I ran munch again set to use up all the memory, after that all the memory was free.

I guess it was just disk caching.

I looked at the linux ate my ram site and did a couple of tests today to see if that was the issue.

First I let bitcoind run until it filled up the entire memory, after that it started using virtual memory.  After it had filled up a part of the virtual memory I stopped it.  It released part of the memory, but still most of it was still used.  Then I ran munch, the c program on linuatemyram, to see if it ran in physical memory, but it ran in virtual memory instead of reusing the ram that bitcoind had tied up.

Unfortunately, free and top do not help much because they report the usage of the entire shared server(90 GB), while I only have 3 GB.  I have been using the display on the control panel webpage to see how much is used.

I installed bitcoind 0.9.3 on a debian server and tried to download the blockchain, but it seems to be leaking a lot of memory.

while running bitcoind, ps says the process is using only about 700 Mb but the memory usage display on the sever keeps going up and up over time.  After I stop bitcoind most of the memory is still reported as being used even though there are no other processes using nearly that much memory.

Also the rate that it is downloading blocks has slowed to something like 1 per second.

Is it possible that this is caused by an error in my bitcoin.conf?

Another issue related to transaction speed is the blockchain bloat.  Once blockchain pruning has been implemented then it will be more practical to increase the transactions per block.

One thing that I see as a major issue is the centrilization of mining.  The proof of work mining system as implemented creates an advantage for economies of scale.  To be a profitable bitcoin miner it is helpfull(necessary?) to have a huge asic farm.  It would be preferable if mining was done as part of the decentrilized network, but it seems to be headed towards a situation where only a few big players will be able to mine.  Then there will be nothing stopping these few players from colluding to controll the network.

So far, all of the alternate methods proposed have the flaw that the miner is required to own coins before they can be a miner.  Also, none of them solve the problem of centralization that is begining to happen in the bitcoin network.

Another method I have been thinking about is a full node lotto.  The only requirement to be a miner would be to run a full node.  One full node is randomly selected to recieve the payout for each block.  There are two ways I can imagine this being implemented.

First would be to include a list of all miners on the blockchain.  Based on the hash of the previous block or a random number, one is selected to be the miner of the next coin   (max_possible_hash / hash =  total_miners / winning_miner).  The problem with this method is that it would require miners to be registered, possibly decreasing anonymity.

Another way would be to have a crawler(or several) traveling from full node to full node.  Upon arriving at a node it builds a block and does some sort of verification to see if it can be a valid block(eg. pick a rondom number, if number is less than threshold, block is valid).  If so the node publishes the block and recieves the block reward.  If not the crawler moves to a different node.  The problem with this is that there needs to be some method of ensuring that the crawler script is being executed honestly.

In additon to this it would be good to have a way to ensure that a person or organization could only run a single node, or this would just turn into a proof of diskspace scheme that would be vulnerable to centralization.  One way I have considered is to only allow one miner per IP adress, but this has the obvious flaw of destroying anonymity.  Not sure if there is a solution for this.

So far I have found four methods proposed for generating new blocks:

1. Proof of work --  The most common method that has worked well so far, but some problems seem to be apearing.  One is the high cost of maintaining the network compared to other methods.  The other is the centralization of hashing power(If GHash and BTCGuild cooperated, they could perform a 51% attack right now)

2. Proof of stake -- https://en.bitcoin.it/wiki/Proof_of_Stake --  This method has been developed to eliminate the high cost of mining.  It may be able to make a network with lower transaction fees than a proof of work network.   The main objection to this is that it is possible for the big holders of coins to monopolize the mining, creating a situation where the rich get richer and the poor get poorer.

3. Proof of burn -- https://en.bitcoin.it/wiki/Proof_of_burn --  This is a method where a miner has to send coins to an invalid address in order to purchase a virtual mining rig.  The argument here is that the burnt coins are not a real expense from a global perspective (whereas electricity is a real expense for the global economy) because the price of coins will increase to compensate for the deflation.  I have my doubts about destroying coins being healthy for the currency, but could imagine a scenario where the "burnt" coins are instead used as part of the reward for future mining.  I don't believe that anything like this has been implemented in a real coin.

4. Proof of bet -- https://bitcointalk.org/index.php?topic=131230.0  --  This is a method where the miner must pay to try to mine a block.  The chance that the block is accepted is proportional to the amount that the miner paid.  If the block is accepted the miner recieves the block reward.  If the block is not accepted the miner loses the coins paid.  This seems to be in the idea stage of development, with much work requred to make it a real thing.

If anyone knows of any other methods or has an idea please share it.

It seems to me that it would be a good idea for the bitcoin network to have another minting process implemented as a backup in case something goes wrong with the current method.  I don't know if there would be a way to implement something slowly and painlessly, or if a hard fork would be the only way to do it.  It is clear that other methods of transaction verification have not been developed sufficiently to be usable at this point, but it is also clear that bitcoin mining has been becoming more centralized and therefore less trustworthy.  It may be nessesary at some point in the future to change the mining process.

rGfhSHdFg1QEXzLvWGgwuW4Zgz8BYDVhkz

ok

hi

Looks like there is a lot of misinformation about this.

So if you don't reuse your addresses then you will have two layers of encryption, but if you do you will only have one?  But this only applies if you reuse the address after you have spent money from it.

If i am understanding right, this is also related to why when you spend coins, the entire balance in that account is spent, but whatever change you need is sent back to you in a new address.

If I spend money twice or three times, will the client or the network automatically change the account that is being spent from every time?

I have signed up as a seller on localbitcoins.  It looks like there are people from around the world who are willing to trade bicoin for cash.  The site itself does not buy or sell bitcoins, it is mostly just a means of helping people get in touch.  I have never actually made any exchanges that were arranged there but it looks like a lot of people are.

I have heard that you are not supposed to reuse your bitcoin addresses.  I have been looking around to but have been unable to find any good information about this.  I am interested in finding both the technical reasons for this and the practical implications.

At this point my understanding of the issue is that it is ok to reuse a recieving address(public key hash), but after you have spent any of the bitcoins then you need to generate new addresses. 

Does this mean that if someone is making regular payments to me over a long period of time, I will have to send them a new address for each payment?

I am using the Bitcoin-Qt client, if I am going to send or recieve money multiple times do I need to take special steps to ensure security, or is this all handled by the client.  I have noticed that the recieving addresses used before I spent money are still available for use, is this ok?

I am really confused about this issue, hope someone can make tis clear.

Thanks