Thoughts on DDoS and Decentralization

Matters.News suffered a massive DDoS attack last night. Although Matters.News has encountered DDoS several times in its history, this time was the longest and the most traffic, lasting 9 hours before and after, with a peak of 50 million requests every five minutes.

The success of this attack also lies in its decentralization. The IPs that initiate the request are evenly distributed all over the world, which makes it impossible for us to block the attacker's traffic by geographical location. In the end, we can only distinguish real users and robots through CAPTCHA indiscriminately, so that users can access it normally.

There is a way of resisting DDoS, which is also decentralized, placing the services that process requests on edge nodes. Each request will be distributed to the nearest server, and even if the regional server where the attacker's traffic is concentrated is no longer responding, users in the remaining regions can still use it normally. This is essentially how Google, Facebook and other large services defend against DDoS, but small teams and services generally cannot afford the corresponding costs. Forcing the cost of a service up is itself a common and effective attack.

A more thorough form of defense is also a more thorough decentralization. If in a system, transfers are made through the full-node client of the blockchain, and content data is transmitted through IPFS Desktop or Planet , then there is no service that can be DDoS. The problem is that users don't want to use it, and we haven't been able to do it as well as Matters.News, although we tried hard.

DDoS attacks can show the usefulness of decentralization. Even if the Matters.News web page is inaccessible, it does not affect the assets on the blockchain and the content in IPFS. But this kind of "usefulness" is very rare. On the contrary, we have long learned that decentralization is not a requirement in most cases, and most users do not care.

Many people who understand technology will intuitively support decentralization, and even become an aesthetic or idea. There is also analysis behind this intuition, as decentralized networks are more connected, robust and scalable. Information systems also follow the principle of evolution. Those systems that can better convey and carry information for people will be more likely to survive, while those systems that are more likely to be disturbed and blocked will gradually fade out.

Surviving decentralized networks, trustless networks, and Web3 will eventually bring about new organizational forms and economic structures. This is a commonplace. These predictions are true, but for every project and team, they don't help. It is as if the survival and reproduction of each animal in the forest constitute the direction of evolution, but knowing the direction of evolution does not help the survival and reproduction of each animal.

I think back to the 1980s when the personal computer era was just beginning, and entrepreneurs needed to prove to users and investors that this was the future of the world. The only thing they can get is all kinds of spreadsheet software, one by one, the black and white, unusable early Excel. Teams scrambled to take over the spreadsheet market and needed to prove they didn't build a clunky calculator at an exorbitant price.

Now that personal computers are everywhere, everyone can see that the new world opened up by personal computers is far more than Excel. But in the midst of technological change, it is always full of blood and foam, and it is always impossible to hit mosquitoes with cannons. Those trends and ideals, grand and ethereal, have little to do with the creation of every moment. If you were in the 1980s, all you could do was get Excel done.