September 23, 2016

How an Integrated Microwave-IP/MPLS Solution Reduces Latency

The point of this post is to determine the amount of latency reduction possible with a one-box integrated microwave router solution when compared to a two-box (separate router + microwave) offering. By how much does the one box solution improve latency?

Latency is important to all network operators. The lower the end-to-end delay the better it is for all types of applications.

For example, latency is critically important to mobile network operators (MNOs) for LTE Advanced features like coordinated multi-point (COMP) and MIMO, which require extremely tight latency. CRAN architectures are also demanding tighter latency from the backhaul.

In addition, latency-sensitive applications like Teleprotection, SCADA, and simulcast in private markets such as public safety, utilities, and the federal government will greatly benefit from low latency network performance. For other customers, low latency is critical for synchronization and HD video transport.

And latency also impacts connectivity requirements for the emerging Internet of Things (IoT).

Fortunately, the recently introduced CTR microwave router, which integrates IP/MPLS packet handling and microwave radio functionality into one device can help all operators. By having these capabilities together in one box, the CTR microwave router can decrease latency by up to 43 percent compared to a separate IP router and microwave radio configuration. With this improved response time, carriers can deliver a superior user experience (UX).

According to the experts, users expect a latency delay of no more than 100 milliseconds on 4G devices. However, as wireless networks move toward 5G, users’ expectations for application response time will drop drastically on the order of 10x to less than 10 milliseconds. Let’s look at this a bit more in-depth.

Latency in Packet Networks

Firstly, latency is the sum of many factors such as:

• Propagation delay
• Serialization delay
• Routing and switching
• Queuing and buffering

For every packet microwave backhaul connection, an IP router processes the IP packets and forwards them. At each step, the router adds latency. And this is in addition to latency from switches and microwave radio units that packets accumulate. Hybrid networks that contain discrete IP routers and microwave radios necessarily have to endure this latency. In-the-lab results can tell us just how much latency is involved.

Hybrid Microwave Backhaul Latency

Secondly, to explain the impact latency has on hybrid microwave network backhaul, let’s consider a configuration based on an MPLS L3VPN. Aviat Networks conducted measurements in its laboratory of just such an MPLS L3VPN microwave backhaul configuration.

In this hybrid scenario, the MPLS L3VPN is created between the edge routers (Figure 1). The microwave links provide a Layer 2 segment between adjacent routers and do not participate in IP routing or MPLS label switching.

Figure 1: L3VPN latency test topology for a hybrid network (click to enlarge)

Overall, the latency testing for the hybrid microwave network in Figure 1 produced the following results when measuring different size IP packets:

• 64-byte packets: 560 microseconds average latency
• 500-byte packets: 665 microseconds average latency
• 1500-byte packets: 903 microseconds average latency

Converged Microwave Backhaul Latency

Furthermore, in a converged microwave backhaul scenario the third-party routers are removed and an MPLS L3VPN is created between edge CTR microwave routers. In this configuration, the microwave routers perform IP routing and MPLS label switching (Figure 2).

Figure 2: L3VPN latency test topology for a converged network (click to enlarge)

When conducting latency testing on the converged microwave network featured in Figure 2 low latency numbers were achieved in contrast to the hybrid network:

• 64-byte packets: 320 microseconds average latency
• 500-byte packets: 418 microseconds average latency
• 1500-byte packets: 623 microseconds average latency

Bottom Line Results

In summary, comparing the Aviat laboratory results for IP/MPLS functionalities on the hybrid network (Figure 1) and converged network (Figure 2) gives us the following bottom line latency improvements:

• For 64-byte packets, latency is decreased by 42.8 percent
• For 500-byte packets, latency is decreased by 37.1 percent
• For 1500-byte packets, latency is decreased by 31 percent

Conclusively, you can see that evolving microwave backhaul toward a converged model will derive significant positive impacts on latency. And with 4G LTE and the oncoming 5G requirements for short app response times, MNOs and other network operators will be able to deliver on the heightened UX expectations.

If you’d like to talk to someone about reducing your latency by up to 43 percent, please click here.