Topic: What Happens to the People Wanting to Run a Node in 20+ Years? (Read 243 times)

No, I don't believe that's true. It is hard to predict the direction in which technologies will develop in the future. But, even with the current technology, we are good for the next 10 to 20 years.

That's true but remember that they have to do it only once. You don't download the entire blockchain every time you run your full node!
Think of it as installing a new operating system. It takes a lot of time to format your disks, re-partition them, install the OS, update it, install all your software and maybe update those too. If it is not your first time and your internet speed, your hard disk, etc. are slow then it could easily take a full day to do all these. BUT you only do it once, every future time you start your computer the operating system starts from where you left off.

Initial Block Download is the same. You do it once then each time you restart your node it starts from where you left off and the amount of data your node has to process is much less.

It's largely hobbyists and serious investors that run full nodes now. The cultural change to have individual users run full nodes does not yet exist, but could if bitcoin grows.

Whether a general consumer needs 14 TB+ in that time, I couldn't say. Video games, movies, photos, and home videos may take up a lot of space, but we also improve our compression algorithms. But perhaps it will be the everything controlled by central microsoft/google servers so you won't have to have home storage if you are at any place with consistent internet.

You can get 28 TB for about 800 EUR/USD now, so I expect that price to drop in twenty years. It's not necessarily the case, as production does not always improve. Computers are so ubiquitous that I'd be surprised if disk space didn't drop substantially. Maybe it's like the light bulb and comes to an equilibrium with planned obsolescence.

Most users don't require a txindex. Chainstate is about 7-8GB right now(?), certainly less than 10GB. Block data grows way faster than chainstate, the UTXO set is constantly having entries added and removed but the block data grows linearly.

Note:This figure only considers blocks data. If we include chainstate and txindex, it might have crossed 400GB right now.

This is spot on, and I believe this will actually be an issue in the future. There's often comparisons between storage options today, and storage options 20 years ago. However, that's because old computers were limited even for the general consumer. We are now probably reaching a point where bigger storage is possible, and could potentially become cheaper over time. However, for it to get cheaper you would need mass production of it. If you general consumer doesn't need 14TB+, then the production will only be for those that can actually use it, most likely commercial, and a few hobbyists. So, we could actually see things start to stagnate  for general consumption, simply because we won't need the higher storage values. In fact, we've probably already seen this happening, with the introduction of chromebooks, which are primarily used for browsing the internet, and with commonly used software now taking a online cloud storage point of view, i.e microsoft office there might be no need to install programs yourself in the future, reducing the need for local storage, at least for the general consumer.

In the future, if you were to run a node you might actually need equipment, which gets rarer, and rarer as time progresses, simply because the industry might cater to the general consumer, which could take a more commercially hosted application approach.

With the accelerating pace of technological change, I believe soon it will be possible to run full nodes even on mobile devices. In the meantime, am pretty sure some solutions will be introduced to reduce the blockchain size or the time it takes to download it and synchronize with the network.

By running a pruned node you still need to download all the blocks. It only helps solving the storage issue as you will keep only the most recent blocks. Besides, we will still need others to run full nodes for the network to work.

The size of the Bitcoin blockchain is increasing by linear progression (about 100GB every two years), while technological advancements in the IT industry continue to grow exponentially. For example, a decade ago, most personal computers had disks drives with a capacity of 320GB or 500GB. Nowadays, most new PCs come with drives that hold 2TB or even 4TB. In fact, storage capacity is doubling every few years and data transfer speeds are increasing even more rapidly.


https://www.statista.com/statistics/647523/worldwide-bitcoin-blockchain-size/

So, unless the rate at which new blocks are generated dramatically changes, there is no concern that blockchain data will surpass the limits of technological advancement in the foreseeable future.

I wrote ROM. There is an actual idea there.

If bitcoin continues to increase in value and distribution, we want more people to run nodes. A typical node contains all of the bitcoin transactions that have occurred, up until the recent ones. ROM, as hardware, and only requiring a one-way process ("read-only memory", read, not write), could be created and distributed as a hardware plug in -- perhaps more cheaply than HDD space. This would provide an immutable transaction history, yet one that a node could verify. The core software currently doesn't support a separated database structure, but we might see it in the future if such ROM hardware became available.

It might also be good if core supported splitting/putting the database history into multiple drives anyways..

Current bandwidth is more than sufficient to download the entire blockchain within a few hours at max. Problem is that there exist a ton of bottlenecks which can't be eliminated that easily. I don't think there is a particularly huge issue with an increase in the rate of transactions in the future; assuming transactions can take place off-chain and we can sustain an increase in the efficiency of the various components. I think Moore's law isn't that relevant anymore. We're already reaching 7nm nodes.

There are more concerns within the system other than your bandwidth and your HDD.
Reducing block size is counterintuitive. We need bigger block size to support LN, even if we were to scale to our closest payment system competitor. It's quite evident that on-chain capacity isn't ideal right now, outright blocksize increase cannot be the only solution.

You are assuming that technology will stay where it is ....and that is wrong. We already have very fast bandwidth in some of the most remote countries in the world. People thought 1mb and 5mb was fast a few years ago and now people are having between 10mb and 100mb and that is still on the slow side.  

The storage capacity is also getting bigger and bigger ....so storing the Blockchain for hosting a full node ..will just grow as the Blockchain grows. (SSD's with a 100TB will soon become the norm.... mark my words)  

https://timesofindia.indiatimes.com/gadgets-news/worlds-biggest-ssd-costs-40000/articleshow/76878101.cms

https://www.databarracks.com/future-of-processing-and-storage/storage.html

In 20 years size of hard drives will be multiple times bigger and size of bitcoin blockchain will be trivial, and download speed will probably be much faster so people won't need to wait for months to download full node, or they can just download pruned node.
There is also possibility of some code updates and changes that would make it easier for everyone to download and run full Bitcoin nodes.

Blockchain size did grow a lot in the past few years and I believe it outpaced the development and price reduction of disks.  That being said, the size itself is probably a larger problem now than the bandwidth.  You could clearly see around you that bandwidth is becoming an incrementally smaller issue with the development of new internet technologies and All Time Highs of download and upload speeds coming every few months or years.

The blockchain size should theoretically increase small enough for future technology to become cheaper at the same time.  A $50 Solid State Drive of today will have so much less storage space than a $50 Solid State Drive of 2030 will.  Think floppy disks, CDs and today's USB disks or microSD cards.  With time, future technology becomes more affordable.

-
Regards,
PrivacyG

I've changed my two statements of ROM to HDD now.

ROM is not RAM.

20-22 years ago ISDN was fast. 25 years ago 56k was fast 40 years ago 1200 bps was fast and a few years before that 300 and 110 baud was nice. And a 5 MB HDD was huge and 64K of RAM was lots with a 1 MHz CPU.

20 years from now we will look back and think about how little storage, RAM, and CPU speed we had.

Given the block chain grows relatively linearly and computing power and storage are growing more (Moore) and linearly, and one get get fiber internet today (and for years) the initial sync will be improving over time.

10 years ago 1TB SSD were insanely expensive. Today they are under $100 USD
Bandwidth price is falling in most places. 5G is allowing very high speeds for downloads even when you do not have the ability to connect to a local provider.
RAM & CPU pricing is falling too.

So there it's no big deal.
If you want to run a node on the PC that you just pulled out of the trash that is one thing, but if you are going to spend some money then it will be fine.

Same as now, you can run a node on a slow internet connection on a RPi3 with a slow drive attached to it.

Or you can get a 4th gen core i5 with 16gb ram and a 512GB SSD for under $250 and not worry about it.
You can probably save $40 going for a similar generation AMD based PC.

10 or 20 years from now the systems will be different but theory will be the samw.

-Dave

Yes, people will have to wait and download the whole blockchain, hopefully the bandwith will increase by that time. I would be more concerned about longer perspective, like 100 or 200 years, if our storage and bandwith technology will stagnate, eventually running a node could become a big problem, and the network would be slowly becoming more decentralized. Perhaps a blocksize reduction would be needed at some point in the future to combat this. But this is a problem for future generations.

IPC increases.
Bitcoin at its current form cannot handle transactions on-chain that is sufficient to support mass adoption. We can use second layer but I doubt we won't increase the onchain capacity through some optimization. Currently, the blockchain size growth is roughly ~100GB per year. It's decent but you have to also take into account that most people won't buy an extra HDD for Bitcoin and for the near future, most consumer computers won't include that much storage because there isn't a huge market for it.

It might not seem like a lot but for the average Joe, having something like this hog up their HDD is a far greater problem for them.

The technology is so fast developing and you find optic fibre connections these days and SSD upto 2 TB also and expandable RAM and system processor are becoming really fast that if someone wants to run a full node even after 20 years it would not be a problem for them and there could be more possible options then.For instance at this time the total blockchain is around 352.74 up from 352.61GB yesterday and rising at the rate of approximately 24% annually as last year the total blockchain was around 286.44 GB which means nearly 60GB increase.So we can assume that the size would in TB's at that time but most probably it would be downloaded at faster rate within a week or so if not any isssue.You would also need extra space for further download and system requirements would also increase.But yes we can say that full nodes might be less operational and light weight clients will be more.

Ten years ago, processors were.. 3 GHz. And today, processors are.. 3 GHz.

Wait.

Let me restart.

The necessary storage/transfer will be mostly linear scaling due to the limits on the number of transactions.

So today, a full node takes maybe 300 GB. Ignoring the first few years, we can estimate 300 GB / 5 years. In twenty years, all transactions will be encompassed within < 2TB of storage. Think how fast memory/HDD develops. You have 2 TB of HDD now, easily accessible. And that's the limit within 20 years. That's not really that much.

Yes, it might take a few weeks to download, but bandwidth will increase too.

Also, keep in mind that many people don't use a full bitcoin node to interact with the chain.

That is assuming bootstrapping the entire blockchain will still be the norm 20 years from now. Who knows, perhaps there is a way to truncate or compress the blockchain in a trustless manner in the future.

Even if there isn't, the current blockchain takes about 8 hours to fully synchronize on my computer. We're still seeing quite significant improvements in both IPCs of CPUs and HDD disk size, don't think it'll be a huge problem even without improvements.

Pruned nodes don't solve the problem of having to download and validate the entire blockchain.

So Bitcoin has been around of a little over a decade, and we can see it already takes people a considerable amount of time to download the entire Bitcoin blockchain with the purpose of running a full-node.

What happens to the folks wanting to run a full-node in 20+ years when downloading the entire blockchain takes months and months to complete?

Will they just have to suck it up?

Will pruned nodes be their only option?

Do we already have solutions for this?