Page 30 - FIGI: Security Aspects of Distributed Ledger Technologies

ensure in this context keep all these transactions looking for a specific transaction ID (but not seman-
on-chain and the settlement near instantaneous. tic equivalents) that a transaction had not completed
• Greater certainty around the concepts of settle- when, in fact, it had.
171
ment and settlement finality applied to crypto-as-
sets is needed. Risks:
• Use of the transaction assurance, for example By deliberately launching transaction malleability
insurance of custodians attacks on multiple exchanges at once, perhaps using
• There may be a need to distinguish between per- software deliberately designed to create mutant
missioned and permissionless DLTs in that respect, transactions could cause short-term problems for
in particular, specific governance issues with per- the market as any uncertainty or doubt about market
missionless DLTs, which makes them less suitable stability will have an effect on market prices, espe-
to the processing of financial instruments, at least cially with such an illiquid, volatile asset class.
in their current form. 161
• Central Bank DLT prototypes have used the BFT Mitigation and Recommendations:
consensus protocol to ensure finality of pay- Cost-based prevention, e.g. consensus algorithms
ments. 162 make it expensive to perpetrate this attack.

8.3.2 Issue: Changes In The Order Of 8.3.3 Issue: Accuracy of Oracle Input/Output
Transactions Data

Dimensions Affected: Consensus, Data Model Dimension Affected: Data Model

Specific Threat: Transaction (Data) Malleability Specific Threat: Oracles are compromised
A transaction (data) malleability attack lets some- Blockchain applications are unable to directly access
one change the unique ID of a Bitcoin transaction and retrieve information from sources outside of the
before it is confirmed on the Bitcoin network, making blockchain. An oracle serves as a conduit between
it possible for someone to pretend that a transac- an external data source and blockchain applications,
tion didn’t happen. The goal then is to deceive a such as smart contracts and DApps. 172
163
merchant or payor into paying twice for the same In contrast to the blockchain philosophy which
transaction by leading the target into believing that mandates operation in a decentralized, trustless
the original transaction failed. The founder of Mt. environment, using an oracle introduces both a trust-
164
Gox claimed that transaction malleability was a ed intermediary and trusted data source with the
primary cause of the spectacular heist of USD 473 possibility both will be provided from a single, cen-
million of Bitcoin stolen from the exchange. The tralized source.
165
claim was analyzed and separately confirmed as a
problem in the Bitcoin protocol, currently fixed in Vulnerabilities:
166
a soft fork and in the SegWit solution (which is still Corrupted data is seeded into/out of DLTs via inse-
167
not fully adopted within the Bitcoin network) as cure oracles
168
well as the Lightning Network. 169 While oracles generally provide critical input and
output capabilities for data on a DL, they are also the
Vulnerability: weakest link as they are not secure. They may give
The vulnerability lies mainly with DL protocols rise to greater opportunity for liability and damag-
such as Bitcoin (and Litecoin) which use transac- es if faulty data is used and there are losses, which
170
tion identification (‘TXID’) in the process of send- could precipitate a damage claim. 173
ing funds, meaning that instead of withdrawing a Oracles require trust both regarding the ora-
value from an account, the Bitcoin protocol points cle itself (as a trusted intermediary to a blockchain
to a prior input (the ‘deposit’) which is the source application) as well as from the data sources them-
of where an address received funds to match to the selves. An oracle is vulnerable to the presence of bad
existing output (the ‘spend’). The problem allows behavior that occurs at/from its data source and
for the transaction identification to be changed to could impact what occurs on the blockchain,
a variation that is a semantic equivalent before the
original transaction is confirmed on the network. This
lends the appearance to the sender, who may be only

28 Security Aspects of Distributed Ledger Technologies