Build data-driven EdgeWorkers applications with a globally distributed, serverless key‑value store designed for fast, frequent reads and simple operations. EdgeKV keeps data close to your business logic and your users — without schemas, ORMs, or database administration.
EdgeKV is a distributed key‑value store purpose‑built for EdgeWorkers. It stores unstructured or semi‑structured data (strings and JSON) and exposes simple get/put/delete operations. There’s no schema or complex query language to learn — think of it like a dictionary keyed by unique IDs.
How EdgeKV delivers low latency and resilience: - Global distribution with local copies on executing edge servers. Popular data is cached at the edge where your EdgeWorker runs and kept close to users for cache‑like read speeds. - Eventual consistency optimized for read‑heavy workloads. Writes and deletes propagate through a publisher/subscriber‑like mechanism that automatically invalidates stale local copies across the network. The platform targets eventual consistency latency of 10 seconds or less for at least 80% of operations. - A simple data model for scale and organization. Items (key‑value pairs) are grouped by developer‑defined groups within durable namespaces. Namespaces include retention policies and are administered via role‑based access control.
Helpful tooling and integrations: - EdgeKV Helper Library for EdgeWorkers to simplify reads and writes - Administrative APIs for full CRUD and token management - A scriptable CLI integrated with the EdgeWorkers CLI for day‑to‑day admin and automation
Explore technical resources: - Read the EdgeKV User Guide - Browse the EdgeKV GitHub repository and the Helper Library - Review the EdgeKV API reference - Use the EdgeKV CLI
Read how customers use EdgeKV with EdgeWorkers in the Nissen customer story.
Both EdgeKV and Cloudflare Workers KV are globally distributed, edge‑native key‑value stores designed for fast reads and infrequent writes with eventual consistency. Key ways EdgeKV stands out for EdgeWorkers developers: - Tight integration with EdgeWorkers. EdgeKV provides a helper library, CLI, and administrative APIs designed to fit into the EdgeWorkers toolchain and deployment model. - Edge‑cache behavior with automatic invalidation. EdgeKV keeps local copies on executing edge servers and uses a network‑wide invalidation mechanism on writes and deletes, targeting ≤10s eventual consistency for at least 80% of operations. - Clear data organization and governance. Namespaces, groups, and items provide simple structure and lifecycle controls, with RBAC aligned to Akamai Control Center.
If you need strong, cross‑region coordination for hot writes, both ecosystems typically direct you to alternate patterns or services; for read‑heavy workloads, either approach works, with EdgeKV offering direct, opinionated fit for EdgeWorkers deployments.
EdgeKV: - Replicates data globally and keeps popular keys local to the edge servers running your code. - Uses a publisher/subscriber‑like invalidation model so that updates purge stale local copies and converge to the latest value with eventual consistency (≤10s for at least 80% of operations). - Optimizes the data path by leveraging Akamai’s delivery network between persistent storage and the edge.
Fastly’s edge compute ecosystem offers edge data primitives as well. At a high level, approaches across edge networks vary in how they cache or replicate writes, how quickly updates propagate, and the strength of consistency guarantees. EdgeKV’s distinct focus is combining globally replicated persistence with local edge copies and automatic invalidation tuned for read‑heavy, ultra‑low‑latency access from EdgeWorkers.
Use EdgeKV when: - Your app reads data frequently at the edge and can tolerate eventual consistency for writes. - You want JSON/string storage without schemas or complex queries. - You need a fully managed, globally available data layer tightly coupled to EdgeWorkers.
Consider other patterns if: - You require strong consistency across regions for concurrent, high‑write workloads. - You need relational schemas, joins, or partial updates to large values.
To understand key‑value databases and their trade‑offs, see Akamai’s overview, What’s the value of a key‑value store?
Prefer to try before you commit? Start a 30‑day evaluation from the EdgeKV page and follow the guided setup after your account is validated.