Just a quick question, and please direct me if it’s been answered already.
I was just imagining a YouTube on SAFE, and was thinking, “Wait, wouldn’t video streaming be faster if it’s coming from big expensive corporate servers?”
Like, wouldn’t having to pull the video from many different spread out hard drives inherently slow down the video streaming speed?
This would apply to online games and downloads as well.
How does this work? Sorry for asking, it just crossed my mind. But knowing MaidSafe, I’m sure they have a solution
Just reading a paper with some data on ftp transfers using distributed consensus (not quite Safe, but decent for showing generalised paxos and further which does actually do some of the things people tell us is impossible, (I mean Byzantine fault tolerance is surely not possible)) here is a link to that.
TL;DR they seen a 3% decrease in speed (in terms of days to catch up to 100% of the speed at todays increase in bandwidth is probably about a week , I always say design for tomorrow and this is the kind of thing I mean).
It looks nice (visually), that’s for sure. A nicer Vuze client with Vuze running in the background, so to speak.
It relies on the “ordered” fetching, what Vuze has been doing for a few years now (if you pay up).
It doesn’t have to be faster since you can’t watch it “faster” than the speed it needs to playback at (I assume they don’t support unlimited fast forwarding, since for that you always need to wait).
It has to be just as fast as the playback speed is, with some buffering to survive hickups along the way. So it keeps the buffer stuffed and drops in performance are irrelevant as long as the average speed is sufficient before the buffer runs out of juice. And even before that it can switch to a lower resolution to survive a bit longer.
let’s say getting a file from a link into SAFEnet is like 3 times slower than getting it from google. So you want to download a file, 500 MB in size, so you need to get something like 1500 chunks. Now the trick is, if you have like 4 ip-connections into the SAFEnet (will be something like that) to ask them all for different parts. This way you can even be faster than 1 direct link to google. Don’t know exactly how this works on ip-level but it probably will be something like that. Vaults will compete to get you a chunk as fast as possible because this way they make safecoins. Popular files will even be cached. So yes Maidsafe will probably be slower with only one connection. But it can be equal/faster cause of more ip-connections into the network instead of just 1 connection to a server from google or amazon.
I think it should be the same or even faster, if the technology is efficacious enough. Theoretically it being on multiple hard drives is the same as one hard drive, just different communication; it’s all sequential and electron-based. Also basically computation output of many hard drives simultaneously should make for lightning fast potentials given how each one can provide small chunks, especially if each of the vaults is coming from powerful enough computers. (I’m guessing the ranking system could tidy that up. But I’m actually clueless as to how all this works, even if I know a little bit.) I’m guessing it’s like a torrent with many seeders.
@dirvine So let’s say 4 ip-connections. Does this mean with 32 close XOR-nodes that each ip-address serves like 8 closest nodes? Or does each ip-connection into the network provides connection to a different XOR-group?
single node. so we will test connecting to 8 4 and 2 nodes in your group. There are some signature checks to ensure integrity as well that require measuring in real life.
Yes I also had a friend show me popcorn time and it streamed 1080P on a hacked FireTV (Wireless), HTPC at the same time on a ~100Mbps connection with buffering taking ~ 5-30 seconds depending on the popularity. No hick ups…unless it was run through the built in VPN, which seems back to centralized.
If you are using “built in VPN” you’ve been using the fake and less reputable popcorn time project.
The one that took the torch of the original project is popcorntime.io
The one you are using is suspected to have spyware.
it works better than the other and I guess if they want to spy on my friends hacked FireTV then oh well. both are on github and the later in discussion seems to have a better following.
, I always say design for tomorrow and this is the kind of thing I mean).
I can vouch I seen it working and its excellent, a friend of a friend showed me and ran away