Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Sia – Decentralized cloud storage network

blog.sia.tech

Post reply on HN
Print view View original

151–160 of 199 posts

Re: Sia – Decentralized cloud storage network

#151

Post by lmpostor »

modeless wrote: post #72
Ajedi32 wrote: post #35

Earlier quoted context omitted.

Doesn't AWS also have the same problem? Someone could upload encrypted illegal content, then later law enforcement could discover the key. Why hasn't AWS gotten in trouble for that yet?

No large cloud provider allows anonymous usage. They take down content on request and assist law enforcement in catching the non-anonymous people doing illegal things on their platforms.

Same argument for Bitcoin, cash etc. You have a thousand degrees of slippery slope in what constitutes "enable". Does paying your workers in cash for tips enable tax evasion?

Also, I'm not sure what scenario in which you would be able to find the encryption key, yet not have access to the wallet controlling the contracts? Even if you did, I'm not sure why hosts could not assist in cancelling contracts in a similar fashion to cloud providers taking down content on request.

Re: Sia – Decentralized cloud storage network

#153

Post by camjohnson26 »

dymk wrote: post #148
camjohnson26 wrote: post #126

Earlier quoted context omitted.

Child porn is the emotional argument, the only possible case where most people might agree that some content can be illegal. A more representative scenario is illegal copies of music or a political pamphlet in China.

Sorry, how is child porn only an appeal to emotion?

Never said it was "only" an appeal to emotion, just that it's not representative. It's meant to stop discussion before it starts.

Re: Sia – Decentralized cloud storage network

#155

Post by tech_man7 »

Taek wrote: post #82
rjeli wrote: post #75

Earlier quoted context omitted.

SGX has always seemed like the Netflix DRM problem to me.. eventually, pixels have to show up on a screen somewhere, right? Am I mistaken in this line of thinking? Homomorphic encryption solves this, but it’s throwing an awful lot of performance out the window.

I think the idea behind SGX based cloud computing would be that you send encrypted data and code to the enclave, which then decrypts the data, does the computation, encrypts the result, and then returns the encrypted result. The enclave gets to see everything that's going on, but in theory nothing can view or corrupt what's going on inside of the enclave. I'm personally pretty skeptical that SGX is a secure platform,…

Yes, this is the idea. Enclave technology should improve over time, whether it's SGX or something else. And Golem is working with Intel, Texas A&M and others in creating an environment called Graphene (https://grapheneproject.io/) where applications can easily be run on SGX.

Re: Sia – Decentralized cloud storage network

#156

Post by Taek »

b_tterc_p wrote: post #150
Taek wrote: post #92

Earlier quoted context omitted.

Ajedi32 is not correct. Data is verified probabilistically on the Sia network. The blockchain has access to the Merkle root of the data that the host is supposed to be storing. The blockchain will request that the host provide a 64 byte segment of the data (chosen randomly) along with a Merkle proof that the data is part of the Merkle root. If the host can provide the data and the proof, the host is rewarded as thoug…

How does punishment work? What stops one bad actor from agreeing to collect infinite data from a variety of sources and tanking both the trust and profitability of data hosts? Also what about bandwidth constraints on the host end

When a host agrees to accept data, they put up out-of-pocket money. This makes it expensive for a bad actor to accept an infinite amount of data, as each piece of data requires more collateral to be put forward by the host.

Before a renter creates a contract with a host, the renter will perform some measurements on the host and determine if the host is suitable. A renter in China will chose different hosts than a renter in the US, because the latencies and throughputs of each host will be different.

Re: Sia – Decentralized cloud storage network

#157

Post by dymk »

camjohnson26 wrote: post #153
dymk wrote: post #148

Earlier quoted context omitted.

Sorry, how is child porn only an appeal to emotion?

Never said it was "only" an appeal to emotion, just that it's not representative. It's meant to stop discussion before it starts.

A question that you can’t answer isn’t the same as one meant to shut down discussion

Re: Sia – Decentralized cloud storage network

#158

Post by Scaevolus »

theonlyklas wrote: post #3

Although the cost of cloud backups are egregious compared to the cost of raw storage, I'll pay a premium for the peace of mind that comes from letting someone else be held liable for storing my data and credentials to access it. I can easily upload/download files to google, iCloud, or Dropbox from almost any device knowing only my email and password, which I find preferable to having to remember an arbitrary 29 word…

Sia's seeds are ridiculous-- the 29 words provide 300 bits of entropy. 100 bits would be a sufficient security margin against brute forcing, assuming a modern memory-hard KDF like Argon2.

With a 100 bit password, assuming every flop of the 1.8 exaflops of the Top500 supercomputers tested a new password, it would still take 25,000 years to crack. Key stretching should add at least 30 bits of security by taking a billion operations--

    $ perf stat argon2 asdfasdf -id -m 16 -t 16
Here's what 100 bits of security margin looks like with a more sophisticated scheme (abbrase): "Hope raised between unpleasant bellows. Devil rode sullenly, refugees waiting." => (first three letters) hopraibetunpbeldevrodsulrefwai.

Re: Sia – Decentralized cloud storage network

#159

Post by khc »

Taek wrote: post #114
carlosdp wrote: post #105

Earlier quoted context omitted.

I agree in that they are probably comparing the wrong aspects of S3. You're not going to be able to compete on the speed anytime soon probably, and maybe even the capacity if it really takes off. But, anyone here that works on an Amazon-based internet company knows that S3 is the most unreliable (in terms of availability, not storage) universally used product they have, next to EC2. A good chunk of incidents can be a…

> You're not going to be able to compete on the speed anytime soon probably, and maybe even the capacity if it really takes off Sia today running in the default configuration gets as much as 300mpbs upload and download, which I believe is comparable to S3. Latencies are several seconds, which is not comparable to S3 however this is a matter of software optimization, not a fundamental limitation of the network archite…

300mbps is no where close to s3 throughput. Try 300MBps per node and much higher when you access the same bucket in multiple nodes
Post reply on HN