Major Characteristic #1
The Lightning Network utilizes HTLCs (hash time locked contracts), which prevent users from being able to obtain their funds at their sole discretion (outside of the parameters and restrictions of the HTLC).
In the passage cited above, one of the main constraints of the Lightning Network is outlined plainly (while being cast as an innovation in the paper).
Since LN is a state channel, rather than a side channel, it must work around the limitations of the Bitcoin protocol entirely in order to facilitate the protocol’s (LN) functionality.
One example of this can be found in the fact that only SegWit bitcoins can be used on the network because they are the only UTXOs that are resistant to potentially being exploited via transaction malleability.
Above is a brief description of transaction malleability. As noted in the prior screenshot from the Lightning Network whitepaper, transaction malleability would obviously significantly impact the integrity of the HTLC that the LN state channel is dependent upon because a change in the transaction’s hash would essentially render the HTLC invalid entirely.
Thus, the adoption of the Lightning Network will always be intuitively tied to the adoption of SegWit transactions on the Bitcoin protocol (if it is to remain a safe and uncompromised option).
Recent data pictured above shows that only above 41% of all network transactions are SW outputs.
The vast majority of the network’s wallet addresses are legacy as well.
This has led to Bitcoin developers considering more extreme measures to push Lightning Network’s adoption to the forefront.
Obviously these measures have been met with great resistance from the community.
After all, there is no guarantee that the network will be used as much as the Bitcoin developers propose that it should, even after such a soft fork is made to the protocol.
In addition, this type of soft fork (like SegWit), would be entirely irreversible unless the protocol were hard forked.
When considering the extremes that the Bitcoin development community must go to in order to facilitate the implementation of LN, it stands to reason that there may be better alternatives.
Other Downfalls of the Lightning Network
Confused about how Lightning payments really work? This article explains the hash- and time-lock contracts that make LN “trustless” with simple diagrams. At the end, we reveal Lightning’s dirty little secret: small payments aren’t “trustless” at all. https://t.co/Yhg3Mk2Rfi
QUIZ] Lightning channels are beads on a string stretched between two people. For the network below, what is the maximum amount Alice can send to Frank in one LN transactions? Assume 1 bead = 1 coin, ignore fees, and AMP is not supported. Answer in the poll below.
Like @el33th4xor said, LN coins have position-dependent value. The coin Bob holds with Carol is worth more than the coin he holds with Alice. The former coin he will likely spend; the latter he will likely not. If on-chain fees are $10, the coin with Alice is worth ~$10 less.
QUIZ + BONUS] The Lightning Network illustrated below consists of 8 users each with 3 open channels. All channels contain 3 coins. Assume TXs adhere to the LN V1.0 protocol and ignore fees. By how many coins can the right side’s combined wealth increase using LN payments alone?
Of course, the criticisms above are just a few of the pitfalls of the Lightning Network.
This all does not include the notable limitation of 16 million satoshis per channel (approx. .16 $BTC).