Every good work of software starts by scratching a developer’s personal itch.
—Eric S. Raymond, The Cathedral and the Bazaar
In his book The Cathedral and the Bazaar (O’Reilly), Eric S. Raymond explains that “Given enough eyeballs, all bugs are shallow.” The reason open source software development produces such consistent quality is simple: crap can’t hide.
In this era of custom database and website development, people are not only tired of hearing that their telephone system “can’t do that,” but quite frankly just don’t believe it. The creative needs of the customers, coupled with the limitations of the technology, have spawned a type of creativity born of necessity: telecom engineers are like contestants in an episode of Junkyard Wars, trying to create functional devices out of a pile of mismatched components.
The development methodology of a proprietary telephone system dictates that it will have a huge number of features, and that the number of features will in large part determine the price. Manufacturers will tell you that their products give you hundreds of features, but if you only need five of them, who cares? Worse, if there’s one missing feature you really can’t do without, the value of that system will be diluted by the fact that it can’t completely address your needs.
The fact that a customer might only need five out of five hundred features is ignored, and that customer’s desire to have five unavailable features that address the needs of his business is dismissed as unreasonable.[183] Until flexibility becomes standard, telecom will remain stuck in the last century—all the VoIP in the world notwithstanding.
Asterisk addresses that problem directly, and solves it in a way that few other telecom systems can. This is extremely disruptive technology, in large part because it is based on concepts that have been proven time and time again: “the closed-source world cannot win an evolutionary arms race with open-source communities that can put orders of magnitude more skilled time into a problem.”[184]
One of the stumbling blocks of the traditional telecommunications industry has been its apparent refusal to cooperate with itself. The big telecommunications giants have all been around for over a hundred years. The concept of closed, proprietary systems is so ingrained in their culture that even their attempts at standards compliancy are tainted by their desire to get the jump on the competition, by adding that one feature that no one else supports. For an example of this thinking, one simply has to look at the VoIP products being offered by the telecom industry today. While they claim standards compliance, the thought that you would actually expect to be able to connect a Cisco phone to a Nortel switch, or that an Avaya voicemail system could be integrated via IP to a Siemens PBX, is not one that bears discussing.
In the computer industry, things are different. Twenty years ago, if you bought an IBM server, you needed an IBM network and IBM terminals to talk to it. Now, that IBM server is likely to interconnect to Dell terminals though a Cisco network (and run Linux, of all things). Anyone can easily think of thousands of variations on this theme. If any one of these companies were to suggest that we could only use their products with whatever they told us, they would be laughed out of business.
The telecommunications industry is facing the same changes, but it’s in no hurry to accept them. Asterisk, on the other hand, is in a big hurry to not only accept change, but embrace it.
Cisco, Nortel, Avaya, and Polycom IP phones (to name just a few) have all been successfully connected to Asterisk systems. There is no other PBX in the world today that can make this claim. None. Openness is the power of Asterisk.
In the past few years, it has become clear that standards evolve at such a rapid pace that to keep up with them requires an ability to quickly respond to emerging technology trends. Asterisk, by virtue of being an open source, community-driven development effort, is uniquely suited to the kind of rapid development that standards compliance demands.
Asterisk does not focus on cost-benefit analysis or market research. It evolves in response to whatever the community finds exciting—or necessary.
After Mark Spencer attended his first SIP Interoperability Test (SIPIT) event, he had a rudimentary but working SIP stack for Asterisk coded within a few days. This was before SIP had emerged as the protocol of choice in the VoIP world, but he saw its value and momentum and ensured that Asterisk would be ready.
This kind of foresight and flexibility is typical in an open-source development community (and very unusual in a large corporation).
The Asterisk-Users list receives over three hundred email messages per day. Over ten thousand people are subscribed to it. This kind of community support is unheard of in the world of proprietary telecommunications, while in the open source world it is commonplace.
The very first AstriCon event was expected to attract one hundred participants. Nearly five hundred showed up (far more wanted to but couldn’t attend). This kind of community support virtually guarantees the success of an open source effort.
So what sorts of things can be built using Asterisk? Let’s look at some of the things we’ve come up with.
Asterisk can be used as a fantastic bridge between an old PBX and the future. You can place it in front of the PBX as a gateway (and migrate users off the PBX as needs dictate), or you can put it behind the PBX as a peripheral application server. You can even do both at the same time, as shown in Figure 27.1, “Asterisk as a PBX gateway”.
Here are some of the options you can implement:
- Keep your old PBX, but evolve to IP
Companies that have spent vast sums of money in the past few years buying proprietary PBX equipment want a way out of proprietary jail, but they can’t stomach the thought of throwing away all of their otherwise functioning equipment. No problem—Asterisk can solve all kinds of problems, from replacing a voicemail system to providing a way to add IP-based users beyond the nominal capacity of the system.
- Find-me-follow-me
Provide the PBX a list of numbers where you can be reached, and it will ring them all whenever a call to your DID (Direct Inward Dialing, a.k.a. phone) number arrives. Figure 27.2, “Find-me-follow-me” illustrates this technology.
- VoIP calling
If a legacy telephony connection from an Asterisk PBX to an old PBX can be established, Asterisk can provide access to VoIP services, while the old PBX continues to connect to the outside world as it always has. As a gateway, Asterisk simply needs to emulate the functions of the PSTN, and the old PBX won’t know that anything has changed. Figure 27.3, “VoIP-enabling a legacy PBX” shows how you can use Asterisk to VoIP-enable a legacy PBX.
Many people confuse Interactive Voice Response (IVR) systems with automated attendants (AAs). Since the automated attendant was the very first thing IVR was used for, this is understandable. Nevertheless, to the telecom industry, the term IVR represents far more than an AA. An AA generally does little more than present a way for callers to be transferred to extensions, and it is built into most proprietary voicemail systems—but IVR can be so much more.
IVR systems are generally very expensive, not only to purchase, but also to configure. A custom IVR system will usually require connectivity to an external database or application. Asterisk is arguably the perfect IVR, as it embraces the concepts of connectivity to databases and applications at its deepest level.
Here are a few examples of relatively simple IVRs an Asterisk system could be used to create:
- Weather reporting
Using the Internet, you can obtain text-based weather reports from around the world in a myriad of ways. Capturing these reports and running them through a purpose-built parser (Perl would probably eat this up) would allow the information to be available to the dialplan. Asterisk’s sound library already contains all the required prompts, so it would not be an onerous task to produce an interactive menu to play current forecasts for anywhere in the world.
- Math programs
Ed Guy (the architect of Pulver’s FWD network) did a presentation at AstriCon 2004 in which he talked about a little math program he’d cooked up for his daughter to use. The program took him no more than an hour to write. What it did was present her with a number of math questions, the answers to which she keyed into the telephone. When all the questions were tabulated, the system presented her with her score. This extremely simple Asterisk application would cost tens of thousands of dollars to implement on any closed PBX platform, assuming it could be done at all. As is so often the case, things that are simple for Asterisk would be either impossible or massively expensive with any other IVR system.
- Distributed IVR
The cost of a proprietary IVR system is such that when a company with many small retail locations wants to provide IVR, it is forced to transfer callers to a central server to process the transactions. With Asterisk, it becomes possible to distribute the application to each node, and thus handle the requests locally. Literally thousands of little Asterisk systems deployed at retail locations across the world could serve up IVR functionality in a way that would be impossible to achieve with any other system. No more long-distance transfers to a central IVR server, no more huge trunking facility dedicated to the task—more power with less expense.
These are three rather simple examples of the potential of Asterisk.
This little gem is going to end up being one of the killer functions of Asterisk. In the Asterisk community, people find themselves using conference rooms more and more, for purposes such as these:
Asterisk is still too much of an über-geek’s tool to be able to serve in the average home, but with no more than average Linux and Asterisk skills, the following things become plausible:
- Monitoring the kids
Parents who want to check up on the babysitter (or the kids home alone) could dial an extension context protected by a password. Once authenticated, a two-way audio connection would be created to all the IP phones in the house, allowing Mom and Dad to listen for trouble. Creepy? Yes. But an interesting concept nonetheless.
- Locking down your phones
Going out for the night? Don’t want the babysitter tying up the phone? No problem! A simple tweak to the dialplan, and the only calls that can be made are to 911, your cell phone, and the pizza parlor. Any other call attempt will get the recording “We are paying you to babysit our kids, not make personal calls.”
- Controlling the alarm system
You get a call while on vacation from your mom who wants to borrow some cooking utensils. She forgot her key, and is standing in front of the house shivering. Piece of cake: a call to your Asterisk system, a quick digit string into the context you created for the purpose, and your alarm system is instructed to disable the alarm for 15 minutes. Mom better get her stuff and get out quick, though, or the cops’ll be showing up!
- Managing teenagers’ calls
How about allocating a specific phone-time limit to your teenagers? To use the phone, they have to enter their access codes. They can earn extra minutes by doing chores, scoring all As, dumping that annoying bum with the bad haircut—you get the idea. Once they’ve used up their minutes…click…you get your phone back.
Incoming calls can be managed as well, via caller ID. “Donny, this is Suzy’s father. She is no longer interested in seeing you, as she has decided to raise her standards a bit. Also, you should consider getting a haircut.”
[183] From the perspective of the closed-source industry, their attitude is understandable. In his book The Mythical Man-Month: Essays on Software Engineering (Addison-Wesley), Fred Brooks opined that “the complexity and communication costs of a project rise with the square of the number of developers, while work done only rises linearly.” Without a community-based development methodology, it is very difficult to deliver products that at best are little more than incremental improvements over their predecessors, and at worst are merely collections of patches.
[184] Eric S. Raymond, The Cathedral and the Bazaar.