3 ms·
While I get their reasons, I think it's disgusting that someone is running critical infrastructure out of what amounts to a centralised (single point of failure
by selfhoster11 4y ago
While I get their reasons, I think it's disgusting that someone is running critical infrastructure out of what amounts to a centralised (single point of failure) consumer communications app being run by a social media. Something like this should only ever be a stopgap measure until proper, dedicated, supported infrastructure is put in place.
- sneak 4y agoWhat makes you think WhatsApp is centralized or has any single points of failure? What makes you think that dedicated infrastructure run by the train people would be less centralized or more reliable than that run by Facebook?
- anthk 4y agoCustom LOCAL XMPP servers could do the same. Whatsapp has had connection issues globally.
- selfhoster11 4y agoI'm sure the WhatsApp backend is redundant and HA and all that jazz behind the scenes. WhatsApp as a service, however, is not. Facebook could discontinue it, pivot it to something else, start making undesirable changes, or sell it to someone who would make changes like that. This wouldn't happen with a system that's under the subway system's own management. > What makes you think that dedicated infrastructure run by the train people would be less centralized or more reliable than that run by Facebook? For one, it avoids all of the issues above. For two, it is much, much easier and cheaper to achieve enough nines for a couple thousand users than for a couple billion users. For three, they don't need to have the expertise to run it themselves in-house - they can subcontract it to someone who knows how to run internal systems like that. The third item does put a supplier risk into the chain, but you already had that with Facebook; and at least you can swap out the supplier if you don't like them.