Page 56 - ITU Journal Future and evolving technologies Volume 2 (2021), Issue 1

ITU Journal on Future and Evolving Technologies, Volume 2 (2021), Issue 1

ties. This attack is not necessarily about depleting hop transfer. The directional capacities in the payment
the channel capacities but guessing the channel ca‑ channels are not advertised but the total capacity in the
pacity of a node. Continuously learning the chan‑ channel is known for a sender to calculate a path. This
nel capacities will eventually yield more complicated provides a partial channel balance privacy. The sender
privacy attacks as discussed in the attack on sender encrypts the path by using the public keys of the interme‑
privacy. diary nodes by utilizing “onion‑routing” so that the inter‑
mediary nodes only know the addresses of the preceding
• Relationship Anonymity. In some cases identities and the following nodes. None of the intermediary nodes
of or may be known. This is a very valid can guess the origin or the destination of the message by

case for retailers because they have to advertise their looking at the network packet.
identities to receive payments. However, if an at‑
tacker can relate the payer to the payee, not only the 4.3.2 Raiden network
spending habits of the sender but also the business
model of the recipient will be learned. In such cases, Shortly after LN, Ethereum foundation announced Raiden
the privacy of the trade can be preserved by hiding Network [5]. Raiden is the equivalent of LN designed
the relationship between the sender and recipient. for transferring Ethereum ERC20 tokens and provides the
Speci ically, who‑pays‑to‑whom information should same privacy features. Although Ethereum is the second‑
be kept private. Some of the PCNs utilize onion‑ largest cryptocurrency, that popularity is not re lected
routing to forward the transactions to the destina‑ well in the Raiden Network. As of June 2020, Raiden
tion node. Onion‑routing is a source‑routing proto‑ serves with 25 nodes and 54 channels. The advantage of
col where the source of a message encapsulates the Raiden over LN is, due to tokenization, users can gener‑
data with the keys of the intermediary nodes like a ate their own tokens to create a more lexible trading en‑
stacking doll. An intermediary node can remove only vironment.
one layer from the incoming message to see the next
node to which the data is to be forwarded. Hence, in 4.3.3 Spider network
a distributed network, an intermediary node will not
con idently be aware of who talks to whom. Spider network [16] is a PCN that proposes applying
packet‑switching based routing idea which is seen in tra‑
• Business Volume Privacy. For a retailer, publicly ditional networks (e.g., TCP/IP). However, it is known
disclosed revenue will yield the trade secrets of its that in packet‑switching the source and the destination of
business, which must be protected by the PCN. In the message should be embedded in the network packet.
that sense, the privacy of every payment is impor‑ The payment is split into many micro‑payments so that
tant. Such payment privacy can be attacked as fol‑ the channel depletion problem gets eliminated. The au‑
lows: In a scenario where two or more nodes collude, thors also aimed to have better‑balanced channels. In this
the amount of a transaction can be known to the at‑ PCN, there are spider routers with special functionalities
tacker. In another scenario, if the recipient is con‑ that communicate with each other and know the capac‑
nected to the network via a single channel through ities of the channels in the network. The sender sends
the attacker, then it will track all of the lows towards the payment to a router. When the packet arrives at a
the recipient. router, it is queued up until the funds on candidate paths
are satisfactory to resume the transaction. The authors
4.3 State‑of‑the‑art PCNs and their privacy do not mention privacy and plan to utilize onion‑routing
evaluation as a future work. The micro‑payments might follow sep‑
arate paths, which would help to keep business volume
In this section, we brie ly describe current studies that private if the recipients were kept private. Additionally,
either present a complete PCN or propose revisions to the hijack of a router will let an attacker learn everything
the current ones, then analyze their privacy capabilities in the network.
based on our threat model. We provide a summary of the
assessment of the current PCNs’ categorizations and pri‑ 4.3.4 SilentWhispers
vacy features in Table 1. Although our main interest in
this paper is speci ically payment channel networks, for SilentWhispers [9] utilizes landmark routing where land‑
privacy in permissionless blockchains, the readers are ad‑ marks are at the center of the payments. In their attack
vised to have a look at [7]. model, either the attacker is not on the payment path
or a landmark is HBC. Here, landmarks know the topol‑
4.3.1 Lightning network (LN) ogy but they do not know all of the channel balances.
When the sender wants to send money to a recipient,
LN [4] is the irst deployed PCN that utilizes Bitcoin. It she/he communicates with the landmarks for her/his in‑
started in 2017 and by June 2020 serves with more than tent. Then landmarks start communicating with the pos‑
12.000 nodes and 36.000 channels. Nodes in LN uti‑ sible nodes from “sender‑to‑landmark” to the “landmark‑
lize “Hashed Time‑Locked Contracts” (HTLC) for multi‑ to‑recipient” to form a payment path. Each node in the

40 © International Telecommunication Union, 2021