This is the artifact evaluation submission #28 for our paper "Byzantine ordered consensus without Byzantine oligarchy".
We made 4 claims which can be found in Figure 3 of our submission. We copy-and-paste them here.
-
claim1: Archipelago achieves higher throughput than its baselines at the cost of increased latency (latencies are competitive when batching commands).
-
claim2: Archipelago’s throughput degrades similarly to its baselines when n increases, but Archipelago can scale up each node for higher throughput.
-
claim3: Archipelago incurs modest network overheads over its baselines.
-
claim4: A geo-distributed deployment increases latencies, but Archipelago’s peak throughput remains similar to a single datacenter setting.
Update: the latest repos are here: https://github.com/Pompe-org
We implemented 2 artifacts validating the claims.
-
Archipelago-HotStuff: the git repo is here. Please read its default
README.mdfor instructions and contact Yunhao Zhang (yz2327@cornell.edu) for any questions about this artifact. This artifact validates all claims. -
Archipelago-Concord: the git repo is here. Please switch to the add_archipelago branch and read
eval/README.mdfor instructions. Contact Qi Chen (cheqi@microsoft.com) for any questions about this artifact. This artifact only validates the first two claims.