Scaling microwave capacity to new levels
Today’s modern microwave systems use techniques such as advanced packet compression and higher-order modulation to increase channel capacity. However, at a certain point, the only way to get more microwave link bandwidth is to increase the number of radio channels used.
Modern methods that combine 2 or more microwave channels to create a higher capacity virtual link have several names, including channel bonding, radio link aggregation (LAG) and multichannel. Although all methods use multiple channels to scale microwave capacity, implementations and efficiency levels can differ. In this article, a multichannel radio LAG packet microwave system is one that merges radio LAG concepts with new techniques that are optimized for microwave. These systems offer a new way to bond microwave channels together to create higher capacity and more reliable microwave links.
Multichannel radio LAG is designed for microwave
Multichannel systems are designed to work with features such as:
- High-order adaptive modulation (AM)
- Packet compression
- Carrying legacy time division multiplexing (TDM) services as packet traffic together with IP traffic
Multichannel systems give network operators new flexibility when it comes to designing microwave links and new ways to increase microwave capacity and availability:
- A multichannel approach creates a virtual link from 2 or more underlying channels. The resulting capacity is the sum of every channel’s capacity.
- The individual channels in the multichannel bundle can have different profiles for frequency bands, modulation levels and capacities.
- Adaptive modulation can be enabled across all channels in the multichannel bundle. This creates room to increase capacity and service availability according to network design parameters.
- Because modern packet microwave systems packetize legacy TDM traffic, legacy and new IP traffic can use a multichannel virtual link as a whole.
- The rigid association between the capacity a service requires and the capacity a radio channel offers is removed. For example, a packetized synchronous digital hierarchy STM-1/OC-3 circuit can be spread across the channels in a multichannel bundle whose total capacity matches the capacity required.
- Microwave link protection can move from a traditional N+1 spare channel approach to a more effective multichannel N+0 approach. An N+0 approach uses the entire virtual link capacity to increase availability.
In contrast, standard LAG techniques suffer from limitations when used in microwave environments:
- There is a rigid association between a flow and a specific channel in a virtual link. This is because standard LAG hashing algorithms use IP or Ethernet header fields to consistently map a flow to a channel. If these fields do not vary much in value, some channels in a virtual link can become congested while others are only lightly used. Low utilization is a particular challenge when packets are encapsulated in IPsec. In these cases, there is not enough variety in the fields used for hashing algorithms to optimally spread the load across a bundle of channels. As a result, the same channels in the bundle are always selected, leaving the other channels underutilized.
- Every channel in a virtual link must support the same capacity. In microwave networks, this is seldom possible due to the effects of adaptive modulation on individual channels.
These limitations mean that channel capacity must be equal to, or greater than, the highest flow bandwidth. This constrains link dimensioning because radio capacities are typically not correlated with IP service flow capacities. As a result, in some network environments — LTE backhaul networks, for example — channel bundles are underused. Multichannel radio LAG eliminates these issues because it:
- Distributes traffic load evenly based on algorithms that do not leave channels underutilized or impact services, even in the event of a channel failure
- Does not require each channel in a bundle to have the same capacity as the most demanding service
The multichannel radio LAG engine
The multichannel engine is the key component in a multichannel system. It distributes packets over the channel bundle in an optimal way, while preserving the correct packet sequence for each flow (Figure 1). Each flow is distributed according to current channel capacity levels. However, each channel can have a different profile and bandwidth capacity. For example, a 2-channel bundle can be comprised of a 14-MHz channel and a 28-MHz channel.
In Figure 1, N is equal to 4 channels. Throughput could reach 3 to 5 Gb/s depending on factors such as packet compression efficiency, modulation format and link dimensioning. Multichannel links with an N value between 8 and 10 channels can be used to scale capacity to the 10 Gb/s range.
The multichannel engine is aware of traffic flow quality of service (QoS) requirements to ensure that service level agreements (SLAs) are maintained. When multichannel link capacity varies, the multichannel engine uses the real-time status of the entire virtual link to adjust traffic distribution across the channel bundle and improve spectral efficiency.
Spare capacity, not spare protection
Unlike traditional N+1 techniques to scale microwave link capacity, multichannel systems do not require spare protection channels to protect link capacity. Instead, multichannel systems use the concept of spare capacity across a bundle of active channels.
When adaptive modulation is used, a channel does not have to be in an ‘on’ or an ’off’ state; it can be in a partially working state, although at a reduced capacity. In the rare case where the capacity available in the multichannel bundle is lower than requested, high-priority committed traffic is preserved and only best-effort traffic is discarded.
From a network design standpoint, the probability of delivering the committed traffic is very high. That’s because the degradation on one channel can be compensated for with the excess capacity available on other channels in the bundle. Traditional N+1 link protection mechanisms do not support the spare capacity concept when scaling and protecting microwave links. If channel capacity drops, all traffic is moved to a dedicated protection channel, stranding any remaining capacity on the degraded channel.
Increase microwave capacity and availability
There are 2 ways to take advantage of the benefits that multichannel provides:
- Increase availability and maintain the same capacity as a traditional microwave system
- Increase capacity and maintain the same availability as a traditional microwave system
Figure 2 shows what is possible when the goal is to increase availability while maintaining capacity levels. It compares the availability of a traditional 3+1 system to a 4+0 multichannel system.
The behavior of a multichannel system is represented by a curve that is associated with the entire set of channel capacity and availability levels. It is not represented by a pre-defined single point as is the case with a traditional N+1 system.Availability in a multichannel system is increased from about 10-5 to about 3x10-7 when compared to the capacity of an N+1 system. This is equivalent to 99.9999% availability.
If the goal is to maintain the same availability value, the multichannel system offers 25% more capacity than a 3+1 system, with the same number of channels used. This extra capacity can potentially lead to link redesign to reduce antenna size requirements, which can further minimize network total cost of ownership (TCO).
The benefits of multichannel are also realized when traditional TDM applications are packetized using circuit emulation.
Multichannel systems enable better designs
In multichannel systems, the capacity needed between any two points in a network becomes the main design factor, enabling better network designs. Figure 3 compares the committed capacity (CC) and best-effort capacity (BEC) in a traditional 3+1 link design to an incremental multichannel design:
- Committed capacity must always be provided, for example to carry circuit-emulated traffic or premium data traffic.
- Best-effort traffic can be interrupted, for example when adaptive modulation levels are decreased to support communication during bad weather.
Traditional link design A traditional 3+1 link design provides room for committed capacity only. Typically, a 3+1 system uses 4 28-MHz radio channels (3 active channels and 1 protection channel) with fixed 128 quadrature amplitude modulation (QAM) to carry 3 STM-1/OC-3 circuits. There is no room for occasional, best-effort traffic.
Step A: Multichannel 4+0 link design In step A, an existing 3+1 arrangement is upgraded to a multichannel 4+0 configuration. Actively using the fourth channel increases the microwave link bandwidth by 25% when modulation levels are not affected by adverse weather.
Step B: Add adaptive modulation Step B is a multichannel 4+0 design that uses adaptive modulation with 1024 QAM. The higher modulation rate increases the microwave channel and link capacity by 30%. Because adaptive modulation is active on all 4 channels, both committed and best-effort capacity are increased. Step C: Add flexible spectrum use Step C adds more flexible use of spectrum to step B. Instead of using 4 28-MHz channels, the following channel widths are used:
- 1 56-MHz
- 1 28-MHz
- 2 14-MHz
This type of deployment shows how multichannel systems can take advantage of available spectrum and potentially decrease right-of-use costs for channels.
There are also cases where moving from an N+1 to an N+0 multichannel design reduces the number of channels required. In an N+0 design, spare capacity is better used so the spare channel associated with a N+1 design is not needed. Multichannel configuration parameters can also be tuned to improve the link budget, which in turn improves network availability. With the introduction of multichannel packet microwave systems the progressive decommissioning of N+1 installations is possible.
Move ahead with a multichannel microwave system
Multichannel allows packet microwave systems to address the microwave capacity and availability demands of modern IP networks. Compared to traditional N+1 and N:1 mechanisms, multichannel systems offer a more flexible, efficient and reliable approach to scaling microwave capacity. With multichannel, microwave link designers can focus on the actual capacity required, with less emphasis on the availability of frequency bands and channel spacing. This gives link designers more flexibility. Most importantly, it helps network operators make better use of scarce radio spectrum and decrease network operational expenses.
Editor’s Note: The author would like to thank Scott Larrigan for his contribution to this article.
To contact the author or request additional information, please send an e-mail to firstname.lastname@example.org.