RPC Reinvented: Asynchronous, Durable, and Agent-Ready

What if your remote calls could run for weeks and survive server crashes? This new standard combines durable execution with RPC to make your applications truly fault-tolerant.

#1about 1 minute

Understanding the reliability challenges of traditional RPC

Traditional Remote Procedure Calls (RPC) are not inherently reliable for long-running operations because they don't handle process crashes or network failures.

#2about 2 minutes

The trade-offs of using event-driven systems for reliability

While event-driven systems offer reliability through queues, they sacrifice the simple API contract of RPC and lead to complex, callback-heavy code.

#3about 3 minutes

Building a fault-tolerant RPC with state persistence

A reliable RPC can be achieved by persisting request and response state in a database, allowing different process instances to handle recovery after a crash.

#4about 3 minutes

Combining durable execution with RPC for a better developer experience

Nexus RPC uses the concept of durable execution to preserve process state, enabling developers to write simple, synchronous-style code for long-running, fault-tolerant operations.

#5about 2 minutes

Applying Nexus RPC to build reliable AI agent tools

Nexus RPC is ideal for AI agents as it allows them to reliably invoke external tools that may be long-running, overcoming the limitations of traditional short-lived RPCs.

#6about 5 minutes

Demo of a fault-tolerant AI agent using Temporal

A live demonstration shows how an AI agent running in a Temporal workflow can invoke a long-running tool and survive process crashes without losing state.

#7about 2 minutes

Summary of Nexus RPC and durable execution benefits

Nexus RPC, enabled by durable execution platforms like Temporal, provides a reliable and simple RPC paradigm for long-running operations, avoiding the complexity of event-driven systems.