Blockstore
Our portfolio company Onename is announcing something today that I think is pretty powerful. It’s a key value store built on top of the Bitcoin blockchain called Blockstore.
Onename initially built Blockstore for their own business but is releasing it to the cryptocurrency community today to solicit feedback.
Here’s how Onename designed Blockstore:
- Based on the Bitcoin blockchain:
- Unlimited data storage: Storing large amounts of data in the blockchain can lead to blockchain bloat, so we decided to use a DHT for data storage while storing only hashes of the data in the blockchain, yielding virtually unlimited storage.
- Give money back to miners: Miners provide critical infrastructure for the Bitcoin ecosystem and we believe that any services built on top of Bitcoin should contribute to Bitcoin mining incentives.
- End-to-end design principle: The end-to-end design principle of keeping the core of the network simple proved to be very successful for the Internet, and we believe this is a good decision for blockchain applications as well. Thus, we use the blockchain for a few basic operations and keeping most of the intelligence client-side.
You can check out the Blockstore codebase and documentation on Github for more information.
Comments (Archived):
Storing only the hash does make great sense for the quantity problem, but does DHT (ignorant here) allow for rigorous security?
…and persistence. As in, will that data record always be there? Is it distributed like the blockchain ledger?
Yes, those are my follow up questions… do you maintain the benefits of the blockchain with DHT?
Yes. Blockchain is used as an index to all data entries and also (more importantly) used to verify that the data you received form the DHT was actually correct.
1. Yes, data is set by default to persist for years.2. Anyone can build an index of the data in the DHT – think google crawlers or the way back machine.3. The DHT has a high replication factor.
Right…but what I think your answer implies but doesn’t specifically state is that initially the ‘value’ part of the data is really stored on one node.It’s not replicated or indexed by other nodes/systems unless someone takes the time/effort to do that (ie. it’s not automatically replicated out to all the other nodes like the blockchain itself is).Not a show stopper…but something that does have to be known and accounted for when building systems with/around this.A key in the blockchain that points to data that is no longer available is basically wasted blockchain space (and unwanted transaction cost)…but I think also inevitable with this approach (but prob. a worthy tradeoff overall).
Just edited to indicate that the DHT has a replication factor. The point is you can get the data from anywhere – any one of the many nodes holding it in the DHT or any API that is mirroring the data in the DHT. The DHT is just a decentralized starting point for the data to be stored.
Cool. +1What are the transaction costs around all of this? Is the cost on the writes and the lookups or just on the writes?Assuming people get comfortable with the tech and the concepts…I think that’s probably going to be the main sticking point and biggest thing that needs to be considered before committing to building it into any major application backend…
I am reasonably competent around software technology and I have no idea what any of this means, nor can I find a cogent definition, for laypersons, of blockchain. Yet it is a constant subject here. I bring this up because I believe that the whole bitcoin community is mired in this kind of arcane language and until it stops writing like this there will never be widespread acceptance. It is pure gobbledygook.
http://www.languagelinker.c…
Hmm. Key-value store technology has been around for a while, perhaps since 2009.http://www.slideshare.net/m… Alternative question should be – How is it better if done on the blockchain & with DHTs?
Perhaps since 2009? Try circa 1952.
In use & wildly deployed with modern tools since when?
Considering that in 1952 there were 3 UNIVAC installations and IBM had just released the 701 for which it had 18 orders, the notion of widely deployed is kinda nonsensical.A quick review of the literature suggests the first implementations were circa 1953.
I was thinking of the modern implementations such as Redis, FoundationDB, Hypertable, etc.. But broadly speaking, you could say they belong to the No-SQL category.
Ok, so you’re only out by 10 years 🙂
Yes, I got what you were trying to say and hence my reply 🙂
Yes, I got what you were trying to say and that’s why I brought up the recent implementations 🙂
Yeah key-value stores have been around since forever. But recent implementations like memcache and Reddis popularized them in web developers. So they’re better understood and much more heavily used now.
Yes, in memory databases have been using key-value store for a long time. I used to be very familiar with Berkeley DB, which was released 20 years ago. http://en.wikipedia.org/wik…The wiki page says bitcoin uses Berkeley DB.
Proper communication is needed for some technologies to breakthrough but at the core it’s still technology and evolving tech at that. Renaming tech is a step that can be handled halfway through its adolescence. Maybe the time is now?
Think of a blockchain as an append-only list of updates to a ledger. Every 10 minutes, one person in the world is able to append to the list and add more “updates” to the list, which are then reflected in the ledger.Ryan’s balance: 100Martin’s balance: 100Fred’s balance: 100Block 1:- subtract 5 from Ryan, add 5 to Martin (signed by Ryan for authorization of the transfer)Ryan’s balance: 95Martin’s balance: 105Fred’s balance: 100Block 2:- subtract 9 from Martin, add 6 to Fred and 3 to Ryan (signed by Martin)- subtract 2 from Ryan, add 2 to Fred (signed by Ryan)Ryan’s balance: 96Martin’s balance: 96Fred’s balance: 108And it turns out you can also store raw data in those blocks if you’d like.P.S. Hope I did my math right.
Wont the blockchain just become really massive? Doesn’t seem like much of a long term thing.
Yes. There are extensive continuous discussions on how to scale up the network (number of transactions) and at the same time keep the size of the blockchain down.In our case, with data embedding, we’re only storing the hash of the file (20 bytes).
This is important for using the Bitcoin/Blockchain technology for building decentralized applications. Let’s assume we’re talking to someone who knows nothing about Bitcoin: all they need to know is that with their private-key they can own their data. “Magic” happens at the Bitcoin/Blockchain layer and people can own their own data. This data can be anything (usernames, profile information, software licenses etc). What we’ve built and released is the “data layer” on top of Bitcoin/Blockchain. It exposes an interface (key-value pairs) that most programmers understand, so now they can easily start building new apps in this brave new world without having to worry about the underlying layer.
My “normie” understanding:Blockchain is the virtual “ledger” created by all of the people working to “mine” new bitcoins. It results from the complexity of the mathematic functions required to mine bitcoins. It is both the output of the actions at any given point and the official public input for everyone going forward. Information can be stored in it. It allows for all sorts of decentralized and public verification which many see as the foundation for a new technical framework similar to how various technical protocols have enable the creation of websites, email, apps, etc etc etc. No one controls it or can easily influence it, which is very different than the current hierarchical internet.Obviously all feel to correct me if/when/where I’m wrong.
A block chain is a database that is distributed across multiple computers around the world. This specific kind of database records the history of all transactions that make changes to the database and synchronizes these changes with all of the copies being stored around the world. Some of these changes are secured using a “proof of work” function like Adam Back’s “hashcash,” others use different methods of byzantine-fault-tolerant consensus to overcome the decades old “Byzantine General’s Problem” (i.e. how to come to consensus on some truth on an untrusted network). If you can figure out how to make that any simpler I am all ears.
Thanks guest, that actually makes sense to me. Maybe you should edit the wikipedia entry for blockchain, which is totally useless!
Your comment is gobbledygook and basically one big blaming statement. Get off your ass and Google the term instead of expecting someone to hand the definition to you in a way you can understand it, you lazy ass.
Troll
I’m just being honest here. I’m really bored with the “layperson’s” argument they can’t understand the technology. Here you are, whining about definitions and ‘arcane language’. What’s arcane is your lack of interest in actually learning about a technology and how it works. I have the thing installed and running on a virtual instance on OpenStack running at my house right this second. I’m actually spending time learning about how it works and how I can use it to build value for everyone moving forward. All you are doing is whining and wanting others to figure out your problems. BTW, if I’m to be labeled as a ‘troll’ for speaking the truth, then so be it. Fantastic comeback, BTW.
What are the typical applications they hope to attract, and will it be a process-based pricing model or other eg based on mining support?
This is geared towards anything that requires globally unique name registration and/or validated storage of data.It can be used for registering domain names, registering user identities (what we do), registering the identities of works of art (what Monegraph or Mine do), etc.
Thanks. For educational purposes, could you elaborate on why this approach is better than non-blockchain ones?
Because before the blockchain, you couldn’t do this in a decentralized way. The domain name system, for example is hierarchical and very centralized in terms of control.Decentralized =- hotbed for permission-less innovation- more secure- users have unique write access to their records- data can’t be tampered with- anyone can independently lookup the information- can’t be censored
Why is this better than Redis, FoundationDB and other existing key-value store platforms? – It would be great to compare vs. existing choices.
For comparison with Redis, this is better for applications which are read-heavy and need high security. By read-heavy I mean that you write the data once and read it many many times. Writes are going to be slower on Blockstore (because the transactions needs to get confirmed on the blockchain), but reads can be made very fast. It has the nice property that the server can’t lie to you about the data (much higher level of security).
It would be seriously impressive if it’s faster at reads than Redis (many Redis systems are just key/value stores in memory and so they offer insanely fast response times).That being said, because it’s distributed it should be *way* more fault tolerant than most any Redis implementation.So there are def. some advantages vs. Redis, but I doubt pure speed would be one…
Reads could be identical because you can keep a local copy in memory and keep that copy consistent with the blockchain/DHT. We already do this for openname-resolvers (memcache server that keeps data consistent with the blockchain for reads) — https://github.com/openname…
Cool – but that’s memcache making the reads identical to Redis speed…not blockstore.To me the comparison is about where the data is stored and if/how it’s replicated (I think you can make a legit case for blockstore there)…but not in the retrieval or processing.
It doesn’t need to be memcached, just in-memory i.e., baked into the implementation of blockstore daemon itself. A read from memory is a read from memory and is going to be comparable 🙂
OK – the question though was why is it *better* than Redis or other current options…speed isn’t an answer to that (at best it sounds like it’s equivalent in terms of speed).Replication, availability, security…those all sound more like the possible *real* answer to me.To me the real question though, and what this whole conversation should *really* be about, is how does the cost to this compare to these other options…tech. debt costs as well as actual transaction costs? What advantages do I really get for incurring those costs?
Security. You’re not trusting any single server/service for the data that you’re getting. The end-users own the data and only they can modify it. Simply not possible in memcache/Reddis.
That I can see/agree with 😉
any risk of false-positives in current design? we can trust both layers independently, but upon lookups, how do we trust the result given?
You don’t have to trust the DHT layer. You can independently verify that the data was correct by checking the hash of the data with the hash in the blockchain. The blockchain layer handles all aspects of the security.
saw this note further down the thread…”Yes, all blockstored servers (daemon) have a full DHT node baked in. So when you fire up blockstored you automatically become part of the DHT network.”
Thanks Muneeb….I was asking all these questions on purpose, to encourage you to further explain things so that more people can understand it.
Wow this reminds me of @terrycojones’ Fluid Info – a publicly writable datastore controlled by nobody. Terry was definitely ahead of his time!
Ah interesting, just checked it out!
I miss Terry and Fluidinfo. I was a big supporter at the time.
I still think there is a lot of win in the idea.
What are DHTs? Aren’t those the harmful chemicals in the pesticides they used on apples in the 1960s?
LMFAO…. Where’s the Rachel Carson of BTC
DHT = distributed hash table
Sounds like sharing tokes with gathered friends.
Seems like some kind of male enhancement product: http://en.wikipedia.org/wik…
TIL:”During embryogenesis DHT has an essential role in the formation of the male external genitalia…”
It’s part of ELLOA, didn’t you know?(Extra Long List Of Acronyms)
DHT is actually an initialism ;)https://www.youtube.com/wat…
LOL, I guess people are familiar with bittorrent but not “DHT”. In the early 2000s there was a lot of interest in this technology. I’m happy to see it make a come back as we start moving back towards a more decentralized Internet.
Even the Bitcoin core Devs consider DHT a joke in 2013. Source: http://bitcoinstats.com/irc…
I thought bitbind.io already provided blockchain-based key-value pair storage. But reading the documentation, it seems bitbind.io has not implemented DHT yet, so there’s at least that one significant difference. Would be interested in a more detailed comparison from someone with a better feel for the code than I have.
As far as I understand, bitbind is more about attaching attributes to own-able and transfer-able concepts and things. It’s more like tagging and hashtags (which don’t have any unique human-readable name registration properties) and less like domain names and the names in blockstore, which are globally unique.And yes, also, it doesn’t use a DHT.
Gareth is apparently using bitbind.io to register properties for unique collectible trading cards, and somehow tying card ownership to a bitcoin transaction. I guess the private keys associated with the transaction output prove association with a particular tagged property, which is different from storing unique key/value record? OK, he has pretty good documentation, and so do you, so I should just go read it and try to understand better.
You can think of Blockstore as providing a simple interface to storing data using the blockchain. Let’s say you want to store a trading card for Tom Brady. You can simply do set(“tom_brady”,”data of card”) and later do get(“tom_brady”) to retrieve it. The private-key of the bitcoin address associated with the transactions owns this card. So a simple set() get() interface (i.e., the key-value store) on top of the most secure and widely-used blockchain.
Basically what I’m getting at is, if I wanted to use blockstore to track unique game assets, is it good for that? I think the answer is yes. 🙂
Yes :)It’s great for any non-fungible token.
It took me a bit to understand what a “key value store” is… I always hate to be outed as a normal, but then I guess we’re all on a journey and should celebrate when take a step forward.For any others who don’t know, it’s basically a different type of database structure. I though this explanation was really helpful: http://www.aerospike.com/wh…My interpretation of this is that Onename has created a clever new way to create a very scalable and flexible database structure in an open framework. By embedding aspects of it into the blockchain (the hashes), they are building a foundation for future blockchain enabled apps. The insight here is in part only tying hashes into the blockchain. The hashes are used as a “door key” to unlock the data without having bloat in the blockchain by the data living there. There are also ways to create monetary value for miners. I’m not sure about the 4th point, but it makes me think about some of the distinctions Benedict Evans is making between Apple and Google’s assumptions about mobile design. I suspect this is applicable across both of their architectures and client side could refer to both the “google cloud” or the “smart ios device.”Looking through the other comments, it’s clear that others struggled with it too, which is comforting and interesting. Agreed on the various comments that some of the jargon heavy components of the blockchain/bitcoin world will need to evolve.
I met the founders of Factom (factom.org) last week. They are proposing a similar hashing and storage network, but with a propriety token for the data storage. People could post a hash of any accurate or inaccurate data, but the time-stamp is network-validated. Over time people could “build a case” for accuracy by hashing the steps of a process to prove they went through the process. But, I get nervous about too many nominal-value transactions in the blockchain…. my two cents.
Factom is very interesting!And hm, personally not a fan of proprietary tokens, since I believe in the strength and value of Bitcoin as a currency and supporting the security of the network by paying miners. That said, everyone has their own approach.I’d point out that most of the operations in blockstore are nominal-value transactions, but the name registrations give big bucks to miners 🙂
Yeah, I’m definitely not “selling” Factom. But I thought it was interesting and applicable to the topic.
I posted this above but since you are answering Qs…Having a system that is dependent of “miners” ? Are miners the equivalent of a 1099 employee with total autonomy. Is there a difference between a uber driver and a miner? Is the role of miners the equivalent of crowd determined surge pricing?
My god does bitcoin have a marketing problem. I mean… I’ve been an entrepreneur building technology for many years, and I even understand the blockchain, mining, all of that. And I think it’s pretty genius.BUT….I can honestly say I have NO IDEA WTF this is talking about. Maybe they don’t need me to.But I’d bet the winner is going to be whoever can “Twilio” this stuff. Put it in terms that business can understand and get excited about while providing a powerful platform for developers to utilize.I’d challenge every blockchain based company to BAN the words “blockchain” “block” “bitcoin” “mining” “miners” etc from any of their public-facing copy.
Andy, you do have a point. This particular blog post we wrote was aimed at the technical crowd, as what we’re releasing here is a developer tool, but our next post will be targeted at the mainstream consumer.
I understand that, and I’m not picking on you. This is something I’ve noticed time and time again with bitcoin-related apps/posts.I don’t consider myself a technical n00b…. I can go through Twilio and understand 99% of it, test the code, etc. I AM your audience, whether you want to admit it or not :)I’m rooting for you guys and trying to help…. good luck!
Good call! We’ll work on it for next time.
You guys are doing awesome stuff…. picks and shovels for the next generation of internet (maybe). Don’t let my critique of the messaging take away from my admiration for the team, product and vision.
Also, what are your thoughts on this explanation of how bitcoin works? http://blog.shea.io/what-is…
Sorry, but in a word, no. It only explains the part of bitcoin that is easy to understand.
Ah ok, what are the parts that are most confusing to you?
Nothing confusing about it, it is quite clear, it just doesn’t tell the whole story.
I REALLY think there is a massive disconnect that is only widening. I’m exactly the kind of person (entrepreneur looking to utilize tech platforms for my own niche/idea) that this kind of thing HAS to appeal to. Write to me every time.Amazon (AWS) fails as well in this respect. Easy for anyone who can and wants to take the time to understand it, hard for the “prosumer” to wrap head around. Perhaps because “prosumer” has more important fish to fry. Part of the issue is coming up with new names for old things. But that’s a strategy to then keep people who learn their way “newly hatched” from easily switching to legacy alternatives.
I agree with you that documentation could and should be better ( so that anyone who wanders into the room isn’t immediately scared off), but I am not sure you are part of the intended audience. I think Onename is building tools that can be used to build things like Twillio for normals ( and smart people like you ). If I’m wrong about that, then I’m doubling down on everything you said. 🙂
You are being far too kind…
But then there’s mo secret handshake.
You can say that again. There was a Google search surge for “WTF is Key-value store” after this post.
Nice! Do you think that Google integrating twitter triffic in search will be a good thing?
i see it sometimes reflected, but not sure if they are doing an experiment, or if it’s for keeps. it’s a subtle signal maybe. social engagement is one signal, but i don’t think it should replace search relevancy.
Having a system that is dependent of “miners” ? Are miners the equivalent of a 1099 employee with total autonomy. Is there a difference between a uber driver and a miner?
Miners can choose to pack blocks however they’d like. If there’s a transaction that they just don’t like, they can choose not to include it. But… if there are mining fees that are enough to entice them, they will usually choose to include them. Decentralized mining ensures that no single entity has a monopoly on the ability to create blocks and include transactions in the blockchain.
Let’s say your product, hypothetically, scales as quickly as facebook.. Where does the supply of miners come from? What does one do to “pack blocks”? Does one do this from the terminal?
That’s the beauty of building on top of the most widely used Blockchain i.e., Bitcoin. People already mine Bitcoin for various reasons. Number of transactions per day on Bitcoin are already reaching 100K per day. We can afford to grow by 100K new users per day 🙂 .. there of course will be scalability challenges are insanely high transaction rates but we already are thinking about potential solutions to those. Our goal with this project was to take some classic scalability principles of the Internet and try to apply them to the blockchain tech
There are 200 million photos uploaded to Facebook per DAY. If each photo required a mining . How does this scale?
Oh you interact with the blockchain only for registering new users or updating sensitive information in your profile e.g., your bitcoin payment address or PGP key. General updates to data don’t touch the blockchain. In fact, our entire argument is to use the blockchain as a “control plane” and use existing Internet infrastructure as “data plane”.
I think this has always been a big issue in tech and has been around forever. The tendency to assume nobody is going to get lost at “DHT” [1] which of course if they are not developers they will. Back in the day this actually served a purpose as it kept people at bay and allowed you to charge them for things they didn’t understand and it made them think you were smarter than they were. Anybody who speaks a difference language can sound smart. Or get them to fixate on a number that you could manipulate to make them think you had a better product (hard drive space, processor speed etc.)But I’d bet the winner is going to be whoever can “Twilio” this stuff.Or maybe Debeers.The “problem” is actually easy to solve. Just get some normals to review all of your marketing and point out things that confused them. Then fix that.But even with “DHT” it needs to be broken down further using analogies that everyday folks understand, that is, if you think it’s important for them to even know what is going on behind the hood, it’s not going to be a definition of DHT.A distributed hash table (DHT) is a class of a decentralized distributed system that provides a lookup service similar to a hash table; (key, value) pairs are stored in a DHT, and any participating node can efficiently retrieve the value associated with a given key. Responsibility for maintaining the mapping from keys to values is distributed among the nodes, in such a way that a change in the set of participants causes a minimal amount of disruption. This allows a DHT to scale to extremely large numbers of nodes and to handle continual node arrivals, departures, and failures.The idea then would be to translate the above into something that makes sense to your Aunt or Uncle. One of the reasons I use so many analogies. Analogies allow everyday people to understand your point and are good shortcuts.[1] http://en.wikipedia.org/wik…
Bitcoin doesn’t have a problem. It’s a protocol not a company. People will use it, or not, depending on their needs and preferences. Your criticism is akin to saying that TCP/IP and OSI layering is too complicated. TCP/IP has a “marketing problem”. Well, no it doesn’t. You have no choice but to use it. OK, I’m exaggerating — for the time being, bitcoin has not established the network effects of the Internet. But it will (or it won’t) regardless of whether we feel threatened by having to learn new acronyms.
Your criticism is akin to saying that TCP/IP and OSI layering is too complicated. TCP/IP has a “marketing problem”.Not really because you didn’t have people trying to sell you on the idea of “TCP/IP and OSI layering” they were selling you on something you could do which used that and they really didn’t pay much attention to those acronyms in their marketing.Look, nobody cares really about how a high speed ski lift works. What they want to know is how long it will take them to get to the top of the mountain again so they can ski down. So the message has to be “3 minutes to the top of the mountain, not 10” that’s a concept that people understand and a benefit that they want. Once you have them at that if you want you can then embellish with some under the hood stuff to try and impress them.Car companies do this. They will talk about 0 to 60 times front and center and then embellish with some technical stuff about the engines that nobody understands (but not before, key point.)http://www.coloradoskihisto…
That’s fair. I wonder, however, if the Internet had existed before the Internet, then there might have been a crowd of obnoxious teenagers trying to hype their half-baked TCP/IP schemes of varying quality and legality.In any event, I draw a distinction between blockstore, and your average trivial hack cobbled together by some script kiddie. The implications of blockstore are very interesting to people who’ve been looking for a decentralized key/value store with arbitrary memory limits. I personally think this could be very useful to manage distributed economies, for example, in multiplayer game design. But I can barely program my Google apps, let alone hack something in C++. So don’t listen to me!
Elevation drop is all I care about when I ski.Whether it be a gondola, a pull rope or whatever.Do you ski?
I used to but I don’t anymore. I’m to worried about getting injured which would prevent me from exercising which is much more important to me than the fun that I would have. Plus my wife never skied and has no interest in doing so. I used to love to ski.That said I’m the type of person who hates to have to wait in line for a long time just to go down. [1] With the type of skiing that I did I cared more about having a long trail that would make the fun last longer.When I first got into skiing I asked some older experienced skiers where I should ski. I had never done it before. They told me “go where the conditions are the best!”. Of course if you are new to skiing that doesn’t matter you are just going to do bunny slopes. So what you want is essentially a nice place that from a newbie standpoint is friendly to beginners and looks and quack like a New England duck.I am always amazed at that “out of box” experience when I know nothing about something and I ask for advice (back in the day, when you couldn’t read things that is). People were rarely ever able to tailor what they said to what was important to a newly hatched.Newly hatched: I was in a wine store the other day and picked the Riesling that had a label that said “staff picks”. And let’s just say the staff isn’t the staff you get in NYC either. But better than a random label purchase so I went with it. (Reason I think you should do “Arnold’s picks”.)[1] That’s one of the things that I liked about having a boat. I could use it whenever I wanted for as long as I wanted with no “governor” to the fun I could have. I’m not an amusement park ride kind of guy. Was just at Disney with the family and didn’t go on a single ride.
I don’t think you should feel so threatened by me saying “tell the benefits not the features.”Again… twilio does this extremely well, and is built upon some pretty complex stuff. Why can’t blockstore? Why can’t bitcoin application providers in general? To me, that’s a challenge worth undertaking that could yield HUGE dividends for these guys.Why does every criticism of bitcoin have to be met with an arrogant “you’re too lazy to learn new acronyms” kind of response?
I am more of a features guy. When I read about a new app, I get really frustrated with these single-page contentless marketing brochures, and immediately start digging for the technical documentation. (Disclaimer: you’d go broke trying to market an app if you mistakenly assume I’m your representative audience.)Also, I don’t know what Twilio is or does. Does that discredit me when it comes to marketing advice on the web? Yes, probably.
I think there is a way to do both really well.
Like Steve Jobs bridging tech / design and selling that?there’s a bit of a dog whistle effect with bitcoin stuff. if you’re wired a certain way, you look at bitcoin (and perhaps blockstore) and think “man, this is awesome shee-it!” and most people are not wired that way. whenever i start talking to my wife about bitcoin (and for the record, my wife is a web technology professional, and knows more about tech than I do) she listens, but pretty soon i notice her eyes are closed and there’s a soft “zzzzz” soundand then there are the bitcoin evangelists that non-bitcoin people seem to think are good. and i look at how they are “marketing it” and they’re obviously scammers, or at best extremely, irresponsibly, naive about money on the internet.so who are the folks doing both of these things — substantive work, and evangelism — really well? Fred Wilson, I suppose, for one. There are not too many that are bridging the gap.
Andreas Antonopolous is a good spokesperson for bitcoin.https://www.youtube.com/wat…https://www.youtube.com/wat…
the thing about this andreas fellow, is that he wants to be everywhere in bitccoin, and that has led him to associate himself with some pretty obviously bad ideas as well as good ones. for example, he aggressively hyped the neobee scam, and is famous for tweeting a link to malicious software to his followers (innocently — he didn’t write or read the software before tweeting).the crowd that I refer to as the “Romanian school” (after their ringleader, who’s a Romanian fellow) has AA on their sh*tlisthttp://cascadianhacker.com/…
Strongly agreed with Andy here.
First example I saw today explaining in normie terms the value of this was on the HN comment thread: Blockstore, by creating transferable digital property in the form of a key/value store, creates an economic incentive to use Bitcoin. Right now, Bitcoin needs all the help it can get. Consumer adoption will not happen if Bitcoin is just trying to be a public ledger of integer values that attempts to compete with MasterCard and Western Union.https://news.ycombinator.co…Not sure I could fully explain/defend this claim, but I can start to put the pieces together if this is accurate.
The claim is pretty much accurate. The innovation here is building a familiar interface (key-value store) for programmers so that they can easily start doing things like transferable digital property, decentralized messaging, decentralized storage etc.
I still think it’s so cool to have founders reply in comments. Call me a rube, but it’s a thrill – thanks.So basically this is an attempt to give developers the building blocks so that other things can be created on top of it? Make sense.Which means Onename is hoping to create an open enough framework for app/feature development that Onename’s “one name” system is embedded inside all of that natively from an identity standpoint?Then you go to the topic of “how do you beat Facebook” or whoever? By building something essential to the next great tech platform transformation completely outside of the walls of their existing scope, even if common identity is something they care deeply about. If (when?) blockchain becomes TCP/IP and if (when?) services get built on top of it, this takes Onename much closer to being deeply and completely integrated inside of it. Of course if truly existentially frightening, one could expect this to be the sort of thing Zuckerberg spends a lot of money to acquire and get rid of the threat. Either way, probably a very fair/good bet on USV’s part. Kudos.As a normie, it would be interesting to hear why this is different than other activities trying to do something similar like Namecoin. I’m guessing the bet is the blockchain concept with bitcoin has enough momentum and traction that other chains or coins will not pick up critical mass, or if the current bitcoin evolves, existing systems will be so tightly embedded in the next version that it doesn’t matter.Either way, cool work, thanks for the comment, and good luck.- Jake (https://onename.com/jakebaker)
Thanks!Yes, a simple and familiar interface for developers to start building apps. Most programmers know how to use a key-value store like memcache, or Reddis.Well the identity protocol itself is open-source and called Openname (http://openname.org), so it’ll be hard for the Zuckerbergs of the world to just buy something and shut it down :-)Yes, precisely. Bitcoin has a lot of developer tools around it. And it has the most secure and most reliable blockchain. If you can figure out how to build on the “main” blockchain then it’s a design point worth exploring
Hear you on the TCP/IP analogy, but remember that it took TCP/IP a long, long time to trickle through to developers. And the world was less connected back then. The bitcoin eco-system was able to develop much rapidly (raised funding quickly, a lot of people started apps/companies on top) so it can be argued that Bitcoin will need to bridge this gap faster as well.
Well, I don’t remember, b/c I was not aware of the Internet until about 1993. Nevertheless, here we are!
Definitely hear you on this one. I haven’t seen something as elegant yet complicated as Bitcoin/Blockchain in the last 10+ years. It sits on the interaction of 3 different research areas. It’ll take a while for these technical concepts to get absorbed and simplified.
Does the infrastructure NEED to be accessible to consumers? Even in the current financial system, there are many concepts that are completely foreign to consumers.I think it matters more if Bitcoin is unable to market itself to the marginal tech-savvy user who is not an early adopter, but would be key to transitioning whatever products are built on the blockchain to the mass market.
I’m not necessarily talking about consumers. I’m talking about people like me…. who are constantly looking for new technology that can help us create awesome stuff in our own niche.
I don’t know if Fred did it intentionally or if the original article was edited after this post… but little things like spelling out DHT (Distributed Hash Table) and then maybe simplifying it (a table that maintains a map of data identifiers to raw data) goes a long way.
lol well said
Oooh, nice idea. Does this work with dedicated sidechains?
Yes. It’s meant to work with any blockchain that allows data embedding.
Let me give you a for instance-see if this is what it means. Suppose I am a concert ticket sales company. I create my own namecoin’s for tickets, and attach a tiny slice of that coin to a ticket to prove it’s the real ticket (create a manifest of sorts). No matter how many hands that ticket goes through, that one slice of namecoin is always attached. There is no need for a clearing house, because the sidechain acts as the clearinghouse. I set up a private sidechain to the blockchain that works with my namecoin. Sidechain is searchable, and transparent and anyone can view transactions-but the data for that sidechain is stored cheaply in the cloud. Is this a reasonable assumption and practical example of what’s possible?
Ryan, you mentioned the possibility of porting to other blockchains like Ethereum. What would be the advantage of that, and scope of that porting?
The idea behind Blockstore is that you separate the key-value interface as a logical/virtual layer on top of a blockchain. We need very basic functionality from the underlying blockchain (ability to embed some data in transactions, and stability/reliability of the underlying network). So in theory, any blockchain that provides these simple features can be used for implementing a key-value store like Blockstore.The main idea is that you really should be building on top of the most secure, and most widely-used underlying network. Kind of like how everyone built their apps on top of the Internet, instead of inventing a new underlying network.
Building applications using decentralized datastores seems like a step in the right direction, but if you’re “keeping most of the intelligence client-side” the application itself isn’t really decentralized, is it?
It is. The decentralization part depends on not having any single point of control/trust which is clearly there.
clearly thereYou probably knew this, but to be clear, I’m speaking hypothetically – not talking about any existing app. I see blockstore as a great way to build applications that cannot lock in users – or their data, which is something I’m very interested in.Clients running blockstored serve as the nodes for the DHT?
Ah yes, we should be more clear on how it’s decentralized because of lack of any central point of control/trust (just like Bitcoin). We’re excited about developers using it to build apps!Yes, all blockstored servers (daemon) have a full DHT node baked in. So when you fire up blockstored you automatically become part of the DHT network.
all blockstored servers (daemon) have a full DHT node baked inso that data is truly “free” ( decentralized) not just the hashes stored on blocks – users are not locked in to my app
This is a great move by OneName because key-value structure will make the transactions ledger much more searchable and quickly accessible.The key value being a unique identifier that unlocks every information cell for a particular transaction.So…,coincidentally, yesterday I was at AWS where we went through key-value stores and the quirks of Hadoop clusters and Redshift wrt Big Data. This is fine for me; there are advantages to learning about databases since high school Comp. Sci. (even if it’s dry and boring and unsexy — unlike UI/UX, haha) and having worked through my own db architecture for Machine Intelligence and analytics reasons.In short, OneName’s providing database infrastructure like the AWS does for millions of clients. This makes sense as a monetization strategy (and I knew it would go the “AWS for Blockchain-Bitcoin” route as soon as I saw the key pair generation pem/ppk OneName does).Yes, blockchain and Bitcoin do need technical people who can translate and market the nerd tech language into Joe-Jane Main Street lingo!
Are you guys on IRC? What channel? Let’s chat
Slight of hand and twist of fate….Another important concept that I realized today. ABS Fred… ABS!
We’re building a blockchain-based collateral/secured transactions registry in Africa and were planning to build a decentralized data store… this might have just saved us a bunch of time.
this is genius
Listening to BlockCypher in SF and they are also an AWS for Bitcoin-Blockchain.
True, cost of mining is a potential long term issue for Bitcoin.
That’s true but that’s also true of Amazon Web Services. Tech requires energy
Think about all the emissions that it replaces by replacing banking infrastructure, etc.Also, theoretically, Bitcoin mining could be zero emission. Eventually we could switch from hash-calculation-based “proof of work” mining to “proof of stake” mining, which simply means you get to mine by having control over a percentage of the monetary supply. Zero emissions.
They say the avg. number of MHash/s are 2000. When I look at their source it shows several programs that do hundreds or thousands and millions of MHash/s. Am I misreading something?Also, I thought the number of machines needed to calculate a hash was relative to the volume of transactions. Is this inaccurate?
Datacenters (which power the Internet) already consume about 3% of US electricity. That is a *lot more* compared to how much electricity is being consumed by Bitcoin miners. The issue here is that people are OK with “wasting energy” when it comes to reliability of the Internet (datacenters are over-provisioned for spikes in load), but Bitcoin mining *feels* like such a big waste because you’re actively doing computations.
Never mind – I see where I misread the article. 🙂
This is the part of the story that seems to be sept under the rug. Also, what happens when a miner dies?
Your empathy is outstanding.
Sniffle. I’m the type of guy who just spent part of my afternoon last week setting up vmware esxi on a used supermicro (had to modify the bios took a while to clear that hurdle) that I bought off ebay just for fun. [1] You need to bring yourself down a notch Charlie. I can teach you how to move amonst the idiots and what they often say that might frustrate you.[1] We have this installed in a few colo boxes and I want to understand how everything ties together.
As I’ve stated I’m not a programmer but I can get some things done that are helpful [1]I have to tell you that nothing jacks me up more than doing something on the computer that solves a problem. But I wouldn’t want to do it 12 hours per day, every day. That’s what I like about my setup I can do what I want when I want, generally.[1] Like the guy in the Viagra commercial “the age of knowing” https://www.youtube.com/wat…
Other mining nodes take over….the mining is done by companies now that have huge mining farms, not as much anymore by individuals, since ASIC.