized systems can handle the speed and volume of transactions on today's financial exchanges better than decentralized systems, but at the price of loss of resilience. Discarding the public blockchain partially addresses this problem, because proof-of-work as currently conducted is far too slow for a high-volume transaction environment: verification by practical byzantine fault tolerance (PFBT) on a private or consortium blockchain is faster. But fully solving it really requires more technological advances.

JPX identifies potential new bottlenecks in a DLT environment, including the fact that running smart contracts slows everything down – a devastating finding for which they did not offer a solution.

But, the exchange didn't ask whether bottlenecks in the present system might also apply in a DLT environment. For example, know-your-customer (KYC) anti-money laundering (AML) regulations are currently applied by financial intermediaries. Strangely, JPX assumes that this would not change in a DLT environment, even though a key benefit of the DLT would presumably stem from the elimination of financial intermediaries if it were used for cash settlement.

There are other strange assumptions, too, such as the notion that a central bank might issue a digital currency for use on a private or consortium DLT, but wouldn't make that digital currency available on its own RTGS system – which would be a competitor to that DLT.

To be fair, this study aims only to prove that the technology has potential for use in capital markets. And although JPX concluded that much more work is needed, they do succeed in this aim, despite the compromises.

Now, there needs to be a second evaluation, which looks at the changes that would be needed to capital market functioning to enable DLT to deliver genuine improvements in cost and efficiency.

The real benefits from DLT will come not from re-engineering capital markets as they currently function, but from re-imagining capital markets for a radically different future.

Boy in a big suit image via Shutterstock

Disclaimer: The views expressed in this article are those of the author and do not necessarily represent the views of, and should not be attributed to, CoinDesk.

AsiaBankingJapan Exchange Group