How To Set Up Optimal Chat for Your Real-Time Application

PublishedNov 17, 2021BySubspace Team
Chat capabilities add an important element of interactivity to real-time apps, allowing users to instantly connect with customer service agents or other users or players, depending on the nature of the application. The best experiences are fast and seamless.
While users demand real-time chat experiences that truly live up to the “real-time” promise, delivering optimal levels of speed and responsiveness may often seem like a complicated feat. But there are some key ways to optimize real-time chat performance.
In this introductory guide, we’ll look at some of the major factors that affect chat speed and responsiveness and explore optimizations that can help deliver the best possible real-time user experiences.
While maximum message delivery time for real-time chat applications is 250 ms, speed is ultimately king. The more effectively developers can optimize speed and reduce latency, the faster—and better—the user experience.
Where chat is concerned, one common speed bump is packet loss, because if a packet is lost, delayed, or damaged, the protocol will resubmit it. This obviously slows down messages.
Just because packet loss is a frequent culprit doesn’t mean you have to live with it. But you will need a strategy to reduce packet transmission problems. Here are some factors to take into consideration for real-time chat applications.

Optimizing the Data Delivery Method

As Steve Jobs once said,
“Details matter. It’s worth waiting to get it right.”
This detail-oriented—but sometimes slower—approach is a good way to describe Transmission Control Protocol (TCP), the first and most stable of the two choices. TCP places priority on delivering every packet, every time. This method, depending on the distance required, can slow down some real-time chat communications, but is very good at delivering volumetric downloads, web browsing, and the like.
UDP, or User Datagram Protocol, offers the flip side of TCP: packets are received and transmitted as they are. With no packet recovery or error correction, UDP is faster, but packet loss may occur, with rates as high as 50% across the public internet. UDP is typically used in gaming, voice and video chat, and live streaming.
TCP and UDP each have their pros, cons, and ideal use cases as far as speed and data integrity are concerned. Almost all ISPs support both protocols, so the choice revolves around the use case. Do you need stability, or faster performance? To learn more about these choices, see our Introductory Guide to TCP/UDP.

Optimizing Publication/Subscription Patterns and Messaging Logs

Another thing that can speed up the chat experience—or drag it down—is the method used for pushing out message data. Pub/sub is one popular messaging pattern that decouples the sender and receiver of messages so they can be sent asynchronously.
In pub/sub, messages are transmitted immediately rather than stored. Accordingly, there are no guarantees about the order in which messages will be received. Apache Kafka, which acts as an event broker between the message publisher and receiver, offers a pull-based approach, in which new messages are appended to a log. This log essentially functions as a queue, in which the receiver tracks its own position and pulls new data whenever it is ready. This can reduce throughput, especially when exacerbated by network congestion.
Optimizing these patterns often depends on the nature of the messages, including the number and volume of topics. Distributed log systems like Kafka may make more sense when there are fewer topics but messages are higher volume.

Optimizing UI Rendering

When a user interacts with an app, they trigger an exchange of data that ultimately results in new information appearing on their screen. But there are two different ways of triggering the new recording, although the end result may appear to be the same. These are called Optimistic and Pessimistic UI rendering.
Just as optimists assume by default that the best-case scenario will occur, so too does Optimistic UI rendering. Optimistic UI rendering works under the assumption a certain server call will follow a specific user action.
In action, this means page changes are rendered immediately based on the anticipated result, even before the app or webpage communicates with the server. The user is only notified in the event that the exchange eventually fails. Taking this “expect the best” approach speeds chat responsiveness because there is no waiting for machines to communicate with each other. After a user performs an action, the page responds instantly.
Pessimistic UI rendering takes the reverse approach. The page waits to identify the actual response before rendering any changes. This improves certainty, but it also has a detrimental effect on real-time performance because the exchange must occur before the response is rendered.
Sure, errors are reduced, but the negative impact of slow load times can be significant. Site abandonment, the resulting effects on SEO—it’s not pretty. That’s why in most cases, Optimistic UI rendering is a good approach for making pages more responsive.
Where does Optimistic UI rendering make the most sense? When it comes to real-time chat, some key opportunities include:
  • Sending messages in the chat window
  • Displaying message reactions (likes, upvotes, etc.)
  • Creating new message windows
  • Deleting, closing, or archiving messages or conversations
It’s important to note, however, that optimism doesn’t mean blind trust. You can feel more confident in taking the Optimistic UI rendering approach—and ultimately deliver the best experience for your users—if you know you have the network reliability to back it up.

Collecting and Understanding Performance Data

Even with optimal setup, sometimes apps don’t perform as well as you’d like them to. Capturing and reading logs may be painstaking work, but it’s an important component to understanding route problems, bottlenecks, and congested routes between locations.
The first step is knowing what data will be helpful for troubleshooting and optimization purposes. Performance tracking should always include:
  • Detailed client-side logs and analytics for low-level protocol retransmission
  • Connection metrics and TCP-level RTT (round trip time) to provide the clearest diagnostics.
  • Alerts when RTT exceeds normal thresholds
Nice-to-have features of performance tracking include:
  • Network-level metadata
  • Geographic region
  • ISP details
  • Last-mile delivery details
With this data captured, you can essentially pop up the hood and take a closer look at what’s going on.


When it comes to real-time chat, there’s no room for latency and lags. Users expect instant connection, and every extra fraction of a second can erode the chat experience—leading to frustration, app and website abandonment, and even loss of customers.
That’s why chat optimization is so important. With just a few changes, it’s possible to reduce or eliminate latency and response time issues and deliver an optimal chat experience.
The right technology partners can also make all the difference, which is why it’s important to consider the network you’re building on. Subspace is purpose-built for delivering the best real-time experiences—we’re a key partner in providing optimal chat capabilities that live up to consumer demands.
For example, Subspace PacketAccelerator helps to enable a better and more responsive chat experience by intelligently selecting the fastest, most dependable path for packets. This accelerates packet delivery and reduces loss and latency, improving overall chat performance and outperforming real-time chat application standards. By how much exactly? In this real-life gaming partnership, PacketAccelerator helped to speed up ping times by as much as 80%, a dramatic improvement in the overall online gaming experience.
At the end of the day, delivering a fast, responsive chat experience requires a multipronged approach. While there are factors developers can account for, your network plays a big role in performance too. That’s why there’s no substitute for delivering chat capabilities on a network designed for interactivity.
Sign up and see how Subspace can help you optimize and accelerate your real-time communication.

Share this post

Subscribe to our newsletter

The world’s fastest internet for real-time applications—period. Every millisecond counts. Learn more in our newsletter.

Related Articles