The above image was made with stable diffusion using the prompt 'interplanetary file system.'
Over the last five years I've written something like a million words. This sounds like a lot, but getting there required writing an average of just 548 words per day every day for five years. About half of what I've written is science fiction. The rest is nonfiction, often on grim topics. So far this feels like a healthy balance between fantasy and reality.
During this time, online censorship has continued to intensify. I've definitely been censored by Facebook and probably Twitter as well. I've also noticed much more apparent search results manipulation by Google. This situation will probably get worse before it gets better.
Hive of course can't be censored, though Hive content/links can be made invisible by the algorithms used by all of the popular sites. Interplanetary file system (IPFS) content can't be censored either. At least, not at the content level. At the network level, according to a recent paper, it may be possible for a malicious actor to make IPFS content impossible to discover. From the paper:
First, we present a content censorship attack targeting the main IPFS DHT-based resolution system. The attack relies on strategically placing Sybil identities in the network so that they replace honest resolvers for a given CID. As a result, downloaders cannot discover provider records for the target CID and are unable to download the content. The attack can be performed from a single, resource-constrained machine at very little cost ($4 using AWS) and makes the provider records unavailable after a time that ranges from a few seconds to up to 48h depending on the initial setup. Currently, IPFS has no mechanisms to counter the attack, threatening the security of systems using IPFS as a storage platform.
Note that IPFS uses content identifiers (CIDs) instead of network addresses to locate content. Network addresses are mapped to content using distributed hash tables (DHTs) which are targeted in this attack. Fortunately, the paper's authors also figured out how to defeat such an attack:
We introduce a mitigation technique that allows us to reliably discover provider records regardless of the number of Sybil nodes placed by an attacker around the target CID. The mitigation replaces the regular put and get DHT operation by hash space region-based queries. Using these, providers always find honest resolvers to store their provider records and querying nodes always discover these honest resolvers and receive true provider records. While introducing an overhead linear in the number of Sybil nodes placed close to the target CID, this mitigation is only enforced when suspicions exist about the existence of an attack, as indicated by our detection mechanism.
The paper suggests that the described attack detection and mitigation measures will be incorporated into IPFS in the near future. For now, node operators can implement these measures manually. Personally, I use Pinata for all of my IPFS storage and gateway needs, and I trust them to stay on top of things like this. But the paper has implications for Ethereum, which uses a similar DHT scheme, and it's not clear how risk to that network will be handled.
In the meantime, I'll just keep writing. A new article of mine appears in the latest issue of International Paneling. It's about the regulator attack on crypto that has set us back years. In my opinion, what would get us headed in the right direction again is an exchange willing to ignore regulators and serve US customers. As risky as this would be for the exchange, it would also be immensely profitable, so I'm hopeful that someone will eventually try it.
Read my novels:
See my NFTs:
