Has anyone ever attempted to build a FPGA hpc cluster using bitcoin mining gear to configure the FPGAs and 40-100G network to load the actual data sets for processing? Custom hardware yes, but could be useful for things like cryptography and the like. Possibly develop some generic RISCV soft processors with four or three pipelines, for a many core CPU on each chip. Could be interesting in practice and could be done multiple different ways, just getting that kind of internal band width stumps me. (200-400+ gb/s) Got several ideas, none are that economical DIY or enterprises for that matter.... Could scale really easily over multiple racks, just the network infrastructure needed is mind boggling. ~Kain Kelly
What is Skywire? Where does it fit in with Skycoin?
Skycoin is a blockchain application platform. We have multiple coins in the platform (Metallicoin, mdl.life, solarbankers.com, etc). We let people launch their own blockchain applications (including coins). There are two parts to Skywire. The first part is the Skywire node. The second part is the hardware. Skywire is one of the first applications we are launching on the Skycoin platform. It is one of our flagship applications that has been in development for several years. Skywire is basically a decentralized ISP on blockchain. It is like Tor, but you are paid to run it. You forward packets for your neighbors and you receive coins You pay coins to other people for forwarding your packets. So it is like Tor but on blockchain and you are paid for running the network. Also, while Tor is slow, Skywire was designed to be faster than the current internet, instead of slower. Skywire is a test application for monetizing excess bandwidth. Eventually the software defined networking technology behind Skywire, will allow us to build physical networks (actual mesh nets) that can begin to replace centralized ISPs. However, the current Skywire prototype is still running over the existing internet, but later we will start building out our own hardware. Skywire is a solution for protecting people’s privacy and is also a solution to net neutrality. If Skycoin can can decentralize the ISPs with blockchain, then we wont have to beg the FCC to protect our rights. Skywire is just a prototype of a larger system. Eventually we will allow people to sell bandwidth, computational resources and storage. On the hardware side, the Skywire Miner is a like a personal cloud, for blockchain applications. It has eight computers in it and you plug it in and you can run your blockchain applications on it. You can even earn coins by renting out capacities to other users on the network.
How would your everyday, average Joe user access the Skywire network? Let's say from their phone…
We designed Skywire and Skycoin to be as usable as possible. We think you should not have to be a software developer to use blockchain applications. Skywire is designed to be “zeroconf”, with zero configuration. You just plug in your node and it works. Its plug and play. Eventually you will be able to buy a Skywire Miner and delegate control of the hardware to a “pool”, who will configure it for you and do all the work, optimize the settings and the pool will just take a small fee for the service and owner of the hardware will receive the rest of the coins their miners are earning. You will just plug in the Skyminer and start earning coins. It will be plug and play. Most users will not know their traffic is being carried over Skywire. Just like they do not know if they are using TCP or UDP. They will just connect their computer to the network with wifi or an ethernet cable and it will work exactly like the internet does now.
Are you completely anonymous on Skywire, or do you need to add a VPN and go through Tor for extra protection?
Skywire is designed, to protect users privacy much better than the existing internet. Each node only knows the previous hop and the next hop for any packet. The contents of the packet are encrypted (like HTTPS), so no one can spy on the data. Since Skywire is designed to be faster than the existing internet, you give up a little privacy for the speed. Tor makes packets harder to trace by reshuffling them and slowing them done. While Skywire is designed for pure speed and performance.
Will Skywire users be able to access traditional internet resources like Google and Facebook over Skywire?
Yes. Most users will not even know they are using Skywire at all. It will be completely invisible to them. Skywire has two modes of operation. One mode looks like the normal internet to the user and the other mode is for special applications designed to run completely inside of the Skywire network. Skywire native apps will have increased privacy, speed and performance, but all existing internet apps will still work on the new network.
How difficult will it be for a traditional e-service to port their products and services to Skywire / Skycoin? Are there plans in place to facilitate those transitions as companies find the exceeding value in joining the free distributed internet?
We are going to make it very easy. Existing companies run their whole internal networks on MPLS and Skywire is almost identical to MPLS, so they wont have to make any changes in most cases.
What is the routing protocol? How are the routes found?
Skywire is source routed. This means that you choose the route your data takes. You can chose routes that offer higher privacy, more bandwidth (for video downloads) or lower latency (for gaming). Skywire puts control of the data back to the user.
I have also understand that the protocols underlying in skywire will be/already are pretty different from the Internet protocols. Taking into account the years of research applied to the current Internet and the several strategies for routing it doesn't seem an easy task to rebuild everything and make it work. Where can be found the information about the routing strategies used in skywire?
The routing strategies are user defined. There is no best routing strategy that is optimal for every user or application. Instead we allow people to choose their routes and policies, based upon the application, time of day, available bandwidth, reliability and other factors. This is actually the way the original internet worked. However, it was scrapped because of the RAM limitations of early computers which only had 4 KB of memory. So the internet was built upon stateless routing protocols because of the limitations of the available computers at the time, not because the networking protocols were the best or highest performance. Today even a cell phone has 4 GB of ram and 1 million times the memory of a computer in the 1980s, so there is no reason to accept these limitations anymore. Our implementation is simpler and faster because we are stripping away the layers of junk that have accumulated. The internet was actually built up piecemeal, without any coherence, coordination or planning. The internet today is a mishmash of different ad-hoc protocols that have been duct taped together over decades, without any real design. Skywire is an re-envisioning of the internet, if it was built today knowing what we know now. This means simplifying the protocols and improving the performance.
How will the routing work if someone from Europe wants to access a video from a node in Australia (for example)? How do the nodes know the next hop if they cant read the origin or destiny of any packet?
If you have a route with N hops, then you contact each of the nodes on the route (through a messaging service) and set the route table on each route. Then when you drop a packet in the route, it gets forwarded automatically. You could have 60 or 120 hops between Australia and Europe and its fine. Each individual node only knows the previous hop and the next hop in the chain. That is all the node needs to know.
Could you estimate a timeline for when Skywire will operate independently from the current ISP infrastructure?
I think Skycoin is a very ambitious project and some parts could take ten or twenty years. Even if we started with a network of a few thousand nodes and we were growing the network over 1% per day, it will still take a decade or two to conquer the Earth. We are going to start with small scale prototypes (neighborhoods), then try cities. I think the first demonstration networks will be working this year.
How will bandwidth be priced in terms of coin hours and who determines this rate?
You could have 40 PHDs each do a thesis on this. The short answer is that an auction model has to be used (similar to Google’s Ad Words auction model) and the auction has to be designed in a way so that the bandwidth prices reach a stable equilibrium. There are parts of Skycoin that are completely open source and public, like the blockchain and consensus algorithm and Skywire. There are secrets like the auction model and pricing, that are designed to protect Skycoin from being forked and to prevent competitors from copying our work. We estimate that if a competitor was to start today, with 2 million dollars a year in R&D, that it would take them a minimum of eight years to develop a working bandwidth pricing model. And from experience in auction models for advertising networks, 80% of the competitors will fail to develop a working model at all. A working, fair, decentralized bandwidth pricing model that was competitive with what we have would take even longer. There are very few people (less than 4) on Earth who have the experience in mathematics, economics, game theory and cryptographic protocols to design the required auction and pricing models. One of Google’s secrets that allows them to dominate the internet advertising industry, is their auction model for ad pricing. That is what allows Google to pay the content producers the most money for their advertising inventory, while charging the advertising buyers the least. Google’s auction models for pricing AdSense inventory are even more secretive and important than Google’s search algorithm. This is one of the most important and secretive parts of Google’s business. Even companies like Facebook, with billion dollar war chests have been unable to replicate to close the algorithm gap in this area. Expertise in these algorithms and their auction and pricing models is one of the reasons that Google has been able to extract advertising premiums over Facebook. Even if a competitor raises a billion dollars and hires all the PHDs in the field and they had ten years to do research, I doubt they would be able to develop anything close to what we have now. The history of bandwidth markets is very interesting and Enron tried to do a trading desk for bandwidth and bandwidth futures and it completely failed. The mathematical stability and predictability of the pricing of bandwidth under adversarial conditions is one of the major problems. For instance, one of our “competitors” suggests that people will be paid coins if someone accesses their content. So why don’t you just put a website and then have 2000 bots go to it, to get free coins! How are they going to stop that. Or if they are pricing bandwidth, if the price is fixed and the price is too low, then people will not build capacity and bandwidth will be insufficient and the network will be slow. Or if the price is variable and adjusts with demands, what will stop someone from buying up the capacity for a link (“Cornering the Market”) to drive the price up 50x on links they control and extort money out of the other people on the network with a fake bandwidth shortage? The pricing algorithm has to be stable under adversarial conditions. It is a very difficult problem, harder than even consensus algorithm research. Even if a competitor had unlimited funding and unlimited time, it is unlikely that they would find a superior solution to what we have and that alone nearly guarantees that we are going to win this market. It gets even more difficult if you need price stability and you admit any type of bandwidth futures, that allow speculation on future prices. This is a kind of problem like Bitcoin consensus algorithm that can only be solved by an act of genius. We have a lot of experience in this area. It is hyper specialized and a very difficult area and is one of the areas that will give Skycoin a strong sustainable advantage.
Will there be a DNS for Skywire to register .sky domains?
Of course. We will definitely add some kind of DNS and name system eventually. Remembering and typing public keys is too difficult. We want to make it as easy as possible. We want people to be able to register aliases (like screen names) so that people can send coins to aliases instead of having to type in addresses every time. This will let people send 5 Skycoin to “@bobcat” instead of sending coins to “23TeSPPJVZ9HvXh6iYiKAaLNQroKg8yCdja”. This will be a revolution in usability.
When operating a Skyminer, will people in my surrounding area see it as a Wifi option on their devices?
You can configure it to expose a wifi access point. It depends on what you are trying to do.
While I plan on running a DIY miner regardless of the payout, will one of the first 6000 DIY miners built to the same spec as the official miner receive a worthwhile payout in Sky coin? What is the requirement for a DIY miner to get whitelisted (and earning Skycoin) on the Skywire testnet?
The reason we have white-listing on the testnet, is to stop too many nodes from joining the network at once. The network can only support so many nodes until we upgrade certain infrastructure (like the messaging/inter-process communication standard). Eventually, all DIY miners will be whitelisted, but there will probably be a queue.
The Sky team is developing antennas by their own instead of buying or using technology already developed, why is such an effort necessary?
You can of course, buy any commercial antenna or wifi system and use it for Skywire. We are developing our own custom antennas, to push performance limitations and experiment with advanced technology, like FPGAs (Field Programmable Arrays) and SDR (Software Defined Radio). Existing wifi has a huge latency (15 milliseconds per hop). We need to make several modification to get that down to 0.5 millisecond per hop. We have several custom PCB boards in development. We have a few secret hardware projects that will be announced when they are ready. For instance, the Skywire Miner was in development for two years before we publicly announced it. Some of our next hardware projects are focused on payments at the point of sale and improving usability, not just the meshnet.
So back in January Steve was asked a question in the skywire group: "Steve, I am not a tech savage, so how can I understand better the safety running a miner if people on the network do DeepWeb stuff? So i will receive and redirect data packets with crazy things and also there is around 128 GB of storage on my miner. How can i have peace of mind of that?" He replied with "If you don’t run an exit node to the open internet it won’t matter you can run relay nodes if you’re worried about it, or proxy specific content." This seems to goes counter to what you mentioned regarding end-to-end encryption with Skywire. Will some people only be relay nodes and some will be exit nodes as well?
I think the question is wrong. You only store content for public keys that you explicitly subscribe to. This means if you do not like particular content or do not want it on your hardware, then you can just blacklist those public keys or don’t subscribe to them. Data never goes on your machine unless you requested it. If you are holding data for a third party such as forwarding packets, it’s always going to be encrypted, so will look like random noise. There will never be anything in the data that causes legal liability. It will look the same as the output of a random number generator.
If using the skyminer, how much bandwidth will be necessary to run it at its best? And what about the router? It's true it has only 100mbits output? Is a 1gigbits connection necessary to reach toprates?
Hold on!!!! Let us get the software and test net running first, lol. We will know once we know what works for the testnet.
What will the price be for future Skynodes (formerly called Skyminers)?
We are working on ways of reducing the cost, such as by buying our own factory, doing custom PCB boards and using different materials. The cheapest Skywire Miner node will be about $30 for a single node miner. We will have a very cheap personal Skywire “hardware VPN” node also. The miners we are shipping now are for powering the network backbone and have 8 computers and are about $800 each. We sold people the miners for 1 BTC each so they can support development, but gave them a Skycoin bonus equal to about 1 BTC worth of Skycoin. Then that money, went to fund the cost for developing the newer hardware.
Fpga Miner Diy. 400 GH/s Bitcoin Miner springs to life with Raspberry Pi Bitfury’s 400GH/s Bitcoin mining rig uses 16 individual hashing boards controlled by a single Raspberry Pi – BItfury ASIC chips use a 55nm process and are sold running at an estimated 1. Fpga miner Fpga miner. FPGA proven. com within 3 days of receiving the miner as per the courrier company's records. AlGharbi Fatema ... Open Source FPGA Bitcoin Miner. From Bitcoin Wiki. Jump to: navigation, search. A miner that makes use of a compatible FPGA Board. The miner works either in a mining pool or solo. This is the first open source FPGA Bitcoin miner. It was released on May 20, 2011. Contents. 1 Software needed; 2 Compiling. 2.1 Altera; 2.2 Changing the clock speed; 3 Programming the FPGA. 3.1 Altera; 3.2 Using ... Build Your Own Portable Bitcoin Mining Rig / Battlestation: Greetings Everyone ! This is XK ; For a really long time since i bought and build my own Bitcoin Miners , I have always wanted to build something more "dedicated " for mining tasks rather than using classical Server cabinets and so on. Something… I plan to make an fpga mining device for bitcoin mining just for the fun of it. I thought understanding bitcoin as whole would be the first step to take. However I realize that understanding the whole verification and transaction process in bitcoin would take a lot of time. Time that I really dont have. So I just wanted to know what would be the most important things to know related to bitcoin ... Open Source FPGA Bitcoin Miner secworks secworks github. SHA256 Algorithm. SHA256 / By ozzy. The OG hashing algorithm, SHA256. Used as the proof of work algorithm for bitcoin and a few other coins (including one of the five POW algorithms used for my favorite coin, digibyte). Pseudo Code From Wikipedia Note 1: All variables are 32 bit unsigned integers and addition is calculated modulo 232 ...
I picked up a few PCI FPGA Cards on eBay for 99p which, apparently, can mine BitCoins at a speed of 21 Ghash/s (once they're correctly configured!) DIY FPGA Installing and Configuring the KU115 for Crypto Mining - Duration: 45:24. Erik Szewczyk 3,590 views. 45:24 . $0 Ethereum Mining Upgrade - How To Hack Your BIOS - Duration: 12:35. UFD ... This is the video showing the final project of EC551 in Spring semester,2015. This project, Bitcoin mining, is done by group "Bitminers". The group members a... DIY FPGA Installing and Configuring the KU115 for Crypto Mining - Duration: 45:24. Erik Szewczyk 3,476 views. 45:24 . Tu iPhone es tan (in)seguro como tu Windows - Duration: 22:51. Chema Alonso ...