mohit-goel/bcr

Byzantine Chain Replication

PLATFORM:

• DistAlgo version -- 1.0.9
• Python Implementation CPython 3.6.3
• Operating System: macOS Sierra 10.12.6 (16G29)
• PyNacl: 1.1.2

INSTRUCTIONS:

• Install Python3
• Install Distalgo: "sudo pip3 install pyDistAlgo"
• Install PyNacl: "sudo pip3 install pynacl"

MAIN FILES.

• bcr.da : This file contains code for master to read configuration file from command line and initiate Olympus and Clients.
• client.da :This file contains code for clients to send client requests and receive client responses, configuration request & response to Olympus
• olympus.da : This file contains code for olympus to initiate replicas and hsndling configuration requests from clients
• replica.da : This file contains code for replica to handle client requests and other replica messages.

SINGLE PROCESS SETUP

python3 -m da --logfile --logfilename testCaseName.txt --logfilelevel 'info' --message-buffer-size size_Bytes bcr.da configurationFileName.txt

MULTI-NODE SETUP

• python3 -m da -r --message-buffer-size=1280000 --logfile --logfilename logdirect/replica0.log --logfilelevel info -n ReplicaNode0 -D replica.da
• python3 -m da -r --message-buffer-size=1280000 --logfile --logfilename logdirect/replica1.log --logfilelevel info -n ReplicaNode1 -D replica.da
• python3 -m da -r --message-buffer-size=1280000 --logfile --logfilename logdirect/replica2.log --logfilelevel info -n ReplicaNode2 -D replica.da
• python3 -m da -r --message-buffer-size=1280000 --logfile --logfilename logdirect/client.log --logfilelevel info -n ClientNode -D client.da
• python3 -m da -r --message-buffer-size=1280000 --logfile --logfilename logdirect/olympus.log --logfilelevel info -n OlympusNode -D olympus.da
• python3 -m da -r --message-buffer-size=1280000 --logfile --logfilename logdirect/master.log --logfilelevel info -n MasterNode bcr.da perform900.txt

WORKLOAD GENERATION:

class PseudoRandom:

   def select_random_elements_of_list(self, list, num_required_elements, seed):
       len_of_list = len(list)
       sample = []
       random.seed(seed)
       for i in range(num_required_elements):
           index = random.randint(0, len_of_list - 1)
           sample.append(list[index])

       return sample

num_required_elements : number of requests to be generated

list : specify the list present in configuration.csv from which request will be picked up randomly

We are using python in-built 'random' utiltiy to generate 'num_required_elements' numbers corresponding to one of the indexes in the list.

PERFORMANCE EVALUATION

• RAFT : 8.78 seconds
• Single-host configuration : 61 seconds
• Multi-Host setup with replicas on different host : 60 seconds