Table of Contents
To achieve virtualization in networks without hosts being aware of the underlying physical architecture tunnelling is widely used. Network virtualization overlays provide multi-tenancy services in cloud data centers for existing network equipment. IP tunnelling is used to logically segregate each virtual traffic. This technology allowed creation of virtual networks and essential to support technologies such as VXLAN, NVGRE and STT (Stateless transport tunnelling).
In today’s topic we will learn about Stateless transport tunnelling, its features, packet architecture, and its uses.
Stateless Transport Tunnelling
Stateless transport tunnelling is comparatively new to network virtualization compared to its other counterparts VXLAN and NVGRE and its original sponsor was Nicira. It is meant for an overlay network in a multi-tenant network which is under tenant control. STT achieves better performance using TCP segmentation offload (TSO) functionality. The main purpose of STT is to enable efficient network virtualization by encapsulating and forwarding packets between virtualized environments over a physical network.
Features of Stateless Transport Tunnelling
- SST uses MAC-IN-IP tunnelling
- STT context IDs are 64 bit in size to allow much larger number of virtual networks and more broad range of service models
- STT achieves performance gain by leveraging TCP segmentation offload (TSO) technique of NIC cards. Using this technique it reduces overhead of sending multiple small requests
- It is stateless
- STT packets support unicast between endpoints of tunnel and does not require to use TCP windowing scheme having TCP synchronization and flow control requirements
- Ability to be implemented as a software switch with NIC hardware acceleration benefit
- Relieves server CPU of significant load in high bandwidth (10G+) systems
Stateless Transport Tunnelling Packet Structure
Let’s look at the structure of a Stateless Transport Tunnelling packet. The outer header of the packet contains MAC and IP address for sending the unicast packet to the destination switch acts as VTEP. The STT header follows the outer header having STT context header of 64-bits which can be further subdivided and used for other purposes.
The STT software switch inserts header information into TCP lookalike packet with required virtualization features. This allows guest OS to send frames up to 64-K to the hypervisor which are then encapsulated before being sent to NIC to perform segmentation.
Use Cases for STT
- Ideal for software implementation in Hypervisors
- Uses Ethernet over TCP over IP tunnel
- Solves the problem of efficient transport for large 64-KB storage-blocks
Related FAQs
Q.1 How does STT differ from other tunneling protocols like VXLAN and GRE?
- VXLAN uses UDP for transport, while STT leverages TCP.
- GRE is a simpler protocol without features like TCP segmentation offload (TSO) optimization, which STT supports.
- STT is specifically optimized for virtualization workloads.
Q.2 What transport layer protocol does Stateless Transport Tunnelling use?
Stateless Transport Tunnelling uses TCP as the transport protocol.
Q.3 Is Stateless Transport Tunnelling suitable for WAN or internet connections?
Stateless Transport Tunnelling is generally designed for data center networks rather than WAN or internet connections. Using TCP-based encapsulation over WAN can lead to performance issues due to TCP retransmissions.
Q.4 What are the key advantages of using Stateless Transport Tunnelling?
- TCP Segmentation Offload (TSO): It leverages the TSO feature of NICs to improve performance.
- High Throughput: Designed for high-speed data center networks.
- Compatibility: Works well with existing network infrastructure and TCP-based systems.
Q.5 How does Stateless Transport Tunnelling achieve high performance?
Stateless Transport Tunnelling offloads packet segmentation tasks to the NIC hardware, reducing CPU load and increasing throughput.
Q.6 Does Stateless Transport Tunnelling support multicast traffic?
No, Stateless Transport Tunnelling relies on TCP, which does not natively support multicast traffic.
Q.7 What are the potential challenges of using Stateless Transport Tunnelling?
- Dependency on TCP can introduce latency or retransmission overhead in case of packet loss.
- Lack of support for multicast traffic.
- May not perform as well in high-latency or lossy networks.
Q.8 Can Stateless Transport Tunnelling work with existing network equipment?
Yes, Stateless Transport Tunnelling can work with most standard network equipment since it encapsulates traffic in a TCP payload.
Q.9 Where is Stateless Transport Tunnelling most commonly used?
Stateless Transport Tunnelling is commonly used in cloud computing environments, virtualized data centers, and software-defined networking (SDN) architectures.
Q.10 Can Stateless Transport Tunnelling be used for non-virtualized networks?
While technically possible, Stateless Transport Tunnelling is optimized for virtualized environments, making it less ideal for non-virtualized use cases.
Q.11 What troubleshooting steps are recommended if Stateless Transport Tunnelling fails?
- Verify TCP connectivity between endpoints.
- Check for MTU mismatches that might cause fragmentation.
- Inspect firewall rules to ensure TCP traffic is allowed.
- Use packet captures to debug encapsulation issues.
Q.12 Is Stateless Transport Tunnelling supported by major virtualization platforms?
Support for Stateless Transport Tunnelling may depend on specific platform implementations. Platforms like Open vSwitch (OVS) and certain SDN solutions often support STT.
Q.13 How does Stateless Transport Tunnelling handle packet fragmentation?
Stateless Transport Tunnelling relies on the underlying TCP stack to handle packet fragmentation and reassembly.
ABOUT THE AUTHOR
I am here to share my knowledge and experience in the field of networking with the goal being – “The more you share, the more you learn.”
I am a biotechnologist by qualification and a Network Enthusiast by interest. I developed interest in networking being in the company of a passionate Network Professional, my husband.
I am a strong believer of the fact that “learning is a constant process of discovering yourself.”
– Rashmi Bhardwaj (Author/Editor)
