Centralization of the mining process may be a good thing.

[jdbtracker]

As mining operations get bigger and bigger and competition gets smaller and smaller, people may choose to start using their Asics in a different way.

The centralized pools and miners may wish to collectively agree to maximize profits by analyzing the mining process and gaming the system to control the difficulty with variable ASIC Farm operation.

The miners or pools could implement a ramp up for the network using the Fees to control when the miners should pour the brunt of their power, there by increasing the amount of transactions per second within the network.

We can increase the 7 transactions/second to a far larger number, basically we can scale the network to a maximum of 10k transaction per second if we begin managing the mining network, obviously this has to be done by consensus, but in the long run a variably controlled mining network will net higher return. This will eventually naturally happen when the network begins to half the reward amount, probably some time in 2020; At that time the Bitcoin mining lottery will be only 6 coins/block, with a maximum Bitcoin price of 1 million dollars/coin that is 864 coins per day, the networks growth will be very difficult at 1million/coin. And this may be reached sooner because of the 1mb/block transaction limit, we will reach this within the next year as business and people begin using it more and more...

If you concur with what I am saying, then please begin researching Mahout machine learning algorithms to control your Pool or Farm; It is part of the Hadoop infrastructure, On a pool it would be easiest to apply and let the algorithms do the data mining for you. With this Mining managers can control the amount of transaction they wish to include in blocks to maintain consistency of the network, control of what % of  your mining operation is active at any given time to scale the transaction limit as necessary, variable predefined  minimum fees based on urgency instead of people guessing how much to use, variable VPS connections to scale transaction volume to specific areas of the network(basically building tunnels to take care of regional variation in transaction volumes to maximize revenue and minimize block difficulty), ASIC VPS Business prioritization to minimize their fees for bulk transactions, off chain transaction processing delay for when blocks are hard to fill(Intentional transaction delays, a sort of hourly or daily clearing of low priority bulk transactions to off peak hours), tier control to move transactions to Farms or ASICs with the greatest probability of finding a block(a probability calculator to see which network miner is closest to finding a block). etc, etc, if you have any more suggestions we should consider it.

These charts are the variables I'm considering, the difficulty would change considerably over a two week period with maximum Hashing all the time; They show that with transactions alone there is a constant ebb and flow to them that changes on a week to week, day to day and hour to hour movement.

Transaction volume

30 day transaction volume, linear, raw data
http://blockchain.info/charts/n-transactions?showDataPoints=true&timespan=30days&daysAverageString=1&scale=0&address=

1 year transaction volume. linear scale
http://blockchain.info/charts/n-transactions?timespan=1year&showDataPoints=false&daysAverageString=1&show_header=true&scale=0&address=

1 year transaction volume, logarithmic scale with a seven day average
http://blockchain.info/charts/n-transactions?showDataPoints=false&timespan=&show_header=true&daysAverageString=7&scale=1&address=

all time transaction History, linear raw, (notice the exponential growth pattern?, these scaling are getting shorter by half everytime... we are getting new users and magnifying it by 10 and doubling that growth rate.
https://blockchain.info/charts/n-transactions?timespan=all&showDataPoints=false&daysAverageString=1&show_header=true&scale=0&address=

All time transaction History with a logarithmic scale applied so you can see the growth pattern more easily.
https://blockchain.info/charts/n-transactions?showDataPoints=false&show_header=true&daysAverageString=1&timespan=all&scale=1&address=

Transaction per block

1 year, logarithmic scale, raw values
https://blockchain.info/charts/n-transactions-per-block?showDataPoints=false&show_header=true&daysAverageString=1&timespan=&scale=1&address=

30 day, linear scale, raw value
https://blockchain.info/charts/n-transactions-per-block?showDataPoints=false&show_header=true&daysAverageString=1&timespan=30days&scale=0&address=

Transaction confirmation time average

https://blockchain.info/charts/avg-confirmation-time?timespan=1year&showDataPoints=false&daysAverageString=1&show_header=true&scale=0&address=R
 
Transaction fee average 1 year, linear, raw  (ideally this should be flat, this would denote a very well managed network.)

https://blockchain.info/charts/transaction-fees?timespan=1year&showDataPoints=false&daysAverageString=1&show_header=true&scale=0&address=

[yatsey87]

Maybe there might need to be some sort of dedicated mining foundation set up eventually, especially if Bitcoin goes mainstream and has to cope with huge demmand. I suppose it depends on how much fees will be in the future.

[jdbtracker]

Maybe there might need to be some sort of dedicated mining foundation set up eventually, especially if Bitcoin goes mainstream and has to cope with huge demmand. I suppose it depends on how much fees will be in the future.

I was hoping people would see the un-spoken possibility of setting up more p2pool style pools with personal network controls for each member alongside the Mahout machine learning algorithms to achieve dynamic voting control of concensus; Different rule sets, mini networks within the vast distributed pool, this type of centralization may be desirable. At least with this setup we can begin constructing pools that can go beyond the 51% mark and destroy any possibiity of a network Hash attack... It's a reversal of how things are now where people are trying to achieve more mining power but staying under 50% to avoid overpowering the network.

We go all the way to 100% in the distributed pool, as more people chose their own mining parameters it would create diversity and a major pool that all attackers would have to go against. It would also destroy the current centralization by the mining farms, by allowing everyone to share the network reward... decentralized creation once more.