Question

In HTTP version \(1.0\), a server marked the end of a transfer by closing the connection. Explain why, in terms of the TCP layer, this was a problem for servers. Find out how HTTP version \(1.1\) avoids this. How might a general- purpose request/reply protocol address this?

Short answer

HTTP 1.0's frequent closing of TCP connections caused inefficiency. HTTP 1.1's persistent connections solved this. General-purpose protocols can implement similar persistent connections.

Step-by-step solution

Understand the problem with HTTP 1.0

In HTTP 1.0, each request/response pair required a separate TCP connection. The server indicated the end of data transfer by closing the TCP connection. This approach had issues like inefficient use of resources and increased latency due to frequent creating and closing of connections.

TCP Layer Impact

TCP connections are not trivial to establish due to the three-way handshake process. Frequent opening and closing of TCP connections increase network overhead and reduce performance due to the additional computational and time costs involved.

Improvement in HTTP 1.1

HTTP 1.1 introduced persistent connections (also known as keep-alive), where multiple requests and responses could be sent over a single TCP connection without closing it. This reduces the number of handshakes and the overhead associated with establishing and terminating connections.

General-purpose request/reply protocol

A general-purpose request/reply protocol can address this by implementing a persistent connection mechanism similar to HTTP 1.1, where the connection is kept alive and reused for multiple request/reply cycles. It ensures efficient resource utilization and reduces latency.

Key concepts

HTTP 1.0

HTTP 1.0 is an early version of the Hypertext Transfer Protocol. It served as the foundation for web communication and was designed to deliver web pages and other resources over the internet. However, it had some performance drawbacks. Each request/response pair required opening a new TCP connection.
This meant that for every action—like loading a webpage or an image—the server had to repeatedly establish and tear down a connection.
Consequently, this led to high latency and inefficient resource usage.

HTTP 1.1

HTTP 1.1 brought several improvements over HTTP 1.0. The most significant change was the introduction of persistent connections, also known as keep-alive.
With persistent connections, multiple requests and responses could be sent over the same TCP connection without the need to close it after every transfer.
This reduced the time and computational resources needed to establish new connections, vastly improving efficiency and reducing latency. Additionally, HTTP 1.1 supports chunked transfer encoding, allowing data to be sent in pieces, which can be helpful for dynamic content generation.

TCP Handshake

The TCP handshake is a three-step process used to establish a connection between a client and a server.
It is called the three-way handshake because it involves three steps:

• SYN: The client sends a SYN (synchronize) packet to the server to initiate a connection.
• SYN-ACK: The server responds with a SYN-ACK (synchronize-acknowledge) packet.
• ACK: The client sends an ACK (acknowledge) packet back to the server, establishing the connection.

This process consumes time and computational resources, making frequent connections expensive in terms of performance.

Persistent Connections

Persistent connections, introduced in HTTP 1.1, allow a single TCP connection to be used for multiple HTTP requests and responses. This approach significantly reduces the need to constantly set up and tear down connections.
The key benefits are:

• Reduced Latency: Fewer handshakes mean quicker subsequent requests.
• Better Resource Utilization: Chances of server overload due to too many connections are reduced.
• Less Network Congestion: Reduced number of packets in the network improves overall stability.

Persistent connections make HTTP communication more efficient and scalable.

Network Overhead

Network overhead refers to the additional resources required to manage data transmission across a network. In the context of HTTP and TCP, overhead includes:

• Time spent establishing and terminating connections
• Computational resources for managing multiple connections
• Bandwidth used for the actual data packets and the packets needed to set up these connections

By reducing the number of connections through mechanisms like persistent connections in HTTP 1.1, network overhead can be significantly minimized. As a result, data transfer becomes more efficient, and network performance improves.