David Vorick, co-founder of Sia, a blockchain based data storage solution, has written a blog in which he declares 2019 as "the year of the 51% attack." He goes on to explain both why this is becoming an issue now as well as some ideas as to what can be done about it.
Vorick explains that the game-theory model that secures Bitcoin's network is based around something called "incentive compatibility." He explains the idea quite eloquently:
"Bitcoin developers strive for something called incentive compatibility. If a protocol has incentive compatibility, it means that the optimal decision for each individual from their own perspective is also the optimal decision for the group as a whole. When protocols are incentive-compatible, individuals can be completely selfish because those selfish actions will benefit the group as well."
He elabporates that this has worked out quite well within the Bitcoin network because anybody attacking the chain to get Bitcoin basically would be destroying the value of their own coins. This however breaks down when the model expands to an altcoin ecosystem that sees many chains using the same algorithms and by extension the same hardware to secure the network. Add onto this the rise of mega mining farms and hashrate marketplaces and you have a recipe for what just happened to Ethereum Classic.
In basic terms, you lose the "incentive compatibility" this way because someone can, for example, use the same hardware from one coin to attack a different chain, a chain they don't already have an investment in. So they do some double spends, sell whatever they can get, and even if it tanks the network it won't hurt them financially.
In a similar vein hashrate marketplaces make the cost of attacking a network significantly lower than ever before. Previously, an entity would need to purchase, set up and maintain enough miners/nodes themselves to pull off such an attack, which would be extremely expensive in almost any case. However, when you can rent hashpower from a farm for just a few hours with just a matter of clicks, the price/reward ratio becomes tantalizing again.
There are many other factors discussed, and the post is well worth a read by anyone concerned about these attacks, but Vorick isn't just pointing out problems, he also suggests some solutions. These include improving means of tracking hardware availability worldwide, which would help exchanges assess if a threat to their network is feasible or not.
He also suggests exchanges creating stronger relationships with mining farms/hashrate marketplaces so that reasonable limits can be put on who can rent hashpower and how much, along with monitoring what it is doing.
One other idea is putting automatic processes in place that would halt trading/blacklist addresses as soon as a block reorganization is detected. This could dramatically reduce the time attackers have to move coins before the network responds.
Again, there is much more detail on all these idea and more in the original post, so go check it out. In the meantime, stick right here with Chepicap for any and all updates on the "year of the 51% attack."