Sometimes you will find yourself tasked with a technical chore that, on the surface, sounds very easy, but turns out to be quite complicated.
Let�s say circumstances warrant a transmitter site�s move to a new location. You find, though, that there are no 950 MHz-band channels available. They�re already all used in your market. To make matters worse, your engineering buddies from across town tell you that the telco facilities that do exist at the new site are unreliable; they fail when it rains or snows (or when the sun is out), and repairs sometimes take days or longer.
�Not acceptable!� you think, envisioning yourself spending Friday night at the site, waiting for a call-back from telco. �There has to be a better way!� And there is.
In this article, we�ll look at other radio link options � such as 11 GHz as prescribed in Part 101; the 952�960 MHz band, as prescribed in Part 101; and, of course, non-licensed, ISM band options as well.
A MORE SOPHISTICATED APPROACH
Many STL systems provide only that, communications in one direction only. Now that IP connectivity is commonplace and expected, it�s time to do the engineering on a fully-duplex system. The 950 MHz legacy systems still have their place � and we use them regularly. By necessity, though, we�ve moved way beyond their native capability. Having high-speed IP connectivity provides many benefits that will be appreciated once it�s up and running.
There�s an important point about IP connectivity that I want to make clear, though, because it represents something of a paradigm shift from what we�re all used to, when it comes to using the legacy gear. The nature of Ethernet, and IP, is that it is easily made to be self-healing. This not only provides an obvious advantage in keeping your station on-the-air, but it also frees you from having to be the one that does the �healing.� Systems can be designed and then configured to do much of your work for you, and then letting you know after the fact, by way of an email (or other remote control functionality).
While what follows may seem like more of an IT topic, it�s important to discuss it (at least conceptually) so that you can really understand what I mean by self-healing.
For the system to work as I�m about to describe, two links need to be established; and, after establishing both links, you will have a choice as to how to configure the self-healing aspect:
� Load-balancing at layer-2. Both ends of both links will terminate in a layer-2 switch. The load-balancing function means that you configure the layer-2 switch to share both links for the passing back and forth of the Ethernet frames. In the event one fails, the layer-2 switch will start passing all of the traffic over the remaining link. You need to be careful not to �over-subscribe� the path, meaning that the normal amount of traffic passing over either link is less than 50 percent of its ultimate capability.
� Using a router to switch routes at layer-3. In this case, both ends of the link will terminate in a router (operating at layer-3). Normally, the router will pass all traffic over one of the two links. You will configure the router to move all the traffic to the backup link in the event that the primary link fails. When the primary link comes back, the traffic will start routing over the primary once again. The primary link would be on one network segment, and the secondary link would be on another.
� Two separate networks, isolated from one another. This is a bit of a turn away from the self-healing aspect; however, since many AoIP codecs have two network ports, and can be configured to switch automatically themselves, you could consider this method. At the far end, some hosts would be on one network, and some on another. If one link were to fail, some hosts (those with only one NIC) would become unavailable to the far end, while others would remain available.
In each of the three cases, your AoIP traffic will continue to flow (with perhaps a very brief gap) after one of the links fails. Using a remote control with SNMP will allow you to learn of the failure; another option would be a use of PRTG, which can be easily configured to let you know of network ports going up and down. So � unlike the old days � instead of being woken up in the middle of the night by a silence detector or panicky jock, now you let your network devices fix the problem for you. During normal business hours, you can see what happened.
Fig. 1: Interconnectivity between the studio and transmitter site, provided by dual IP links, in a self-healing configuration. One link is via radio, and one via wireline. The first of the self-healing system configuration ideas we�ll look at will make use of a licensed IP radio and a telco backup. (See Fig. 1.)
One link providing IP connectivity in this case will be provided by a licensed IP radio operating in the 11 GHz band.
Earlier I alluded to �unreliable� telco connectivity; however, for purposes of this discussion, let�s assume that you can get high-speed connectivity to your new site from your studio facility.
If your telco provider is AT&T, ask them about ASE (ATT Switched Ethernet); if your provider is Verizon, ask them about Ethernet Virtual Private LAN Service; for Frontier, ask about Ethernet Local Area Network; and for CenturyLink, ask for plain old metro Ethernet.
Deciding which of the two links becomes the �primary� and which is the �secondary� may take some doing. In an ideal world, both will be very reliable; but in reality, you will likely find one is better than the other. Clearly, the better one should be your primary.
Fig. 2: Interconnectivity between the studio and transmitter site, provided by dual IP links, in a self-healing configuration. Both links are based on radio. The second of the self-healing configurations is seen in Fig. 2. This is a variation of the first method; the second link is now provided by an unlicensed radio link, operating in one of the Industrial, Scientific, Medicine bands. Why would you want to pay telco for the second link when you use an unlicensed radio? Or, for that matter, why would you even go for a licensed radio?
There are a couple of aspects you need consider. First: Not to cast aspersions on the quality of the ISM radios, but you will find that many of them are very inexpensive and meant to be throw-aways. They are simply not as rugged as some of the radio brands I�ll mention at the end of this article.
That being said, my experience with this type of radio has been very good. We have not had a radio failure, nor an instance of unexplained interference to any of the radio links we use here in Los Angeles. Your mileage may vary, however. Call me conservative, but I would be hesitant to rely 100 percent on this type of radio for mission-critical applications; for a backup, though, I�m comfortable.
Fig. 3: Interconnectivity between the studio and two remote sites. Both sites have redundant paths to and from the studio location. The third self-healing configuration (Fig. 3) is simple variation on the first two; however, in this case, we have three sites, instead of two. An example of this in practice would be one studio location, a main transmitter site (location A) and an auxiliary transmitter site at location B. The studio to A link would be your normal route; in the event that failed, traffic would flow from the studio, to site B, thence to site A. In any case, both sites have a redundant path.
Many manufacturers are now making microwave radios because there is a large market for �backhaul� from cell sites. (Just think about all that data flowing over LTE networks!) That�s an advantage for us because as more brands fight for market share, pricing gets better.
Here are some of the features you�ll need when specifying a radio system:
Layer-2 (Ethernet) Interface. You�re going to treat this radio link as if it were a simple Ethernet cable. This goes for the metro Ethernet connections I discussed earlier as well. In addition to their Ethernet interfaces, some radios still have �auxiliary� TDM interfaces which can carry one or more T1s; so if you need some time to transition between TDM and IP, and you still want to use an older T1 shelf, this is something to consider.
VLAN Priority Support. It�s quite likely that you�ll be using the radio link or the metro Ethernet as a VLAN trunk because it�s a good idea to separate your AoIP traffic from all other traffic to and from the other end. It�s typical that a link such as this will support nearly 100 mbps throughput (or more), so I doubt that anyone downloading a manual (as an example) on the far end is going to interrupt or even slow down AoIP traffic; however, at the very least, consider it to be future-proofing.
The 802.1Q standard defines a system of VLAN tagging for Ethernet frames and also contains a provision for a quality of service prioritization scheme known as 802.1P, which indicates the priority level of the frame. The 802.1Q standard adds this information to the Ethernet header. The priority level values range from zero (best effort) to seven (highest). You can configure these values to prioritize different classes of traffic such as AoIP versus �everything else� you might use to and from the transmitter (or other) remote site.
One more thing to keep in mind when you build a system such as this: Frames for the native VLAN are not tagged, and thus can�t be prioritized. Don�t send your AoIP traffic in the native VLAN.
Adaptive Modulation and Coding. This is a feature of many radio systems, whereby the modulation, coding and other signal and protocol parameters are adjusted to match the conditions on the radio link. For example, if the link fades, the modulation rate is adjusted to lower the throughput of the system. The idea is that higher-priority traffic will still pass through, even in adverse conditions.
In many cases, carrier-grade microwave systems are made up of an indoor unit connected to the outdoor unit mounted directly on the back of the antenna, as shown. Interconnecting cable is something like inexpensive LMR-400.Remote Access. Radio links such as the ones we�ve been talking about are designed to operate in a networked IT environment, and this remote access is a given. You�ll have full HTTP access to all configurations and parameters. SNMP support is a standard feature of this type of system.
From my own experience I recommend having an additional backup network at the remote end, for two reasons: First, if the radio link fails, and that�s all you have for connectivity, then you�ll be completely blind as to what is going on at the far end. Second, even if the radio link is simply being aimed by a rigging crew, it�s really tough to rely on it to tell you its own signal strength at the remote end.
With that backup network in place, you can log into the remote receiver and see what is going on. Nothing fancy is needed � perhaps the old DSL line you�ve been using up until this point. The scenarios depicted in Figs. 1 and 2 will also allow you to have backup access, of course.
Indoor Unity/Outdoor Unit. Many of the older microwave systems use a very expensive waveguide between the radio and the antenna up on the roof or tower. Today it�s more likely that your system will consist of an indoor unit and an outdoor unit.
The indoor unit can live in your rack room; the outdoor unit usually mounts right on the back of the antenna. The two devices communicate through intermediate RF frequencies that can be accommodated through a much more readily available coax, such as LMR-400.
The outdoor unit is also powered over the same cable. It�s typical of ISM radios to be mounted directly on their dishes as well, and to be connected directly via a Cat-5 or Cat-6 cable, which also carries the power. This greatly simplifies installation. Use direct-burial category cable for this type of connection.
Brands to look at for the IP radio systems are many: Ceragon, Aviat, Dragonwave, Exalt, Racom and Proxim are obvious ones. For ISM radios, one must at least look at Ubiqity and Adtran.�
With the licensed radio links you must of course go through the prior coordination notice process with all the local users, since the frequencies are shared. If the PCN process shows that your frequency choice and path design will not cause interference to other users, you can begin the application process.
The process sounds more difficult than it really is, and there are lots of firms that will do that work for you, such as Comsearch, Micronet, V-Soft, RFEngineers.com, Terrestrial RF Licensing Corp. and Jeremy Ruck and Associates.
One advantage of the current crop of ISM radio systems is their diminutive size, which can be accommodated by a simple non-penetrating roof mount, as shown. Let�s completely shift gears now and go back to our 950 MHz topic � except we�re going to consider a band that is just above the one with which we�re accustomed. Part 101.101 describes the spectrum between 952 and 960 MHz as usable for Private Operational Fixed Point-to-Point Microwave Service (OFS for short). I have seen this band used for at least two applications (both of which were inner-city relays):
� Double-hop STL. There was an application of this band at a station I inherited years ago. Two transmitter sites simulcast the same program; and since they were some 50 miles apart, and out of view from one another, a two-hop STL was needed. Again, spectrum was an issue; so the first link was put on the OFS band I referred to earlier. The second hop went the �final mile� to the transmitter site, and was in the 945-952 MHz band we normally use.
� RPU pickup. I know of two instances where inner-city relay was used to return a very remote RPU receiver audio output back to a physical location from which it was more accessible by conventional means.
If you were to decide to use a system such as this, it would still be necessary to go through the normal PCN process. The good news is that the gear we are accustomed to using � and perhaps you have some of it on the shelf � is usable for this application. Sure, it�s a bit old-fashioned, especially with the availability of the inexpensive ISM-band radios. Still, keep it mind as an option.