Hard to say until we get some actual experimental data during beta for example. It depends on what kind of user you are, how quickly the project is growing, if there is a capacity shortage or abundance, etc. If you’re going to store lots of popular media de-duplication will save you tons.
To borrow the thread a bit, I’m wondering about safecoins the currency, how will them be stored? In a data file just like bitcoins? Also how easy will they be to transfer? Would it be possible to design a smartphone app for sending and receiving safecoins just like there is apps for bitcoins today?
Bitcoins are not stored in a data file, your public and private keys are stored in a data file, which allows you to authorize transaction orders from the corresponding Bitcoin address that are sent to the miners who handle that transaction. Bitcoin addresses simply have a numeric balance attached to them.
MaidSafe uses a login system, using a 4-digit PIN, username, and a password. With those three you retreive your MaidSafe Passport from the network, which is stored there in encrypted form (encrypted with your credentials). In that passport are the private and public encryption keys that belong to your account, to authorize different kinds of actions on the network. So you don’t need to save and secure anything on your local system, everything is behind your login credentials, and it can be accessed on any machine that can run the MaidSafe client software and has an internet connection.
Down the hood, SafeCoins exist not as a balance to an address, but each SafeCoin exists as an actual data structure that has a unique ID. SafeCoins belong to an account. They can be transferred to another account using the MaidSafe software client, by filling in a simple form with how much SafeCoins you want to send to a particular person, identified by an actual, readable username. SafeCoin transactions only take mere seconds to complete, and maybe even less than a second.
It’s definitely possible to design a smartphone app for sending and receiving SafeCoins, and it will simply be a matter of logging in and sending for example 5000 SafeCoins to ‘Seneca’. If I understand correctly the MaidSafe devs will actually make that smartphone app for us and will be available for launch (at least for Android, not sure about iOS. Probably both).
There is equally an option (to be proposed and worked out) for App developers to donate safecoin to their initial users, if they are confident they will earn it back through popular use of their app.
Thanks a lot for your answer!
Looking a bit further, don’t you think safecoin could be the real “bitcoin 2.0”? The currency which the general public starts to use in their daily life? Comparing the two this is what I got:
Advantages of safecoin:
Safecoin doesn’t waste energy like bitcoin do. Energy is used to store data.
Safecoin transfers are instant.
Safecoin is anonymous, you can’t track the coins in the blockchain.
Safecoin doesn’t run the risk of a 51% attack.
Safecoin doesn’t need a 27GB blockchain to be used.
Advantages with bitcoin:
Bitcoin is more well disturbed (but it wasn’t in the beginning)
Bitcoin is already up and running (but they seem to be working hard on the safe network)
I did post this question on the bitcointalk forum too but didn’t get any respond.
I’ve read the previous discussions on the subject:
I find some of the answers there, especially by the guys directly involved with safecoin to be very political correct.
I’m pretty much convinced of it.
Yes, and in addition to the ones you listed, this is my main reason: MaidSafe will have considerable appeal regardless of SafeCoin. Most people have no interest in crypto-currencies, but a lot of people do store pictures, music, videos, documents and own multiple machines. Since accessibility of the same files from multiple machines and also backups are important, a lot of people also see the advantages of using cloud services.
However, cloud services ask money and want to make a profit, there’s only so much free space they’ll give you. In addition, some of those people are wary of their privacy and/or legal ownership of their data.
I’m convinced MaidSafe will be quite economically competitive to other cloud services. It will utilize a lot of free storage space that is currently just sitting there on people’s systems. That’s almost pure profit for any casual farmer, so it will drive the costs down. Casual farmers don’t have a profit incentive, they just try to recover some of the costs of their hardware that they bought anyway. Or they can use the SafeCoins they farm themselves to upload their own data to the network. A huge cost saver compared to traditional cloud services.
In addition, de-duplication will save massive amounts of storage space. A lot of data people want to put in the cloud are public files that are rather popular. The network makes sure there are no unnecessary duplicates of the same files. Putting data on SAFE that already exists is thus de-duplicated and doesn’t cost SafeCoin, since no extra storage space is used. This is a huge cost saver as well.
So, MaidSafe will give you better bang for less of your buck. It offers services that are attractive to the general public. Once people start using it, they’ll come into the possession of SafeCoin (hopefully mainly through farming themselves). Once they have it, they’ll find out how easy it is to use and that it is desirable, and will be far more open to getting more and trading with it.
There you have it, a regular citizen not interested in crypto-currencies, in possession of and using a crypto-currency to buy services he or she desires. That is the formula to mass adoption.
This thread brings up a couple questions for me:
If SafeCoins are stored as actual data which simply changes owners, how is it possible to have it capable of divisibility? When a new safecoin is minted, it will check if an existing safecoin already exists with a particular ID. If a safecoin has already been divided multiple times, what exactly will let the minters know that ID already exists?
What kind of algorithm is used for deduplication? All files must be encrypted using the uploaders’ private keys (to prevent others from deleting it). Would that not create different xor IDs of chunks for different people, despite files being identical?
On an additional note, Gangnam Style was 1 PUT and 2 billion GET requests. I hope the paid PUT and free GET system will be sustainable in actual practice!
Oh wow, that’s right. So what would this mean, then?
Now I kind of wish to hear a reply to this
One of the most important aspects of “online life” is privacy, at least the (piece of mind) knowledge that your data is secure. MaidSafe solves this, assuming the NSA doesn’t have a backdoor key to the native encryption system. (Haha, uh heh uhhh) No but seriously, aside from farming, safecoins and this new concept of a distributed Internet, the safe network would solve the red herring that the government has been feeding us, that is “To get security, we need to sacrifice our own privacy”, which is absolute bolocks.
I know that nodes can cache files, but I don’t know whether or not the cached nodes get paid if they serve the files. I think I’ve seen posts that say they do and posts saying they don’t haha. Maybe someone else can clarify. If they don’t get paid for it, then I guess it’s more sustainable (but why would nodes agree to do that for free?)
And since I’ve already piled up some questions here (paging @dirvine!), here’s an additional question: how will nodes that cache popular content be aware when that content is deleted? Are they in constant contact with the data managers or something?
Nodes will be running 24/7 and the cache definitely is caching what is most popular… since the network just uses what is available; certainly since cache is in memory the network will definitely utilize it to its fullest capacity. Therefore these portions of data won’t be coming out of the Hard Disk so no farming reward is granted on cached data it seems based on posts.
Cache data is held as LRU Cache data rapidly increasing and decreasing based on popularity of data. There is no payment for these but for vaults not to cache is bad for them as they have to deal with more traffic.
Like the PKI if a safecoin is subdivided then its written to with data to show this has happened (basically invalidated, but still in the address space). There are a few ways to subdivide, one of the easiest is to use the next n bits of the address space to do it. Then you have coin abc, you node it is divided so the next bits are abc0 abc1 abc2 etc.
Up-loaders use self-encryption which uses data from the file to encrypt the file. Like convergent encryption but takes it a little further. There is a video and blog thing about it here
Indeed, but also figures show something like 95% of data (stats are on line but hard to find, we did a large report a couple of years back) is put and only read for 30 mins and then never accessed again. There is a balance for sure.
In this video, @dirvine already explains how self-encryption works. It also shows that you’re data-atlas and all the other personal data like private keys etc. are self-encrypted as well. So you’re files are encrypted using self-encryption, really not recognizable because the hashes of the chunk A are used to encrypt chunk B etc. Another layer of protection is done by doing a XOR-trick where you really cannot say anything about that data at all. So here you have a data-atlas of all the chunks and their names, your DHT-routing list and private keys etc. And this file, you’re personal file is going through the process of self-encryption as well. So really nobody has a clue on what that data is. When you log in to the SAFEnet, you’re username and pin will be used to get the file for you. The filename is based on your username and PIN. So your not logging “in” to the network, you just ask for you personal file! And guess what, the moment you get the chunks to your computer, you need your password to decrypt it. And when you go offline, this personal file is going to self-encryption again. But when only one bit is changed (and it will) all the chunks will be completely different from the time before.
Start at 5:30 min. And you’ll get a very good explanation about how the self-encryption, datamap and data-atlas work. I think it’s really brilliant.
Yes, it’s seemed like that for a long time to me.
@seneca has hit on the “invisible” killer feature I think. As well as the technical advantages, Safecoin will be adopted by people, hopefully millions and more, who just want to use the network’s other functions: cheap secure storage, secure anonymous sharing and communications, reward based publishing & creativity platform… plus zillions of apps.
When you have a lot of ordinary folk with money in a really low friction easy to use wallet ecosystem, on top of a completely open platform with a level playing field… you have a recipe for mass adoption. IMO
@dirvine Just a side note. It´d be great if the instruction videos had subtitles. Shona´s accent is beautiful, but it can be a bit difficult to hear what is being said, at least for someone like me who doesn´t have English as a native language, and who is not used to hear this accent.
adding this as a reminder, cool video with @viv
Would be great to read this report, because quite frankly, I have a hard time believing it Any pointers as to where I can read about this?
You need to dig on the net for ages I am afraid, it took us a huge amount of digging. Many companies are now selling dedup software so a lot is hidden behind paywalls and register to get report pages. I am not sure, perhaps @nicklambert still has a copy from way back.
After a bit of digging I found it: http://www.researchgate.net/publication/236011831_A_Storage-Centric_Analysis_of_MapReduce_Workloads_File_Popularity_Temporal_Locality_and_Arrival_Patterns/links/0c96051ffb6dff29fc000000
The research analysed ﬁle access patterns of two multi-petabyte Hadoop clusters at Yahoo! across several dimensions, with a focus on popularity, temporal locality and arrival patterns. The study analysed two 6-month traces, which together contain more than 940 million creates and 12 billion ﬁle open events.
The study found that files are very short lived: 90% of the ﬁle deletions occur on files that are 22.27mins − 1.25 hours old.
Other interesting data:
Workloads are dominated by high ﬁle churn (high rate of creates/deletes) which leads to 80%− 90% of ﬁles being accessed at most 10 times during a 6-month period. There is a small percentage of highly popular ﬁles: less than 3% of the ﬁles account for 34% − 39% of the accesses (opens).
Young ﬁles account for a high percentage of accesses, but a small percentage of bytes stored. For example, 79% − 85% of accesses target ﬁles that are most one day old, yet add up to 1.87% − 2.21% of the bytes stored.