Discussion:
[bitcoin-list] Difficulty adjustment mechanism
s7r
2014-09-13 13:15:51 UTC
Permalink
Hi there,

I have an contradictory discussion with an altcoin supporter and want
to bring some solid arguments in a public talk, so require little help
to respond to a question with simplest answer.

The problem is within the difficulty adjustment mechanism, that it
happens at every 2016 blocks. In case the hash power will suddenly
decrease, the 2016 blocks will take a lot of time to solve, therefor
freeze the network in a non-operational way. I know by far this is
just a joke, because this is very unlikely to happen anytime (people
paid for mining equipment and make money) but for the sake of
discussion, let's just assume it 'could' happen.

Can this really freeze the network for unlimited time and bitcoin has
no mechanism to balance it back? A resourceful party with the intent
to attack the network in an irrational way, brings lot of hashing
power and keeps it for 2016 blocks, then removes it leaving the other
2016 blocks to solve at very high difficulty but with low hash power
in the network causing a 'blackout'? Thank you in advance for your
answers.


- --
s7r
PGP Fingerprint: 7C36 9232 5ABD FB0B 3021 03F1 837F A52C 8126 5B11
Jeff Garzik
2014-09-13 14:04:40 UTC
Permalink
This is why testnet has a special rule: If 20 minutes passes without
a block, you may mine a diff-1 block.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi there,
I have an contradictory discussion with an altcoin supporter and want
to bring some solid arguments in a public talk, so require little help
to respond to a question with simplest answer.
The problem is within the difficulty adjustment mechanism, that it
happens at every 2016 blocks. In case the hash power will suddenly
decrease, the 2016 blocks will take a lot of time to solve, therefor
freeze the network in a non-operational way. I know by far this is
just a joke, because this is very unlikely to happen anytime (people
paid for mining equipment and make money) but for the sake of
discussion, let's just assume it 'could' happen.
Can this really freeze the network for unlimited time and bitcoin has
no mechanism to balance it back? A resourceful party with the intent
to attack the network in an irrational way, brings lot of hashing
power and keeps it for 2016 blocks, then removes it leaving the other
2016 blocks to solve at very high difficulty but with low hash power
in the network causing a 'blackout'? Thank you in advance for your
answers.
- --
s7r
PGP Fingerprint: 7C36 9232 5ABD FB0B 3021 03F1 837F A52C 8126 5B11
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (MingW32)
iQEcBAEBAgAGBQJUFEOHAAoJEIN/pSyBJlsR3OgIAL0fvQ6MoftEdqoMZUHpjdYT
bRE6pXY52O5YIznaVljfbtQinApdWQUJ50VpiJd7KX3zOWFeo+KoUHyYlXTwoUEw
fK1cbrA5KC8IogHfTCcQxP/lE0mxgCSW3/vQda71IWaOyU4MUT2BSZ/4wwMctI4f
8tNVPPmzPo/wpO0+NOejowsGLmuzfzy0RMwCTXSbhPJuXr3yvdOTjzkfgy4TsJSU
K4RmBxqDh6LCrMQTpJfvqCUVEyqckG3AITQ2txdnWiMq1Ep94mvXb2L94Galmz00
yQzvo9F+G9HL0KgBAmZMrlFjsm58YDTblMTRKzBipozfuOk113XUecV+jHBvBrU=
=tcCa
-----END PGP SIGNATURE-----
------------------------------------------------------------------------------
Want excitement?
Manually upgrade your production database.
When you want reliability, choose Perforce
Perforce version control. Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
bitcoin-list mailing list
https://lists.sourceforge.net/lists/listinfo/bitcoin-list
--
Jeff Garzik
Bitcoin core developer and open source evangelist
BitPay, Inc. https://bitpay.com/
s7r
2014-09-13 16:00:56 UTC
Permalink
Post by Jeff Garzik
This is why testnet has a special rule: If 20 minutes passes
without a block, you may mine a diff-1 block.
Thanks for your point of view Mr. Garzik.
Why aren't we using this in the real bitcoin network too? If it's good
for balancing the functionality of the network in the context of
sudden hashrate moves? Specially down moves, since if the hash power
goes UP, the network will see blocks are created more often than at
every 10 minutes and adjust the difficulty directly proportional, correct?
Post by Jeff Garzik
there,
I have an contradictory discussion with an altcoin supporter and
want to bring some solid arguments in a public talk, so require
little help to respond to a question with simplest answer.
The problem is within the difficulty adjustment mechanism, that it
happens at every 2016 blocks. In case the hash power will suddenly
decrease, the 2016 blocks will take a lot of time to solve,
therefor freeze the network in a non-operational way. I know by far
this is just a joke, because this is very unlikely to happen
anytime (people paid for mining equipment and make money) but for
the sake of discussion, let's just assume it 'could' happen.
Can this really freeze the network for unlimited time and bitcoin
has no mechanism to balance it back? A resourceful party with the
intent to attack the network in an irrational way, brings lot of
hashing power and keeps it for 2016 blocks, then removes it leaving
the other 2016 blocks to solve at very high difficulty but with low
hash power in the network causing a 'blackout'? Thank you in
advance for your answers.
------------------------------------------------------------------------------
Want excitement?
Post by Jeff Garzik
Manually upgrade your production database. When you want
reliability, choose Perforce Perforce version control.
Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
Post by Jeff Garzik
https://lists.sourceforge.net/lists/listinfo/bitcoin-list
Jeff Garzik
2014-09-13 16:34:03 UTC
Permalink
Why not on mainnet? It is simply much lower priority for bitcoin
today than other problems. It tends not to happen at higher
difficulties for a variety of reasons.

The difficulty adjustment is capped at a factor of 4, up or down, to
address some other attacks that can happen in mining.

Personally, I think there needs to be a mainnet safety rule such as
"if 24 hours goes by without a block, you may mine a block" etc. But
I readily admit I've not thought through all the ramifications of such
a policy.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Post by Jeff Garzik
This is why testnet has a special rule: If 20 minutes passes
without a block, you may mine a diff-1 block.
Thanks for your point of view Mr. Garzik.
Why aren't we using this in the real bitcoin network too? If it's good
for balancing the functionality of the network in the context of
sudden hashrate moves? Specially down moves, since if the hash power
goes UP, the network will see blocks are created more often than at
every 10 minutes and adjust the difficulty directly proportional, correct?
Post by Jeff Garzik
there,
I have an contradictory discussion with an altcoin supporter and
want to bring some solid arguments in a public talk, so require
little help to respond to a question with simplest answer.
The problem is within the difficulty adjustment mechanism, that it
happens at every 2016 blocks. In case the hash power will suddenly
decrease, the 2016 blocks will take a lot of time to solve,
therefor freeze the network in a non-operational way. I know by far
this is just a joke, because this is very unlikely to happen
anytime (people paid for mining equipment and make money) but for
the sake of discussion, let's just assume it 'could' happen.
Can this really freeze the network for unlimited time and bitcoin
has no mechanism to balance it back? A resourceful party with the
intent to attack the network in an irrational way, brings lot of
hashing power and keeps it for 2016 blocks, then removes it leaving
the other 2016 blocks to solve at very high difficulty but with low
hash power in the network causing a 'blackout'? Thank you in
advance for your answers.
------------------------------------------------------------------------------
Want excitement?
Post by Jeff Garzik
Manually upgrade your production database. When you want
reliability, choose Perforce Perforce version control.
Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
Post by Jeff Garzik
https://lists.sourceforge.net/lists/listinfo/bitcoin-list
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (MingW32)
iQEcBAEBAgAGBQJUFGo4AAoJEIN/pSyBJlsR3d0IAISkymj6mxOkifdmp6ujUL4y
7lVoROugxAKTen9Fhg4rtWC10HQkTClJVfmaUYb+3D+oJ6YFjvZeyYT9TxFBrnvC
JfKG6m/yc9yp/R1MwSL81ez8TQvBt1UUVZcxApYW1TWXJDH95ua5IakQDkag/dET
HUtCAabPTDtQf0UaFqcycVXcXRYjvH73pOOD5j4WBeW1M2kd7pLm9Zdh1Up7nWVK
hfISwfq2S39vMBb5474/WP38YymW0izjh9yrxMaNT3MeuxR3PUo5ue9O470+YP5Y
5k03vs+qF3GWYRIy+13x//WeiYPQQxONjxb4+mgcfoYpXjx611VKPpYjZGarTNU=
=JCev
-----END PGP SIGNATURE-----
--
Jeff Garzik
Bitcoin core developer and open source evangelist
BitPay, Inc. https://bitpay.com/
s7r
2014-09-13 16:45:18 UTC
Permalink
Post by Jeff Garzik
Why not on mainnet? It is simply much lower priority for bitcoin
today than other problems. It tends not to happen at higher
difficulties for a variety of reasons.
The difficulty adjustment is capped at a factor of 4, up or down,
to address some other attacks that can happen in mining.
Capped at a factor of 4, up or down? How exactly in a simple example?
Post by Jeff Garzik
Personally, I think there needs to be a mainnet safety rule such
as "if 24 hours goes by without a block, you may mine a block" etc.
But I readily admit I've not thought through all the ramifications
of such a policy.
Post by Jeff Garzik
This is why testnet has a special rule: If 20 minutes
passes without a block, you may mine a diff-1 block.
Thanks for your point of view Mr. Garzik. Why aren't we using this
in the real bitcoin network too? If it's good for balancing the
functionality of the network in the context of sudden hashrate
moves? Specially down moves, since if the hash power goes UP, the
network will see blocks are created more often than at every 10
minutes and adjust the difficulty directly proportional, correct?
Post by Jeff Garzik
Hi there,
I have an contradictory discussion with an altcoin supporter
and want to bring some solid arguments in a public talk, so
require little help to respond to a question with simplest
answer.
The problem is within the difficulty adjustment mechanism,
that it happens at every 2016 blocks. In case the hash power
will suddenly decrease, the 2016 blocks will take a lot of
time to solve, therefor freeze the network in a
non-operational way. I know by far this is just a joke,
because this is very unlikely to happen anytime (people paid
for mining equipment and make money) but for the sake of
discussion, let's just assume it 'could' happen.
Can this really freeze the network for unlimited time and
bitcoin has no mechanism to balance it back? A resourceful
party with the intent to attack the network in an irrational
way, brings lot of hashing power and keeps it for 2016
blocks, then removes it leaving the other 2016 blocks to
solve at very high difficulty but with low hash power in the
network causing a 'blackout'? Thank you in advance for your
answers.
------------------------------------------------------------------------------
Want excitement?
Post by Jeff Garzik
Post by Jeff Garzik
Manually upgrade your production database. When you want
reliability, choose Perforce Perforce version control.
Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
Post by Jeff Garzik
Post by Jeff Garzik
bitcoin-list mailing list
https://lists.sourceforge.net/lists/listinfo/bitcoin-list
Jeff Garzik
2014-09-13 17:00:14 UTC
Permalink
If difficulty is X, range for adjustment is X/4 to X*4.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Post by Jeff Garzik
Why not on mainnet? It is simply much lower priority for bitcoin
today than other problems. It tends not to happen at higher
difficulties for a variety of reasons.
The difficulty adjustment is capped at a factor of 4, up or down,
to address some other attacks that can happen in mining.
Capped at a factor of 4, up or down? How exactly in a simple example?
Post by Jeff Garzik
Personally, I think there needs to be a mainnet safety rule such
as "if 24 hours goes by without a block, you may mine a block" etc.
But I readily admit I've not thought through all the ramifications
of such a policy.
Post by Jeff Garzik
This is why testnet has a special rule: If 20 minutes
passes without a block, you may mine a diff-1 block.
Thanks for your point of view Mr. Garzik. Why aren't we using this
in the real bitcoin network too? If it's good for balancing the
functionality of the network in the context of sudden hashrate
moves? Specially down moves, since if the hash power goes UP, the
network will see blocks are created more often than at every 10
minutes and adjust the difficulty directly proportional, correct?
Post by Jeff Garzik
Hi there,
I have an contradictory discussion with an altcoin supporter
and want to bring some solid arguments in a public talk, so
require little help to respond to a question with simplest
answer.
The problem is within the difficulty adjustment mechanism,
that it happens at every 2016 blocks. In case the hash power
will suddenly decrease, the 2016 blocks will take a lot of
time to solve, therefor freeze the network in a
non-operational way. I know by far this is just a joke,
because this is very unlikely to happen anytime (people paid
for mining equipment and make money) but for the sake of
discussion, let's just assume it 'could' happen.
Can this really freeze the network for unlimited time and
bitcoin has no mechanism to balance it back? A resourceful
party with the intent to attack the network in an irrational
way, brings lot of hashing power and keeps it for 2016
blocks, then removes it leaving the other 2016 blocks to
solve at very high difficulty but with low hash power in the
network causing a 'blackout'? Thank you in advance for your
answers.
------------------------------------------------------------------------------
Want excitement?
Post by Jeff Garzik
Post by Jeff Garzik
Manually upgrade your production database. When you want
reliability, choose Perforce Perforce version control.
Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
Post by Jeff Garzik
Post by Jeff Garzik
bitcoin-list mailing list
https://lists.sourceforge.net/lists/listinfo/bitcoin-list
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (MingW32)
iQEcBAEBAgAGBQJUFHSeAAoJEIN/pSyBJlsRQ7gH/3iB0sy4PRi03dbcq5mKiwc4
aBlG2zt8rVRqX6KkCoeijY4lysg3KYzhyBBUna4bwGAhpvBe0jjybEWFAMqM2h6+
7JzUdcZ9K5HuOtQUY2/4jHAz6jg3LOCBvtJUmLtC1HGSGHSl0tFiRVc2bzWrIOsT
zkq87G3ocfSNiQSRSKqhxM1W4tc1BUHBr6T0ZS2EyArZljnIN5WVfw3ueTPZizdM
q51+krxBjIW5iEkS4pVNChlsMaMca6JLVOY8jJpEWDf0vTX9zsheMdDg0MyDfMn5
ObywubBmV+tawNsHU7DWwliIiCoMr2h2G7KpV8V/np1iVyaSribaIj1hM/rWvh8=
=Q1i9
-----END PGP SIGNATURE-----
--
Jeff Garzik
Bitcoin core developer and open source evangelist
BitPay, Inc. https://bitpay.com/
Michael Wozniak
2014-09-13 16:45:39 UTC
Permalink
The biggest problem with a time based rule is that there is no central time to provide a measure of how much time has actually passed. You can see this issue on testnet already because block timestamps are modified. For example, the current time is (according to my server) 1410626564 and the best block on testnet shows a timestamp of 1410628433 (30 minutes in the future).
Post by Jeff Garzik
Why not on mainnet? It is simply much lower priority for bitcoin
today than other problems. It tends not to happen at higher
difficulties for a variety of reasons.
The difficulty adjustment is capped at a factor of 4, up or down, to
address some other attacks that can happen in mining.
Personally, I think there needs to be a mainnet safety rule such as
"if 24 hours goes by without a block, you may mine a block" etc. But
I readily admit I've not thought through all the ramifications of such
a policy.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Post by Jeff Garzik
This is why testnet has a special rule: If 20 minutes passes
without a block, you may mine a diff-1 block.
Thanks for your point of view Mr. Garzik.
Why aren't we using this in the real bitcoin network too? If it's good
for balancing the functionality of the network in the context of
sudden hashrate moves? Specially down moves, since if the hash power
goes UP, the network will see blocks are created more often than at
every 10 minutes and adjust the difficulty directly proportional, correct?
Post by Jeff Garzik
there,
I have an contradictory discussion with an altcoin supporter and
want to bring some solid arguments in a public talk, so require
little help to respond to a question with simplest answer.
The problem is within the difficulty adjustment mechanism, that it
happens at every 2016 blocks. In case the hash power will suddenly
decrease, the 2016 blocks will take a lot of time to solve,
therefor freeze the network in a non-operational way. I know by far
this is just a joke, because this is very unlikely to happen
anytime (people paid for mining equipment and make money) but for
the sake of discussion, let's just assume it 'could' happen.
Can this really freeze the network for unlimited time and bitcoin
has no mechanism to balance it back? A resourceful party with the
intent to attack the network in an irrational way, brings lot of
hashing power and keeps it for 2016 blocks, then removes it leaving
the other 2016 blocks to solve at very high difficulty but with low
hash power in the network causing a 'blackout'? Thank you in
advance for your answers.
------------------------------------------------------------------------------
Want excitement?
Post by Jeff Garzik
Manually upgrade your production database. When you want
reliability, choose Perforce Perforce version control.
Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
Post by Jeff Garzik
https://lists.sourceforge.net/lists/listinfo/bitcoin-list
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (MingW32)
iQEcBAEBAgAGBQJUFGo4AAoJEIN/pSyBJlsR3d0IAISkymj6mxOkifdmp6ujUL4y
7lVoROugxAKTen9Fhg4rtWC10HQkTClJVfmaUYb+3D+oJ6YFjvZeyYT9TxFBrnvC
JfKG6m/yc9yp/R1MwSL81ez8TQvBt1UUVZcxApYW1TWXJDH95ua5IakQDkag/dET
HUtCAabPTDtQf0UaFqcycVXcXRYjvH73pOOD5j4WBeW1M2kd7pLm9Zdh1Up7nWVK
hfISwfq2S39vMBb5474/WP38YymW0izjh9yrxMaNT3MeuxR3PUo5ue9O470+YP5Y
5k03vs+qF3GWYRIy+13x//WeiYPQQxONjxb4+mgcfoYpXjx611VKPpYjZGarTNU=
=JCev
-----END PGP SIGNATURE-----
--
Jeff Garzik
Bitcoin core developer and open source evangelist
BitPay, Inc. https://bitpay.com/
------------------------------------------------------------------------------
Want excitement?
Manually upgrade your production database.
When you want reliability, choose Perforce
Perforce version control. Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
bitcoin-list mailing list
https://lists.sourceforge.net/lists/listinfo/bitcoin-list
Jeff Garzik
2014-09-13 17:09:03 UTC
Permalink
That is conflating two different things.

Block timestamps are never "modified." Miners are permitted to mine
timestamps within a certain time range (e.g. no more than 2 hours into
the future). That is normal for testnet or mainnet.
Post by Michael Wozniak
The biggest problem with a time based rule is that there is no central time to provide a measure of how much time has actually passed. You can see this issue on testnet already because block timestamps are modified. For example, the current time is (according to my server) 1410626564 and the best block on testnet shows a timestamp of 1410628433 (30 minutes in the future).
Post by Jeff Garzik
Why not on mainnet? It is simply much lower priority for bitcoin
today than other problems. It tends not to happen at higher
difficulties for a variety of reasons.
The difficulty adjustment is capped at a factor of 4, up or down, to
address some other attacks that can happen in mining.
Personally, I think there needs to be a mainnet safety rule such as
"if 24 hours goes by without a block, you may mine a block" etc. But
I readily admit I've not thought through all the ramifications of such
a policy.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Post by Jeff Garzik
This is why testnet has a special rule: If 20 minutes passes
without a block, you may mine a diff-1 block.
Thanks for your point of view Mr. Garzik.
Why aren't we using this in the real bitcoin network too? If it's good
for balancing the functionality of the network in the context of
sudden hashrate moves? Specially down moves, since if the hash power
goes UP, the network will see blocks are created more often than at
every 10 minutes and adjust the difficulty directly proportional, correct?
Post by Jeff Garzik
there,
I have an contradictory discussion with an altcoin supporter and
want to bring some solid arguments in a public talk, so require
little help to respond to a question with simplest answer.
The problem is within the difficulty adjustment mechanism, that it
happens at every 2016 blocks. In case the hash power will suddenly
decrease, the 2016 blocks will take a lot of time to solve,
therefor freeze the network in a non-operational way. I know by far
this is just a joke, because this is very unlikely to happen
anytime (people paid for mining equipment and make money) but for
the sake of discussion, let's just assume it 'could' happen.
Can this really freeze the network for unlimited time and bitcoin
has no mechanism to balance it back? A resourceful party with the
intent to attack the network in an irrational way, brings lot of
hashing power and keeps it for 2016 blocks, then removes it leaving
the other 2016 blocks to solve at very high difficulty but with low
hash power in the network causing a 'blackout'? Thank you in
advance for your answers.
------------------------------------------------------------------------------
Want excitement?
Post by Jeff Garzik
Manually upgrade your production database. When you want
reliability, choose Perforce Perforce version control.
Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
Post by Jeff Garzik
https://lists.sourceforge.net/lists/listinfo/bitcoin-list
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (MingW32)
iQEcBAEBAgAGBQJUFGo4AAoJEIN/pSyBJlsR3d0IAISkymj6mxOkifdmp6ujUL4y
7lVoROugxAKTen9Fhg4rtWC10HQkTClJVfmaUYb+3D+oJ6YFjvZeyYT9TxFBrnvC
JfKG6m/yc9yp/R1MwSL81ez8TQvBt1UUVZcxApYW1TWXJDH95ua5IakQDkag/dET
HUtCAabPTDtQf0UaFqcycVXcXRYjvH73pOOD5j4WBeW1M2kd7pLm9Zdh1Up7nWVK
hfISwfq2S39vMBb5474/WP38YymW0izjh9yrxMaNT3MeuxR3PUo5ue9O470+YP5Y
5k03vs+qF3GWYRIy+13x//WeiYPQQxONjxb4+mgcfoYpXjx611VKPpYjZGarTNU=
=JCev
-----END PGP SIGNATURE-----
--
Jeff Garzik
Bitcoin core developer and open source evangelist
BitPay, Inc. https://bitpay.com/
------------------------------------------------------------------------------
Want excitement?
Manually upgrade your production database.
When you want reliability, choose Perforce
Perforce version control. Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
bitcoin-list mailing list
https://lists.sourceforge.net/lists/listinfo/bitcoin-list
------------------------------------------------------------------------------
Want excitement?
Manually upgrade your production database.
When you want reliability, choose Perforce
Perforce version control. Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
bitcoin-list mailing list
https://lists.sourceforge.net/lists/listinfo/bitcoin-list
s7r
2014-09-13 17:39:32 UTC
Permalink
Post by Michael Wozniak
The biggest problem with a time based rule is that there is no
central time to provide a measure of how much time has actually
passed. You can see this issue on testnet already because block
timestamps are modified. For example, the current time is
(according to my server) 1410626564 and the best block on testnet
shows a timestamp of 1410628433 (30 minutes in the future).
This problem can be addressed, like the problem of transaction time
and order of transactions in blocks. For sure maybe you cannot get
100% time accuracy, but you can know the time with +/- some error
margin for time window.

And as for the other algorithms used by altcois, i don't have one to
give as an example but can't all of them be unstable. And of course,
it is not a problem now and most probably will never be, but if it's
simple to implement why not take safety precautions and be prepared
even for the unexpected / impossible threat? I mean, isn't this what
cryptocoin decentralized money is all about? :)
Post by Michael Wozniak
Post by Jeff Garzik
Why not on mainnet? It is simply much lower priority for
bitcoin today than other problems. It tends not to happen at
higher difficulties for a variety of reasons.
The difficulty adjustment is capped at a factor of 4, up or down,
to address some other attacks that can happen in mining.
Personally, I think there needs to be a mainnet safety rule such
as "if 24 hours goes by without a block, you may mine a block"
etc. But I readily admit I've not thought through all the
ramifications of such a policy.
Post by Jeff Garzik
This is why testnet has a special rule: If 20 minutes
passes without a block, you may mine a diff-1 block.
Thanks for your point of view Mr. Garzik. Why aren't we using this
in the real bitcoin network too? If it's good for balancing the
functionality of the network in the context of sudden hashrate
moves? Specially down moves, since if the hash power goes UP, the
network will see blocks are created more often than at every 10
minutes and adjust the difficulty directly proportional, correct?
Post by Jeff Garzik
Post by Jeff Garzik
wrote: Hi there,
I have an contradictory discussion with an altcoin
supporter and want to bring some solid arguments in a
public talk, so require little help to respond to a
question with simplest answer.
The problem is within the difficulty adjustment mechanism,
that it happens at every 2016 blocks. In case the hash
power will suddenly decrease, the 2016 blocks will take a
lot of time to solve, therefor freeze the network in a
non-operational way. I know by far this is just a joke,
because this is very unlikely to happen anytime (people
paid for mining equipment and make money) but for the sake
of discussion, let's just assume it 'could' happen.
Can this really freeze the network for unlimited time and
bitcoin has no mechanism to balance it back? A resourceful
party with the intent to attack the network in an
irrational way, brings lot of hashing power and keeps it
for 2016 blocks, then removes it leaving the other 2016
blocks to solve at very high difficulty but with low hash
power in the network causing a 'blackout'? Thank you in
advance for your answers.
------------------------------------------------------------------------------
Want excitement?
Post by Michael Wozniak
Post by Jeff Garzik
Post by Jeff Garzik
Manually upgrade your production database. When you want
reliability, choose Perforce Perforce version control.
Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
Post by Michael Wozniak
Post by Jeff Garzik
Post by Jeff Garzik
bitcoin-list mailing list
https://lists.sourceforge.net/lists/listinfo/bitcoin-list
- --
Post by Michael Wozniak
Post by Jeff Garzik
Jeff Garzik Bitcoin core developer and open source evangelist
BitPay, Inc. https://bitpay.com/
------------------------------------------------------------------------------
Want excitement?
Post by Michael Wozniak
Post by Jeff Garzik
Manually upgrade your production database. When you want
reliability, choose Perforce Perforce version control.
Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
Post by Michael Wozniak
Post by Jeff Garzik
https://lists.sourceforge.net/lists/listinfo/bitcoin-list
------------------------------------------------------------------------------
Want excitement?
Post by Michael Wozniak
Manually upgrade your production database. When you want
reliability, choose Perforce Perforce version control. Predictably
reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
Post by Michael Wozniak
https://lists.sourceforge.net/lists/listinfo/bitcoin-list
s7r
2014-09-13 15:54:23 UTC
Permalink
Hello,
If the hashing power of the attacker is in the order of 50%, then
he can disrupt the network in several ways, including
double-spending and the blackout attack you are describing.
Bitcoin makes only the assumption that there is a majority of
honest node. If you break the assuption, you disrupt the network
for as long as the assumption is broken.
Back to your question, If the hashing power of the attacker is in
the order of 10%, the average time interval between blocks is
increased by 10%, i.e 1 minute when he pulls out of mining.
That is until the next difficulty adjustement, roughly 2 weeks and
36 hours later. Hardly a "blackout".
Indeed, but the scenario changes if the power is 80% if the hashing
power. If immediately after a difficulty adjustment the hashing power
drops by 80%, it will take way lot longer until the next difficulty
adjustment. The next 2016 blocks will be solved much much slower
preventing the correct functionality of the network, regardless if all
the remaining peers and miners are honest.

This is what matters, an assumption of this type of attack but with
60% of the hashing power - 10% won't do a lot of damage and hardly
create a blackout as you say.
Pierre
Hi there,
I have an contradictory discussion with an altcoin supporter and
want to bring some solid arguments in a public talk, so require
little help to respond to a question with simplest answer.
The problem is within the difficulty adjustment mechanism, that it
happens at every 2016 blocks. In case the hash power will suddenly
decrease, the 2016 blocks will take a lot of time to solve,
therefor freeze the network in a non-operational way. I know by far
this is just a joke, because this is very unlikely to happen
anytime (people paid for mining equipment and make money) but for
the sake of discussion, let's just assume it 'could' happen.
Can this really freeze the network for unlimited time and bitcoin
has no mechanism to balance it back? A resourceful party with the
intent to attack the network in an irrational way, brings lot of
hashing power and keeps it for 2016 blocks, then removes it leaving
the other 2016 blocks to solve at very high difficulty but with low
hash power in the network causing a 'blackout'? Thank you in
advance for your answers.
------------------------------------------------------------------------------
Want excitement?
Manually upgrade your production database. When you want
reliability, choose Perforce Perforce version control.
Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
https://lists.sourceforge.net/lists/listinfo/bitcoin-list
Loading...