HTTP - Connection Management

---

---

HTTP connection management is a set of techniques used to manage the connections between a client and a server. These techniques are used to improve the performance and efficiency of HTTP communication.

Connection Management Techniques

Some of the key techniques used in HTTP connection management are listed below:

• Persistence Connection
• Connection Pipelining
• Multiplexing

Persistence Connection

Connection persistence addresses the limitation of previous version where a single request and response can be sent in each connection. It took time to establish a new connection. Connection persistence model refers to, when a single TCP connection is used for sending multiple HTTP requests and responses without closing the connection after each exchange. This helps in reducing the overhead of establishing a new connection each time for every request.

Features of Persistence Connection

• Keep-Alive Header: To achieve persistence connection and keep connection open, HTTP/1.0 uses Keep-Alive header.
• Persistent connections are set by default in HTTP/1.1. Connection is only closed if the client or server specifies Connection: close header.
• Persistent connections reduces latency and improves resource utilization for both client and server sides by avoiding the setting TCP connections for each request and response.

Limitations of Connection Persistence

• Persistent connection consumes server resources such as memory and file descriptors when maintaining many idle connections.
• Under heavy load, DoS attacks can be conducted.
• Proper time configuration is required, as poor time management can lead connection to close early or can be left idle for too long, these can cause inefficiencies.

Example

Request Message:

GET /index.html HTTP/1.1
Host: www.example.com
Connection: keep-alive

Response Message:

HTTP/1.1 200 OK
Content-Type: text/html
Connection: keep-alive

Connection Pipelining

Connection pipelining technique sends several requests one after other over a single connection, without waiting for response, which reduces the latency in the network. It was introduced in HTTP/1.1 to improve performance by reducing idle time between successive requests.

Features of Connection Pipelining

• It allows sending of multiple request without waiting for their corresponding responses.
• Multiple requests are sent by client sequentially and are processes in the same order of the request sent.
• It reduces the latency as multiple requests can be sent without waiting for responses.

Limitations of Connection Pipelining

• Head-of-Line Blocking: The major drawback of connection pipelining is head-of-line(HOL) blocking. In this, if one request in the pipeline is delayed then it delays all other responses in queue.
• Limited Support: Due HOL blocking, connection pipelining is not widely adopted.
• Inefficient Error Recovery: After the delay or failure of any request, it blocks other requests and error recovery process is difficult in pipelining.

Example

Request Message:

GET /index.html HTTP/1.1
Host: www.example.com

GET /about.html HTTP/1.1
Host: www.example.com

Response Message:

HTTP/1.1 200 OK
Content-Type: text/html



HTTP/1.1 200 OK
Content-Type: text/html

Multiplexing

Multiplexing is a feature introduced in HTTP/2, allowing multiple parallel requests and responses to be sent simultaneously over a single connection. In multiplexing, responses are independent and can be sent in any order.

In HTTP/2, streams get divided into frames, these frames can be then sent in any order and then reassembled at the receiving side. So, by using this mechanism, delay in one stream does not affect others. This solves the problem of Head-of-Line(HOL) problem of connection pipelining.

Features of Multiplexing

• Priority: HTTP/2 allows prioritization of streams, so that critical resources can be delivered faster.
• Stream-Based Communication: It allows the server to send responses in any order, based on the dependency of the streams. Every request and response pair is assigned with a unique stream ID, allowing independent handling of multiple streams.
• Header Compression: The header compression 'HPACK' compress the headers, which helps in reducing size and improving performance.

Limitations of Multiplexing

• Complexity: Multiplexing is complex to implement such as for managing prioritization and stream dependencies.
• Resource Utilization: For handling multiple streams, more resources are required on the server side.
• Compatibility: For multiplexing, there should be support of HTTP/2 support on both sides. Not all servers and clients support HTTP/2, which limits the adoption of multiplexing.

Differences Between Connection Management Techniques