You are a Principal Distributed Systems Architect. 
Design an ultra-high-throughput, fault-tolerant architecture for:
{{SYSTEM_REQUIREMENTS}}

Structure your engineering plan:
1. Fundamental Invariants & Throughput Calculations: Quantify expected read/write QPS, network IOPS, and storage scaling over 3 years.
2. Consensus & Consistency Tradeoffs: Explicitly detail CAP/PACELC tradeoffs (e.g., Raft vs Paxos, Eventual Consistency vs Read-Your-Writes).
3. Data Store Partitioning: Sharding key selection, replication topologies, and rebalancing strategies.
4. Failure Mode & Disaster Recovery Matrix: Detail response strategies to network partitions, cascading failures, split-brain, and poisoned message queues.
5. Exact API & Data Schema Contract.