[LIST] Altcoins that solve the scaling problem (or attempt to solve it)

[d5000]

Bitcoin's extreme blockchain congestion, coupled with high transaction fees, of the last week has shown us that it a solution to the scaling problem is needed.

There are, however, some altcoins that already attempt to solve it and have, in part, some interesting solutions to offer. Some of them are still in the testing stage, others are already available.

The list is still a draft and not complete! Everybody that thinks (s)he knows an altcoin that also does solve that problem can post his/her "candidates" here. I will review it and if I think it adds something interesting, I'll add it. I'll only add coins that have at least a testnet running and a estimated release date. If you spot an error, I'm open for feedback.

Already running:

Cryptonite (XCN)

What they solve: IBD, Storage
How they solve it: Mini-Blockchain
Bitcointalk thread

Cryptonite is already running since 2014. It combines Bitcoin's code with an account tree, and periodically (approximately once per week) blocks are "pruned" from the chain while the account tree is updated. Only a "proof chain" with the block headers is conserved. The result: a very fast-syncing (~1 hour) and very small blockchain which almost doesn't grow.

Downsides: They had a critical flaw in the code until September 2017, but it has been fixed.

IOTA

What they solve: IBD, Storage, Latency
Hoy they solve it: Tangle (DAG) snapshots
Bitcointalk thread

IOTA is one of the most experimental and different cryptocurrency projects. Apart from their well-known usage of a DAG instead of a blockchain, which drastically reduces latency between nodes, they offer also a scaling solution: Once a transaction has been stabilized, it can be "forgotten" by the network and instead the state of the affected accounts is saved on a snapshot. So the general way they deal with the problem is pretty similar to XCN's.

Downsides: IOTA still uses a centralized node, the "Coordinator", that is part of the validation process. However, the IOTA developers plan to shut it down once the network has matured.

ARDOR

What they solve: IBD, Storage
How they solve it: Child chain snapshots
Bitcointalk thread

ARDOR is the second generation of the NXT project. It carries on a novelty: Almost all transactions are not recorded on the main chain, but on "child chains". These are a bit like sidechains, but they temporarily (for 1440 blocks) become validated and stored by the main chain - after this timeframe they are pruned. Child chains can be stored by archival nodes if it's needed.

Downsides: Not fully open source (JPL)

Lisk (LSK) (originally Crypti)

What they solve: IBD, Storage
How they solve it: Sidechains
Bitcointalk thread

Lisk is the only coin I have on my radar that provides a mechanism for pegged sidechains. Pegged sidechains are also planned for Bitcoin, but only centralized ones (e.g. Blockstream Elements) have been delivered until today. Lisk's solution, however, is also not totally decentralized; the sidechains are issued by an entity that must care for the security of the chain via Delegated Proof of Stake (DPOS). But it's a step in the right direction, because the peg itself seems to be managed by an internal mechanism and not by the issuer.

Downsides: The mechanism is not totally decentralized.

Some other interesting future projects (without release date)

- Lightning Network (testnet already working) / Raiden network
- Ethereum's Plasma child-chain architecture (planned) and "sharding" (planned)
- Drivechains
- IOT Chain

"What they solve" section

- IBD: Initial Blockchain Download. That means that new full nodes must download less data to participate. "Light" (SPV) nodes do not count as almost all cryptocurrencies offer them.
- Storage: Full nodes have to store less data on their hard disk.
- Latency: There is less latency between nodes that transfer transactions or blocks.

[CoinHoarder]

Lisk (LSK) (originally Crypti)

What they solve: IBD, Storage
How they solve it: Sidechains
Bitcointalk thread

Lisk is the only coin I have on my radar that provides a mechanism for pegged sidechains. Pegged sidechains are also planned for Bitcoin, but only centralized ones (e.g. Blockstream Elements) have been delivered until today. Lisk's solution, however, is also not totally decentralized; the sidechains are issued by an entity that must care for the security of the chain via Delegated Proof of Stake (DPOS). But it's a step in the right direction, because the peg itself seems to be managed by an internal mechanism and not by the issuer.

Downsides: The mechanism is not totally decentralized

I'm not sure this is entirely accurate. My understanding is unlike Bitcoin sidechains which are pegged- Lisk sidechains are not pegged. Otherwise, what would be the financial incentive to build a token/dApp on top of Lisk? Especially when they could build on another smart contract platform and receive said benefits.

[Btc4Proxies]

Great post! You should look into ITC as they solve the security issues and inefficient centralized architecture of IoT.

How they solve it: DAG+PBFT

This is from their doc:
"ITC is designed to solve the severe safety problems of present IoT. ITC applies a combination of asymmetric encryption of cryptography, semi-homomorphic-encryption ciphertext computing technology and distributed architecture without data center. Therefore, ITC can not only protect the users' devices from being attacked by hackers, ensuring its controlling right safety; but also defend users and their devices' data safety sovereignty and privacy security. For example, the data of intelligent device, such as the camera, can only be checked by users themselves.

ITC is a safe light operating system for IOT based on blockchain. ITC solution integrated blockchain technology, asymmetric encryption technology of cryptography, computing technology of semi homomorphic encrypted ciphertext, and distributed architecture without data center, with the aim to resolve the server's security problems in IOT, to meet the demand of highly concurrent usage scenarios, and to achieve the connectivity of all things.

ITC is the abbreviation of IOT onchain Token, which is used to support the decentralized operation system of IOT. As the measure of value transfer in the ecosystem, any value transfer about the use and ownership as well as the content on intelligent devices must be settled by ITC."

They say they are the IOTA of china. Currently trading on like 3 exchanges and have 3x'd since their ICO a few months ago. https://coinmarketcap.com/currencies/iot-chain/

[d5000]

I'm not sure this is entirely accurate. My understanding is unlike Bitcoin sidechains which are pegged- Lisk sidechains are not pegged. Otherwise, what would be the financial incentive to build a token/dApp on top of Lisk? Especially when they could build on another smart contract platform and receive said benefits.

I had read in the Lisk documentation there was a mechanism to "lock" Lisk and transfer their value to the sidechain. I remember it as a "one way peg", not a "two way peg" (the "Holy Grail" attempted e.g. by the Drivechain mechanism). Unfortunately, I've lost the link to that document, so I couldn't re-check it.

But maybe it's true that even then Lisk should not be in this list - because we have already one-way-pegged sidechains in Bitcoin (Blockstream/Rootstock). I put it "on hold" for now, until I find the document.

Great post! You should look into ITC as they solve the security issues and inefficient centralized architecture of IoT.
How they solve it: DAG+PBFT
They say they are the IOTA of china. Currently trading on like 3 exchanges and have 3x'd since their ICO a few months ago. https://coinmarketcap.com/currencies/iot-chain/

What is the project status? On Coinmarketcap they're still listed as a "token". If there's at least a public testnet running and a rough release date in the coming months, I will list it here.

[CryptosapienZA]

Interesting list. I have never heard of cryptonite..

To add on your list is Raiblocks. It uses what they call a block lattice. "Users send funds by creating two blocks: one send block on their personal blockchain and one receive block on the recipient’s blockchain. Users receive funds by “pocketing” any outstanding receive blocks into their personal blockchain. Users do not have to be online to receive funds (this is a common question concerning RaiBlocks’ novel “pocketing” system). Whenever the user decides to access their funds, the wallet itself will automatically “pocket” any outstanding funds. Pocketing funds essentially means signing the receive block with your private key, so that it can be added to your personal blockchain".
https://hackernoon.com/iota-vs-raiblocks-413679bb4c3e

[d5000]

To add on your list is Raiblocks. It uses what they call a block lattice.

Thanks. I know RaiBlocks a little bit, and have briefly read about their blockchain model. Can you clarify how they solve the scaling problem?

For example: Is every new full node forced to download the entire DAG ("lattice") before he can make a transaction? Or is there some form of "pruning" like in IOTA? Latency between nodes should be better in RaiBlocks than in blockchain-based solutions, but latency improvement alone, in my opinion, does still not qualify as a full "scaling solution".

I also have to investigate if Byteball fits into this list.

[christianb35]

To add on your list is Raiblocks. It uses what they call a block lattice.

Thanks. I know RaiBlocks a little bit, and have briefly read about their blockchain model. Can you clarify how they solve the scaling problem?

For example: Is every new full node forced to download the entire DAG ("lattice") before he can make a transaction? Or is there some form of "pruning" like in IOTA? Latency between nodes should be better in RaiBlocks than in blockchain-based solutions, but latency improvement alone, in my opinion, does still not qualify as a full "scaling solution".

I also have to investigate if Byteball fits into this list.

RaiBlocks completely fix the scaling issues, so far it as no known limits. It is a new blockchain technology, basically each node/wallet has its own blockchain so it doesn't need to wait for anyone before sending a new transaction.

If you run a full node, yes, currently you need to download the full database (about 1.8 GB), They can of course prune the database or make it optional to download it completely and that only apply to when you want to run a full node. But the storage is not reallly a matter? At least it is far from the biggest scaling problem of BTC currently... the BTC Blockchain is what, 100GB? I think we can consider that Hard Drives will grow at the same rate as the blockchain.
For the web wallet or the coming light wallet it wont be needed to download everything of course.

The transaction are validated by a "Delegated Proof of State" mechanism, so no time wasted by a proof of work.

How is it not a full scaling solution? Currently it does faster transaction than IOTA, 0 fee, 0 inflation, it is fully decentralized and there's no limit in the number of transactions: https://i.imgur.com/oRpO8lc.gif

[d5000]

If you run a full node, yes, currently you need to download the full database (about 1.8 GB), They can of course prune the database or make it optional to download it completely and that only apply to when you want to run a full node. But the storage is not reallly a matter?

The problem isn't so much storage. At least in traditional blockchains, when a node "syncs" or "downloads" the chain, then it validates all the block hashes. This consumes CPU and RAM, and the higher the transaction load, the higher the minimum requirements for "full node" hardware. (For a throughput of ~25 tx/s (8 GB blocks), for example, a 16-GB-RAM machine is needed in Bitcoin-style coins, for 50 tx/s it's 32GB, and so on). Additionally, every "initial sync" consumes bandwidth; and in a terabyte-chain that can be already a problem for a node running on a home PC.

So the question must be changed, because I don't know exactly how RaiBlocks works:

Do the nodes have to validate the whole dataset when syncing? (In a DAG or "lattice" instead of validating blocks they would validate transactions - if they spend past valid transactions - and if these do not conflict with the votes of the DPOS validators.)

If yes, then I see not really advantages of RaiBlocks with respect to Bitcoin and other coins, unless they have some system to reduce the "load per transaction".

Iota, as far as I know solves the problem via pruning, because transactions that are fully spent long ago are "forgotten" and nobody must download them.

(I am not biased against RaiBlocks, only I don't know the system very well. They deserve, at least, some respect to do things a bit different than 99% of the rest of the coins. )

[christianb35]

If you run a full node, yes, currently you need to download the full database (about 1.8 GB), They can of course prune the database or make it optional to download it completely and that only apply to when you want to run a full node. But the storage is not reallly a matter?

The problem isn't so much storage. At least in traditional blockchains, when a node "syncs" or "downloads" the chain, then it validates all the block hashes. This consumes CPU and RAM, and the higher the transaction load, the higher the minimum requirements for "full node" hardware. (For a throughput of ~25 tx/s (8 GB blocks), for example, a 16-GB-RAM machine is needed in Bitcoin-style coins, for 50 tx/s it's 32GB, and so on). Additionally, every "initial sync" consumes bandwidth; and in a terabyte-chain that can be already a problem for a node running on a home PC.

So the question must be changed, because I don't know exactly how RaiBlocks works:

Do the nodes have to validate the whole dataset when syncing? (In a DAG or "lattice" instead of validating blocks they would validate transactions - if they spend past valid transactions - and if these do not conflict with the votes of the DPOS validators.)

If yes, then I see not really advantages of RaiBlocks with respect to Bitcoin and other coins, unless they have some system to reduce the "load per transaction".

Iota, as far as I know solves the problem via pruning, because transactions that are fully spent long ago are "forgotten" and nobody must download them.

(I am not biased against RaiBlocks, only I don't know the system very well. They deserve, at least, some respect to do things a bit different than 99% of the rest of the coins. )

I don't really understand the issue. I am too looking for the ultimate scaling solution and so far RaiBlocks seem to be the one to me but the storage/download is not really something that bothered me: People that need to run a full node, well technically it is only those doing the "Delegated Proof of Stake" also called representatives that need to run one and one of the reason why DPoS is a good solution is that it let only those with a good network and hardware in general do the PoS and for them downloading 100 GB or 1TB or 10TB ... once in a life time .. is nothing.
Pruning is always an option, I don't think that it is something that is hard to do, it is just something something that should be done lightly because we maybe loose the transactions history if all nodes do that.
To me, when it comes to storage, the most important thing to watch is "What data is stored" ... if the Blochain tries to store too much data then it may have scaling issues later, at least more than others. But RaiBlocks is clear on that: Do one thing and do it well and its transactions are very minimalist... it is why its the current database is relatively small for the number of transaction it contains.

RaiBlocks is actually very different than all other currencies, its Blockchain is at the same time simple and can scale at the level that matters the most to me: transactions.

[d5000]

People that need to run a full node, well technically it is only those doing the "Delegated Proof of Stake" also called representatives that need to run one and one of the reason why DPoS is a good solution is that it let only those with a good network and hardware in general do the PoS and for them downloading 100 GB or 1TB or 10TB ... once in a life time .. is nothing.

Yes, delegates shouldn't have problems - but it would not be as decentralized as, for example, Bitcoin wants to be. DPOS is a tradeoff between centralization and high hardware requirements for the "masses".

Pruning is always an option, I don't think that it is something that is hard to do, it is just something something that should be done lightly because we maybe loose the transactions history if all nodes do that.

The cryptocurrencies that heavily prune (Ardor, Cryptonite, Iota) assume that they're reaching a level of consistency after certain time that makes keeping the history unnecessary. This is also a tradeoff - you can't "prove", for example, in Ardor or Cryptonite, that you have sent some coins one year ago to another person (you can only prove that the account state is valid). But for most applications that should be OK, it's like with physical cash.

it is why its the current database is relatively small for the number of transaction it contains.

I must investigate further on that. Do you know how large are average 1-to-1 transactions in RaiBlocks? (In Bitcoin, they're between ~190 and ~220 bytes).

I don't rule out that RaiBlocks has a RAM/CPU usage advantage over blockchain-based coins because of the "transaction-is-a-block" characteristic. But I must investigate that further. The problem is that no altcoin (perhaps Ethereum or Ripple) has been really "stress tested" with the levels of transaction throughput that Bitcoin achieves. In the altcoins I list here, however, you can infere logically from their features that they scale better than BTC.

[Ebenezar64]

A very nice article that quickly highlights the reason you are interested in these coins.

Personally I agree on Ardor and IOTA and I have my eye on both of them. IOTA have had some issues but that is normal with such an ambitious project and who doesn't like zero fees?!

I would say I prefer ARK to LISK, I just think the team seems a bit better and I like the Smartbridges that ARK is trying to implement. I am eagerly awaiting the next Core update which will have some cost reductions in Delegate voting and transfers.

If you are looking specifically for Decentralized solutions or Scalable Blockchains the first things that come to mind are:
- Substratum for a decentralized internet
- RaiBlocks or HPB for extreme scalability.

Would love your view on these on how they solve the decentralization and scalability solution.

Will scalability even be a problem if coins start to implement the lightning network?

[d5000]

I would say I prefer ARK to LISK, I just think the team seems a bit better and I like the Smartbridges that ARK is trying to implement. I am eagerly awaiting the next Core update which will have some cost reductions in Delegate voting and transfers.

ARK is definitively a candidate for this list. I will investigate it and evaluate if it fits the requirements. The last time I checked it, its "scalability-relevant" features were still in development. I just looked at their homepage and it seems that has not changed still. The crucial question for me is if their SmartBridge is able to manage advanced techniques like "pegs" so currencies can be "expanded" via a sidechain. That seems to be a similar problem like with Lisk.

If you are looking specifically for Decentralized solutions or Scalable Blockchains the first things that come to mind are:
- Substratum for a decentralized internet
- RaiBlocks or HPB for extreme scalability.

Substratum, afaik, is still in the whitepaper/planning stage. I would like to see it in production (even as a testnet) and then evaluate it. As of RaiBlocks (is it now being rebranded to Nano?), I still have doubts like I explained above. I'll look at HPB, however according their homepage their testnet will start in March.

Will scalability even be a problem if coins start to implement the lightning network?

I have asked that myself, but the LN has some downsides and on-blockchain scaling techniques (even sidechains) have a lot of advantages over it, because on-blockchain TX preserve the "trustless transaction" paradigm that was revolutionary at BTC, while LN is only partly trustless. If Bitcoin achieves decentralized pegged sidechains, however, even BTC would land on this list.

[Ebenezar64]

I have asked that myself, but the LN has some downsides and on-blockchain scaling techniques (even sidechains) have a lot of advantages over it, because on-blockchain TX preserve the "trustless transaction" paradigm that was revolutionary at BTC, while LN is only partly trustless. If Bitcoin achieves decentralized pegged sidechains, however, even BTC would land on this list.

Thanks for taking the time to answer my questions and for following up on your own threads (not something everyone does).  I agree that keeping the "trustless transactions" is an important part to maybe keep ahold of. When you are thining of side-chain solutions are there any existing example out there that are already working?

And do you agree that the sidechains can be considered "altcoin killers"?
What kind of conditions do you think there will be between "chain-transfers"?

The Raiden Network (Ethereum) and Atomic swaps also seem like techniques that have some merit. What are your views on these?

I agree that both ARK, RaiBlocks and Substratum needs to show some proof of concept before we can get overly positive but I still like their ideas and potentials.

[qwizzie]

You can add Dash to that list, it has a detailed roadmap towards onchain upscaling :  https://github.com/dashpay/dash-roadmap