0

I'm about to embark on a blockchain project for supply chain, and am investigating Hyperledger Sawtooth (https://www.youtube.com/watch?v=uBebFQM49Xk) and Hyperledger Fabric (https://drive.google.com/file/d/1OsIoPtlv5X2PWyOAlDn1FCnHCZPyrF57/view) at the moment. it appears that the abovementioned frameworks are capable of "thousands" of transactions per second (tps).

My question has to do with my planned use case. if i'm planning to track metrics of a certain supply, and i need it to be updated every minute, and there are hundreds of thousands of this supply at any point in time, how can this scale? i'm assuming that as things are queued latency increases. At 150,000 event reporting per second (which seems to be conservative based on our calculations), we would incur tens to hundreds of seconds of delay.

is my simple math usable? are there any mechanisms to work around this?

thanks

soler
  • 112
  • 9

1 Answers1

0

As per my understanding with hyperledger, it can support 100,000 transaction per hour(which is approx 28/sec.) with a large cloud setup. I think, the kind of transaction limit you want would be difficult to achieve as of now.

Gaurang Singh
  • 188
  • 9
  • Hyperledger is a consortium under the Linux Foundation with several blockchain platforms. There may be claims for 100000 tps for Hyperledger Fabric (I have not seen that claim for Hyperledger Sawtooth), but I would regard that with suspicion, if such a claim was made. Was it all local transactions on a single host or closely connected hosts? My point is one cannot expect the throughput of a database with a blockchain. – Dan Anderson Oct 12 '18 at 20:32
  • @DanAnderson I have written 100,000 per HOUR, not SECOND. Getting 100,000 per second is still difficult to achieve. – Gaurang Singh Oct 23 '18 at 12:05