Active-Active Data Architecture for High-Availability Enterprise Systems
DOI:
https://doi.org/10.15662/IJEETR.2023.0501006Keywords:
active-active data architecture, multi-region replication, high availability, distributed databases, failover, quorum consistency, conflict resolution, enterprise systems, data platform architectureAbstract
Active-active data architectures underpin a wide class of enterprise systems that must remain continuously available across geographically distributed regions, even in the presence of whole-site failures. Yet the engineering literature still treats active-active as a binary deployment choice rather than a spectrum of design tradeoffs. This article develops an architectural treatment of active-active data systems for high-availability enterprise platforms. We position availability as a graded property rather than a guarantee, distinguish the four core data dimensions — durability, consistency, conflict resolution, and latency — that compose any active-active design, and analyze the implementation patterns that emerge when each dimension is intensified: multi-region synchronous replication, quorum-based consistency, last-writer-wins versus operational-transform conflict handling, and read–write regional steering. We then examine the operational patterns that determine whether an active-active design actually meets its availability objective in practice: symmetric and rehearsed failover, topology-aware observability, symmetry-safe deployment, and schema evolution that remains compatible across regions during rolling upgrades. We close with a suitability framework that maps a given workload — writes-per-second, dependency on strong invariants, tolerable staleness, and regulated-data constraints — to a concrete active-active posture. The result is a practitioner’s map for deciding when active-active is warranted, which flavor, and at what cost, rather than a vendor pitch for any single topology
References
[1] E. Brewer, “CAP twelve years later: How the ‘rules’ have changed,” Computer, vol. 45, no. 2, pp. 23–29, 2012.
[2] S. Gilbert and N. Lynch, “Perspectives on the CAP theorem,” Communications of the ACM, vol. 55, no. 3, pp. 86–93, 2012.
[3] D. Abadi, “Consistency tradeoffs in modern distributed database system design,” ACM SIGMOD Record, vol. 41, no. 1, pp. 9–16, 2012.
[4] H. Howard, D. Malkhi, and A. Spiegelman, “Flexible Paxos: Quorum intersection revisited,” arXiv preprint arXiv:1608.06696, 2016.
[5] R. Yadav and A. Rahut, “FlexiRaft: Flexible quorums with Raft,” in Proceedings of the Conference on Innovative Data Systems Research (CIDR), 2023.
[6] D. Ongaro and J. Ousterhout, “In search of an understandable consensus algorithm,” in Proceedings of the USENIX Annual Technical Conference, pp. 305–319, 2014.
[7] L. Lamport, “Paxos made simple,” ACM SIGACT News, vol. 32, no. 4, pp. 51–58, 2001.
[8] V. Arora, T. Mittal, D. Agrawal, and A. El Abbadi, “Leader or majority: Why have one when you can have both? Improving read scalability in Raft-like consensus protocols,” in Proceedings of the USENIX Workshop on Hot Topics in Cloud Computing (HotCloud), 2017.
[9] M. Shapiro, N. Preguiica, C. Baquero, and M. Zawirski, “A comprehensive study of Convergent and Commutative Replicated Data Types,” INRIA Technical Report RR-7506, 2011.
