Do You Care About Packet Loss or Delay?

TCP (Transmission Control Protocol) and UDP (User Datagram Protocol) are transport network protocols that enable communication between computers over the internet.

In general, TCP is used for activities that require accurate data transmission, such as sending emails and web browsing. On the other hand, UDP is primarily used for fast and efficient transmissions related to things like online gaming, live streaming, and VoIP calls.

5 Fundamental Differences Between TCP vs UDP

Since it enables accurate and reliable data transmissions between two devices, TCP is essential to how computers communicate over the internet.

On the transport layer, TCP’s job is to receive data from an application, break it into segments, and send it to the internet or network layer. The segments are then collected by IP, given a unique address, and routed through the best possible path to its destination.

Here are five keys to TCP:

TCP is a connection-based protocol, so it must establish and maintain a connection between two devices (one-to-one communication only) before it starts sending data between them. It does this through the three-way handshake (SYN, SYN-ACK, ACK).

TCP guarantees packet delivery, which ensures no data is lost. The receiver sends an ACK message for every packet that arrives successfully, and the sender can retransmit lost packets when it doesn’t receive an ACK.

TCP ensures in-order packet delivery by adding sequence numbers to the data packets. This ensures that they are reassembled in the correct order at the destination.

TCP has flow and congestion control, for example, by setting a timer for every packet sent and speeding up the transmission rate if an ACK is received before the timer runs out. TCP will also slow the transmission rate if the receiver can’t handle the incoming packets.

TCP uses a mandatory checksum for error handling. The checksum is a unique value generated by the sender based on the information in a data packet. If it doesn’t match the receiver’s checksum, the packet is dropped, and a retransmission is requested.

UDP is also a transport layer network protocol that’s built for speed, but it isn’t as reliable as TCP. Here are five keys to know about UDP:

UDP is a connectionless protocol, which means it doesn’t need to establish a connection between a sender and receiver before transmitting data. Unlike TCP, the lack of a connection state allows UDP to support one-to-one as well as one-to-many communication methods.

UDP doesn’t guarantee packet delivery simply because it doesn’t confirm if data packets arrive at the destination or not. It also has no mechanism in place to retransmit lost packets.

UDP has no sequencing, which means the receiver can’t reassemble them if they arrive out of order—and there can also be duplicate packets.

UDP has no congestion control, so UDP transmits data as fast as it can regardless of the state of the network. The downside here is that if the receiver is overwhelmed or reaches its maximum processing speed, some packets will be discarded, and there’s no way to recover them.

UDP has an optional checksum. UDP cannot correct complex errors, and it simply discards any corrupt packets it receives without requesting a retransmission.

When To Use TCP vs UDP

TCP features allow it to function as a reliable way of transmitting data. For instance, if a packet is lost, it stops the transmission to resend that packet. Meanwhile, when the receiver gets a corrupt packet, it requests a replacement. Lastly, if the network is congested, TCP can adjust the transmission rate so the receiver can properly process the incoming packets.

This insistence on 100% packet delivery means TCP transmissions are prone to delays. UDP, on the other hand, doesn’t have as many features as TCP—so that makes it lightweight and better suited for scenarios that require low latency.

There are many ways to describe how UDP works (such as “best-effort,” “fire and forget,” and “unreliable”), but its fundamental goal is to get data packets from the sender to the receiver as fast as possible. As a consequence of this, UDP neither confirms the delivery of packets nor reassembles those that arrive out of order. In fact, even if the receiver is struggling to process incoming packets, UDP just keeps on blasting them out.

The decision to use TCP or UDP often comes down to how much you can tolerate packet losses and delays. In other words, it’s a question of whether you prefer accurate or timely transmissions. Each can be great in some scenarios but a hindrance in others.

When to use TCP

TCP is the preferred protocol when accuracy is the top priority over speed. This is why it’s used for activities like the following:

• Web browsing
• Sending emails
• File transfers
• Text messaging

TCP works with other network protocols like HTTP, SMTP, and FTP to make these possible.

When to use UDP

UDP is best for activities that require a steady flow of data, low latency, and can tolerate some degree of packet loss. These include:

• Online gaming
• VoIP services
• Live video/audio streaming
• DNS lookup
• Multicasting

Since TCP intends to deliver every piece of information perfectly, it will stop or pause the transmission to resend every lost or corrupt data packet. UDP won’t.

Similarly, TCP will only resume its transmission when the sender receives a confirmation of delivery. UDP won’t. This is why using TCP for live or real-time communication will often cause unnatural delays and breaks in the transmission, whereas UDP won’t.

On the other hand, using UDP for things like emails and web browsing is less than ideal because there’s a high chance the receiver will get incomplete or scrambled information. That’s why it’s more useful for things like voice calls—since they can handle a bit of jitter. Likewise, if a few frames go missing during a live stream, it won’t really affect the overall quality of the transmission.

It all comes down to what your needs and priorities are when deciding what network protocol to use.

Is TCP More Secure Than UDP?

TCP is often considered to be more secure than UDP but this isn’t entirely true. TCP and UDP, by themselves, don’t offer any form of security to the packets. Both can transmit plain or encrypted packets.

TCP does have superior error-checking capabilities, however, and it’s also less prone to malicious attacks because it uses sequence numbers and acknowledges packet delivery. For instance, if TCP will not acknowledge packets with a spoofed IP address.

Nevertheless, TCP’s three-way handshake is still capable of being exploited in a DOS (Denial of Service) or DDoS (Distributed Denial of Service) attack.

Imagine there’s a hacker who sends a large number of SYN packets to a server. The server responds with a SYN-ACK for every packet it receives and waits for a response that never arrives. This is known as SYN flooding and will slow down the server so much that it won’t be able to process legitimate requests.

Keep in mind that TCP is also susceptible to session hijacking and man-in-the-middle attacks.

As for UDP, it is prone to IP spoofing, where hackers pretend to be legitimate devices. UDP transmissions are thus vulnerable to eavesdropping, interception, and manipulation. Lastly, UDP’s lack of flow and congestion control makes it susceptible to DoS and DDoS attacks as well.

Fortunately, there are many steps companies can take if they want to secure their VoIP communications while using any of these protocols. They include:

• Setting up a VPN: Using a VPN (Virtual Private Network) will add an extra layer of protection to the transmission and make it difficult for hackers to intercept.
• Enabling firewalls: Companies can enable and configure firewalls to detect and filter all incoming packets, thus preventing or limiting unauthorized access.
• TLS/SSL encryption: Transport Layer Security (TLS) and Secure Sockets Layer (SSL) are protocols that encrypt data packets as they are sent out to reduce the chances of eavesdropping and interception.
• SYN backlog recycling: Deleting the oldest half-open TCP connections can free up memory on the receiving device so it can process legitimate requests.

Apart from these steps, companies should run regular data security audits to identify possible weak points—because data breaches and cyberattacks can cost companies thousands to millions of dollars. That said, security audits will also reveal if your system is compliant with industry standards, so they can be worth it for that check alone.

It’s also important to educate every employee on data security, regardless of whether or not they are part of the security/incident response team. Remember that any employee who has access to a device that connects to the internet is a potential security risk that someone or something could exploit.