Connectionless Protocol Definition
A connectionless protocol refers to the communication between two network endpoints without a prior arrangement in which one network endpoint simply sends a message to the other. At the sending end, the device transmits the unit of data before ensuring that the receiving end’s device is ready.
This type of protocol describes most open internet transmissions, although some protocols request a retransmission as needed to allow for error correction. There are a variety of supported connectionless protocols online, including HTTP (hypertext transfer), ICMP, IP, IPX, UDP, and TIPC. A connectionless protocol differs from a connection-oriented protocol, in which both devices must communicate with each other.
What is a Connectionless Protocol?
A connectionless network protocol is an alternative type of data transmission in which a network endpoint sends an IT signal automatically, without determining whether a receiver exists or stands ready to receive it. This stands in contrast to many conventional connection-based data transmission methods, which usually establish device connectivity checks or dedicated handshaking.
A connectionless transport layer protocol is a type of open signal like a radio frequency transmission. It is simply sent or broadcast outward, without as much regard for the recipient. Connection-based protocols have an intended destination and a defined point of origin, more like a cabled connection.
Although connection-oriented protocols function more effectively in certain cases, for data delivery in many systems, a connectionless protocol is adequate. Many connectionless protocols can facilitate multi-casting to a variety of receivers, and these protocols are often favored due to lower overhead costs since there’s no need to implement connectivity protocols such as those relating to handshakes.
Although connectionless transport protocols may drop a minimal number of data packets, in some cases this may not be noticed by receivers. Tools such as elastic load balancers that prevent packet loss can also mitigate against this type of issue.
Connection-Oriented and Connectionless Protocols
Connection-based services or connection-oriented services and connectionless services are the two services that layers provide to layers above them.
Connection-Oriented Services
Users of connection-oriented services follow a sequence of operations:
- Establish connection
- Send information
- Release connection
To provide connection-oriented services, it is necessary to establish a connection first, and then start the communication, send the information, and then release the connection. This type of service is more reliable, because if there is an error on the receiver’s end, the sender can resend the information.
Connection-oriented means that devices perform a handshake when they communicate to set up an end-to-end connection. The handshaking process can vary significantly and may be a simple transport layer protocol TCP synchronization, or a complex negotiation of communications parameters with a modem. Obviously, connection-oriented systems demand on bi-directional communications and cannot function in unidirectional environments.
Connectionless Protocols
In connectionless protocols, users make no effort to establish a dedicated end-to-end connection. Instead, they simply send the information, ideally ensuring there is adequate speed and minimal interference. Just like on a citizen band radio or walkie talkie, if the message is garbled, the only recourse is to ask for a clarification. Advantages of connectionless protocols include simpler form, lower overhead, no need for circuit setup, and ability to multicast.
Connectionless services designate the intended destination of each message or packet of information. The system routes each packet independently, just like postal packages, so there’s no way to be sure you’ll get all of the information in correct order.
There are advantages and disadvantages of connectionless and connection-oriented protocols. The difference between connectionless and connection-oriented communication breaks down as follows:
| Connection-oriented protocols | Connectionless protocols |
| authentication is needed | no authentication necessary |
| checks whether message is received, sends again if error occurs | no guarantees of delivery |
| more reliable | less reliable |
| stream based | message based |
| analogous to telephone | analogous to postal |
| long and steady | bursts of communication |
| congestion is not possible | congestion is possible |
| packets follow the same route | packets do not follow same route |
| high range bandwidth required | low range bandwidth adequate |
Both connectionless communication and connection-oriented communication typically connect two or more than two devices providing services typically offered by the network layer.
Connection vs Connectionless Protocols: Layers and Services
In computer networking and the open system interconnection (OSI) model, the transport layer is the fourth layer, responsible for end-to-end network communication. The transport layer brings logical communication to application processes that run on different hosts.
Those hosts run inside a layered architecture composed of various protocols and other network components. The transport layer and its protocols offer host-to-host communication services for applications, including connection-oriented communication, flow control, multiplexing, and reliability. Protocols are simply sets of rules that control the meaning and format of messages, frames, or packets that the server and client exchange.
The OSI model of general networking and the transport layer of the internet protocol suite, which serves as the internet’s foundation, have distinct implementation details. There are five classes of transport protocols defined for the OSI transport layer. They range from TP0, with the least error recovery, to TP4, which offers more error recovery and is intended for less reliable networks.
The Transmission Control Protocol (TCP) is the best-known transport protocol of the internet protocol suite. User Datagram Protocol (UDP) is a connectionless protocol used for less complex transmissions, TCP is the more complex protocol, used for connection-oriented transmissions based on its stateful design incorporating data stream services and reliable transmission. Together, UDP and TCP comprise nearly all internet traffic.
Services are the operations that any given layer in the OSI Reference model is able to provide to the layer above it. A service defines the operation itself that a layer can perform, but the service does not specify any details of implementation.
In the transport layer, data on the host computers is delivered to the appropriate application processes. The system forms data segments by statistically multiplexing data from different application processes, and adds destination and source port numbers to each transport layer data segment’s header. The port numbers together with the source and destination IP addresses constitute a network socket—an address for identification of the process-to-process communication. The session layer supports this function in the OSI model.
Certain transport layer protocols support virtual circuits—for example TCP does, but UDP does not. This means these protocols provide connection-oriented communication using a packet-oriented network. TCP also supports error recovery for end-to-end reliable communication through error detecting code and automatic repeat request (ARQ) protocol.
The much simpler UDP protocol uses packets called datagrams instead of segments. It does not provide reliable communication or virtual circuits.
These differences mean that TCP may be used for various protocols, including email transfer and HTTP web browsing, while UDP is more likely to be used for functions such as broadcasting and multicasting, where retransmission to many hosts is not possible. Offering shorter latency, higher throughput, lower costs, and other benefits of connectionless communication, UDP is often used when packet loss can occasionally be accepted during real-time multimedia communication—for example, for online computer games, IP-TV, and IP-telephony.
Datagrams are referring to connectionless protocols. A connectionless protocol layer lacks an acknowledgment scheme and sequence number, meaning the protocol is connectionless. However, it can also be called datagram.
Clients and servers that use a perfect channel to communicate share a dedicated point-to-point channel—or an illusory version of such a connection. First, the client and server must establish the connection to communicate and transmit the data, afterwards closing the connection. The channel guarantees that all data transmitted over the channel is sent and received in the same order.
When clients and servers use datagrams to communicate, they send and receive completely independent packets over the network. The packets do not have a dedicated point-to-point channel and travel on whatever route is available.
In contrast, a connection-oriented protocol layer has an expected acknowledgment process and sequence numbers, and if they are not received, retransmission takes place. This is how connection-oriented protocols guarantee delivery of data. They are more reliable protocols.
There are two main protocols at the transport layer of the OSI Reference Model: the connection-oriented TCP, and the connectionless UDP. TCP works based on a set of rules and uses them to negotiate before establishing a logical connection and sending the data. It is therefore used for applications that demand these types of connections and service features.
UDP is preferred for use with applications that don’t have those features or connections, but do need faster performance. By not negotiating and then establishing a connection, the UDP saves time.
Some refer to this as a kind of circuit-switching. However, a TCP connection is still sending data as packets, not actually creating a circuit between devices. There is still potential for mixed pieces of data, data loss, or other problems that come with packet switched communications. In other words, TCP and other connection-oriented protocols do not eliminate the need for circuit-switching technologies.
A packet switching network follows connectionless networking protocols that divide messages into packets before sending them, including Wide Area Network (WAN) protocols and TCP/IP. This is distinct from circuit switching technology.
Does the VMware NSX Advanced Load Balancer Platform Support Connectionless Protocols?
Yes. The VMware NSX Advanced Load Balancer supports both connection-oriented (TCP) protocols, as well as connectionless protocols such as UDP and HTTP.
Before sending any TCP connection to the server, the VMware NSX Advanced Load Balancer rewrites the client IP address, regardless of which TCP profile type a virtual service uses. Along these lines, the platform rewrites the destination address from the virtual service IP address to the server IP address. The Service Engine source IP address is always what the server sees.
Learn more about the VMware NSX Advanced Load Balancer’s support for connectionless protocols here.
For more on the actual implementation of load balancing, security applications and web application firewalls check out our Application Delivery How-To Videos.