Decentralized networks are defined by the lowest common denominator. If you want to push a large amount of data throughout the network, you need to push it everywhere, even to the slowest devices. Also, the mempool synchronization is not, and cannot be, exact. If a peer asks you to synchronize the mempool with them, neither of you know which transactions the other party already has in their pool and which not. In many cases, you end up resynchronizing the entire mempool. Which means that the block size limit now on Bitcoin is not in fact 1 MB, but 1 MB times the amount of users times the amount of mempool block of mempool volume in between blocks. And every time a block sync happens, you need the whole block and mempool to travel across the entire world, to every node. You also need to remember that the mempool synchronization takes away resources shared with block synchronization, so the larger the block limit is, the more unwanted forking happens because the information about new blocks arrives later to many miners – again, exponentially. If anyone with a single-chain solution tells you they can transfer thousands of transactions per minute and satisfy millions of users, they are lying or using cheap tricks.
So the blocksize limit is a technological problem, or more specifically a Bitcoin-related design flaw. Is there a theoretical solution? Yes, since at least 2014, and it’s called sidechains or childchains, and these have been inspired by nothing else than altcoins. Why? Because altcoins in themselves are the scaling solution for Bitcoin, where Bitcoin is only used as the common chain with some market arbitrary value of the altcoin in question. Once you buy an altcoin, you are no longer putting any pressure on the Bitcoin blockchain, until you decide to sell your altcoin and use BTC as a mediator again. Childchains/sidechains are just like that, but only use one network, one dev team, and one common token. It’s like the whole bitcoin/altcoin market virtualized.
Who is working on this technology at the moment? I am not aware if there is any working solution yet, but quite certainly Ardor, or NXT 2.0 development team is releasing their testnet in Q1 2017. It will be spiced up with additional features already present in NXT like the decentralized exchange. Also, Ethereum made some announcements lately and would like to push their solution public early 2017.
Why should you want to use something like this instead of another altcoin? Three reasons – support, price stability and security. Simply, if you use an unknown altcoin because your main coin like BTC has reached it’s technological limit, you are putting yourself at risk of the altcoin holders, who may decide to dump on you. Also, you don’t know how long the chain will be alive, you don’t know if the devs have capabilities of solving issues, etc.
disclaimer: no proofreading done