heh, this is funny considering it’ll be easier to saturate the IO on cloud vendors and DDoS quite a chunk of modern services than to do any of the categories they bring up https://www.bis.org/publ/arpdf/ar2018e5.htm Chapter V of the Annual Ecomomic Report 2018. Cryptocurrencies promise to replace trusted institutions with distributed ledger technology. Yet, looking beyond the hype, it is hard to identify a specific economic problem which they currently solve. Transactions are slow and costly, prone to congestion, and cannot scale with demand. The decentralised consensus behind the technology is also fragile and ...
I think it would be a lot cheaper to DOS bitcoin by spamming transactions than to DOS the likes of VISA or Mastercard
for probably ~0.5 BTC you could monopolize all the space in a block. at 6 blocks/hour & $6400/btc, that is ~$19,000/hour to make bitcoin transactions cost prohibitive
if you mine the blocks yourself it costs nothing
but that is a big if
it would be hilarious to run a mining operation which didn't both including any transactions
still "it costs nothing" ignores opportunity cost.
just like claiming mining on self-sustaining solar power is "free electricity". You could be selling that electricity to the grid
right, if you are mining that much you would probably really prefer bitcoin do well
since you have millions of dollars of sha-256 hardware
One of magic things about PoW - the method of legitimate use and “exploitative” use depends on basically the same hardware - so there is always a balanced opportunity cost for either
Vs. knowing how to exploit the VISA network translates into little to no experience into how to run a payment processor company
spamming transactions doesn't require any PoW
Miners make decisions on whether to include the spam
so?
miners will include the transactions with higher fees.
Then either spammers pay high enough fees to subsidize miners, or spam slows down
correct
which, like I said, is not very expensive
System is self correcting in this way
try and DDOS VISA for < $20,000 hour
also: visa runs its own infra for the critical bulk, did not mean what got implied here
most meaningfully big things actually run their own beefy infra for the bulk and use the cloud for distributing peak loads
maybe, maybe not
Over time, I think most will move to cloud
depends, if you can save your on premises staff into the margin of amazon, sure, but at a larger scale the price of electricity dominates again and you save yourself the premises and staff back from the amazon margins on that
and as transitions both ways are with a lot of inertia, and the company is older, makes sense to play a bit of both games
of course, you assume your company can be as efficient in electricity as a cloud provider
I'm incredulous
but yeah, big legacy companies are going to be hybrid for a while at least
they already have big capex investments. no reason to throw those out
fresh high margin services or consumer oriented companies will prolly never leave the cloud just for the HR simplicity
yup
in a lot of countries governmental subsidies flip that dynamic
government subsidies to run your own DC?
no, just for new tech sector electricity
hmmm, not familiar with that. you have a link to an article discussing it I can check out? or just something you've run in to in your work?
did not find the tekes (.fi) program in english, but apparently this is now a thing for bitcoin miners as well: https://www.theatlantic.com/technology/archive/2018/03/bitcoin-mining-arbitrages-cheap-electricity-into-money/555416/ A close-up portrait of bitcoin miners in eastern Washington reveals the real hustle at the heart of cryptocurrencies.
here’s the current eu summary, but that’s a bit of a slog to try to make sense of or to find an example: https://ec.europa.eu/energy/sites/ener/files/documents/ECOFYS%202014%20Subsidies%20and%20costs%20of%20EU%20energy_11_Nov.pdf
thnx
but i’ll hold the cloud is an uncanny valley of premises and staff vs. electricity
IIRC both github and gitlab went off the cloud at some point for the dynamic, though the gitlab wtiteup on it revealed their issue was more their ceph architecture than the lack of bang per buck in the cloud