About Orderful - Modern EDI for the Supply Chain

Orderful is an EDI (Electronic Data Interchange) service provider for retail, logistics, warehousing, and the broader supply chain. EDI is the set of protocols that moves orders, invoices, load tenders, and inventory files between trading partners. Its earliest forms date to the 1950s and 60s, and it came into its own in the 1970s. Orderful's mission is to embrace EDI for what it does well and bring it into a modern stack: the fastest, most modern EDI platform available.

Use Cases: Durable EDI Transactions and Agents

Orderful runs DBOS in two places:

High-volume transaction processing. Orderful's typical path: a trading partner sends an order to a supplier, an invoice or inventory file to a customer, a load tender to shipping companies, and Orderful processes and delivers it. Orderful processes millions of EDI workflows per day.

The Orderful AI agent harness. Orderful has launched several AI solutions, including an EDI industry first: an agent harness that asks AI to take on sophisticated, multi-step workflows. Agent plans can run for minutes or hours, then wait days for a human-in-the-loop approval. Mapping those planning cycles onto durable workflows turned out to be a natural fit.

Challenge: Async Semantics That Off-the-Shelf Queues Can't Match

As Orderful grew, they faced a number of software engineering challenges scaling their stack, which used RabbitMQ to manage EDI workflow messages and process flow:

Workflow complexity. EDI workflows are inherently asynchronous, and require support for fanning out, partitioning, retries, and prioritization. Orderful found RabbitMQ workflow semantics to be a poor match for EDI workflow complexity. They decided to evaluate workflow orchestration solutions to replace RabbitMQ.

Self-hosting was a must. Being able to self-host the solution was important in order to ensure control over the infrastructure running customers’ mission critical EDI workflows. 

Cost consciousness. Workflow orchestration systems like Temporal are famously expensive to host and operate. Orderful wanted to minimize, or even altogether avoid adding new backend infrastructure or staff to operate it.

Solution: Just Postgres and Containers

Orderful evaluated a slate of durable execution and workflow orchestration platforms, including Temporal, Restate, Little Horse, and others, and ran a weighted analysis across their criteria. DBOS came out head and shoulders above the rest.

Here’s what mattered to Orderful:

An architecture that already existed. Orderful is a TypeScript shop using RDS Postgres as the backing data store. DBOS needs a database and containers to run, and nothing else.

‍"To implement durable queueing and workflows with DBOS, you don't need anything more than the basic primitives of a compute server and a Postgres server. Anyone can keep those things up and running and administered really well. That was incredibly attractive." — Piers MacDonald, CTO Orderful.

Developer accessibility. Orderful is constantly shipping new processing capabilities. Strong TypeScript SDK support, good tooling, and no language bridging kept DBOS accessible to the whole engineering team.

Self hosting. The free, open source DBOS Transact library runs anywhere. DBOS Conductor, the management control plane for DBOS Transact applications can also be self-hosted, giving Orderful the control and insights they needed over customers’ EDI workflows.

Observability through Conductor. Orderful’s motivation for moving the golden path to DBOS was insight: how long these things take to run, whether they need recovery, and what failed and why.

“DBOS Conductor gives us insight into how long workflows take to run, if anything failed and why, if any workflow needs to be recovered,” said Orderful software engineer, Calvin Ng.

Results: 2M Workflows a Day, With 4–5M in Sight

Orderful's mainline processing path now runs on DBOS and handles roughly two million EDI workflows a day using a fair queuing pattern. Delivery is next, and the team is also looking at unifying its ingestion workflows later this year, along with a two-to-three-times expansion in step counts per workflow.

‍"In EDI, you can't have things fall between the cracks. You need the ability to trace it end to end, and we don't think we could do that without DBOS." — Piers MacDonald

The partnership has been hands-on. Orderful and DBOS share a Slack channel, and the team has used it for everything from scaling guidance to feature requests, including a unified workflow view across multiple system databases, now under active discussion as Orderful moves multi-region and multi-cloud.

‍“We always put workflow orchestration off because we'd heard that implementing it with something like Temporal was such a heavy lift. Had we known how easy DBOS makes it, we would have done it sooner. DBOS should really be a language primitive."