Question

In a gRPC-based architecture, why does the use of HTTP/2 multiplexing provide a performance advantage over traditional HTTP/1.1 REST interfaces?

Accepted Answer

In HTTP/1.1, the protocol is sequential, meaning a client must send a request and wait for the server to send the complete response before it can send another request over the same connection. This limitation is known as head-of-line blocking, where one slow or large request prevents all subsequent requests from being processed on that connection. To work around this, browsers and clients often open multiple parallel TCP connections to the same server, which consumes significant system resources and slows down connection establishment due to the TCP handshake process. HTTP/2, which gRPC uses as its transport layer, solves this through multiplexing. Multiplexing allows the client and server to send multiple requests and responses simultaneously over a single TCP connection. Each request and response is broken down into small, independent units called frames, which are interleaved, sent, and then reassembled at the destination. Because the connection is never blocked by a single request, the client can push many data packets forward without waiting for previous ones to finish. This eliminates the need for multiple connections, reduces the latency involved in establishing those connections, and maximizes the use of available network bandwidth, resulting in significantly higher throughput and lower overhead for microservices communication.

Home → All Courses → Engineering and Technology Courses → Distributed Systems for Internet-Scale Services → Flashcard

In a gRPC-based architecture, why does the use of HTTP/2 multiplexing provide a performance advantage over traditional HTTP/1.1 REST interfaces?