1. Messages
  2. Performance Difference bet RAFT Orderer and Orderer with Kafka(Latency, Throughput, TPS)

Re: Performance Difference bet RAFT Orderer and Orderer with Kafka(Latency, Throughput, TPS)

soumya nayak <soumyarjnnayak@...>
 

Hi pavan, 

From my experience I did not find much latency. it was very similar in both the cases. 

Fabric - v1. 4.3
Kafka - 4 brokers (all in 4 different azure cloud VM) 
Orderers - 2 ( 2 different vms) 
AZURE VM - 2 VCPU, 8 GB RAM, 200GB SPACE

Raft orderers- 3 nodes (all in 3 different azure VMs) 

Uploaded 21 million records from a sql server to blockchain set up in above both the configurations set up. 

We did 900-1000 transactions as per the block creation cutoff and sometimes we were able to achieve around 1300-1400 trans also for both the above cases. 

Even when I am trying to fetch also records from the 21 million with a load of 30-40 concurrent users iam able to fetch the records in less than a second in both the cases. 

If you can tell us more details about your set up would be great. 

Regards, 
Soumya

On Mon 30 Dec, 2019, 3:34 AM Baohua Yang, <yangbaohua@...> wrote:
Hi pavan

How many kafka nodes are you using?  And is it in the same VMWARE with OSN? 

Would expect the CPU utilization is different here,  the kafka one is higher? 

On Dec 28, 2019, at 20:52, Adhav Pavan <adhavpavan@...> wrote:


Hello Experts,

Did anyone compare performance(Latency, Throughput, TPS) between orderer with Kafka and RAFT Orderer?

I could see here a considerable difference in terms of latency, throughput, and TPS.

I tried with the same setup with the same resource configuration on two different VM(the Only difference is the orderer system).

Note: Used Single orderer in both networks.
Fabric Version: 1.4.4

Orderer with Kafka is more efficient than RAFT. 
I am using the default configuration for RAFT and Kafka

I tried with load generator at the rate of 100tps
WIth Kafka all parameters are fine(latency 0.3 to 2 Seconds) whereas using RAFT, latency is gradually increasing up to 10+ Seconds, the tx failure rate is also high.

What could be the reason for this considerable difference in terms of TPS, throughput, and latency?

Please correct If I am doing something wrong.

Thank you so much.

Heartfelt Regards,
Pavan Adhav

Blockchain Developer
Cell Phone:+91-8390114357  E-Mail: adhavpavan@...
Join fabric@lists.hyperledger.org to automatically receive all group messages.