Difference between revisions of "Antifragility"

From Obyte Wiki
(added wikilink; brief data about Feb tests)
(typo)
 
(5 intermediate revisions by 2 users not shown)
Line 1: Line 1:
 
==Why stress tests?==
 
==Why stress tests?==
Antifragility is a property of systems that increase in capability, resilience, or robustness as a result of stressors, shocks, volatility, noise, mistakes, faults, attacks, or failures. It is a concept developed by Professor Nassim Nicholas Taleb in his book, Antifragile and in technical papers. <ref>[https://en.wikipedia.org/wiki/Antifragility Wikipedia: Antifragility]</ref>
+
Antifragility is a property of systems that increase in capability, resilience, or robustness as a result of stressors, shocks, volatility, noise, mistakes, faults, attacks, or failures. It is a concept developed by Professor Nassim Nicholas Taleb in his book ''Antifragile'' and in technical papers. <ref>[https://en.wikipedia.org/wiki/Antifragility Wikipedia: Antifragility]</ref>
  
It applies to any cryptocurrency in that a coin's main [[platform]] must be frequently stress-tested as it matures, otherwise the live production will break unexpectedly, maybe with disastrous results. Sometimes, like an exchange being severely hacked, it can mean the end of that platform.
+
It applies to any cryptocurrency in that a coin's main [[platform]] must be frequently stress-tested as it matures, otherwise the live production will break unexpectedly, maybe with disastrous results. Sometimes, such as an exchange being severely hacked, it can mean the end of that platform.
  
==Byteball stress tests==
+
==TPS Transactions per second ==
These are community-produced tests, with no "team" input (at present). They are experiments.
+
=== 2017 Comment ===
 +
<blockquote>'''Tonych 2017-02-07:''' You know, there is no architectural limit in the DAGs. Regarding the practical limits, I don't buy into this race to Visa tps. The most pressing issue of crypto is not tps, it is adoption (which we address in the first place). Tps will come second after the first is solved.<ref>https://bitcointalk.org/index.php?topic=1608859.msg17789094#msg17789094</ref></blockquote>
  
===Jan 2018===
+
=== 2019 Comments ===  
@portabella has agreed to do one "next week", i.e. Jan 7-13. He provided the address 3JI5LUOH755Y3XV2IYF25M2YG2MNHZST, but it is not clear what for.
+
<blockquote>'''Tonych 2019-01-06:''' After recent optimizations, I synced a full node in 1.5 days, average speed 30 tps</blockquote>
  
The idea is to see if the platform can cope with a relatively large volume of transactions: 10 tps (transactions per second)? 50tps? 100tps? 500tps? 1000tps for 5 minutes?
+
<blockquote>'''Punqtured 2019-01-26:''' [The 30 tps is for both syncing up a full node and active transactions.] The bottleneck is currently when a full [[node]] establishes order between transactions by looking at their positions in relation to transactions posted by witnesses. This re-ordering of transactions is what is currently the bottleneck. When a full node starts to syncronize, it does exactly the same re-ordering of units as it would [with] "new" units. The node receives the new units from the hub and establishes the main chain. So if the hub feeds a node 30 units per second from the past or feeds the node 30 units per second from the present, doesn't make any difference.</blockquote>
  
===Feb 2018===
+
<blockquote>'''tarmo888 2019-01-26:''' Any time there are more transactions per second than your node is able to process, your node goes into catch-up mode, so temporarily there could be bursts of 1000 TPS and nothing bad would happen if the TPS dropped back to 1 TPS after that. The 30 TPS is averaged over a longer period of time, so Obyte is currently able to sustain 30 TPS over a long time and much higher TPS for short bursts [minor grammatical edits].</blockquote>
@seb486 did a series of 3 stress tests.
 
  
==Comment==
 
<blockquote>'''Tonych 2017-02-07:''' You know, there is no architectural limit in the DAGs. Regarding the practical limits, I don't buy into this race to Visa tps. The most pressing issue of crypto is not tps, it is adoption (which we address in the first place). Tps will come second after the first is solved.<ref>https://bitcointalk.org/index.php?topic=1608859.msg17789094#msg17789094</ref></blockquote>
 
 
==External links==
 
==External links==
  

Latest revision as of 17:20, 26 January 2019

Why stress tests?

Antifragility is a property of systems that increase in capability, resilience, or robustness as a result of stressors, shocks, volatility, noise, mistakes, faults, attacks, or failures. It is a concept developed by Professor Nassim Nicholas Taleb in his book Antifragile and in technical papers. [1]

It applies to any cryptocurrency in that a coin's main platform must be frequently stress-tested as it matures, otherwise the live production will break unexpectedly, maybe with disastrous results. Sometimes, such as an exchange being severely hacked, it can mean the end of that platform.

TPS Transactions per second

2017 Comment

Tonych 2017-02-07: You know, there is no architectural limit in the DAGs. Regarding the practical limits, I don't buy into this race to Visa tps. The most pressing issue of crypto is not tps, it is adoption (which we address in the first place). Tps will come second after the first is solved.[2]

2019 Comments

Tonych 2019-01-06: After recent optimizations, I synced a full node in 1.5 days, average speed 30 tps

Punqtured 2019-01-26: [The 30 tps is for both syncing up a full node and active transactions.] The bottleneck is currently when a full node establishes order between transactions by looking at their positions in relation to transactions posted by witnesses. This re-ordering of transactions is what is currently the bottleneck. When a full node starts to syncronize, it does exactly the same re-ordering of units as it would [with] "new" units. The node receives the new units from the hub and establishes the main chain. So if the hub feeds a node 30 units per second from the past or feeds the node 30 units per second from the present, doesn't make any difference.

tarmo888 2019-01-26: Any time there are more transactions per second than your node is able to process, your node goes into catch-up mode, so temporarily there could be bursts of 1000 TPS and nothing bad would happen if the TPS dropped back to 1 TPS after that. The 30 TPS is averaged over a longer period of time, so Obyte is currently able to sustain 30 TPS over a long time and much higher TPS for short bursts [minor grammatical edits].

External links

References