OS-VoIP | Open Source VoIP » IP PBX

Writing an RFP for an Open Source IP PBX – Part1

Aaron Rosenthal — Fri, 20 Mar 2009 16:02:23 +0000

Most large enterprises would naturally write an RFP for something as critical as their communication systems but also these are the Fortune companies who still haven’t really caught onto the awesomeness of Open Source IP PBX systems, though sooner or later they will and there are some things they should know when writing an RFP specifically for an OS IP PBX. And even if you’re not a Fortune company, you should still write an RFP… honestly, if you’re looking to invest anything over $100K on a phone system, you’d be silly not to have an RFP. Increasingly I’ve found that there is a growing number of large enterprises interested in evaluating an Open Source IP PBX and for those who do, one should understand the differences between an OS IP PBX and a proprietary IP PBX enough to tailor their information in an RFP to fit the realities of an Open Source IP PBX.

I thought I’d write an article about RFP’s for Open Source systems because too few companies write them. The process of writing an RFP can be just as useful in helping a company determine their own specific communication needs as it is for a vendor in determining what those needs are. Most of an RFP, whether for a proprietary or Open Source system, will likely be fairly similar except for a couple key OS VoIP areas which include – interface requirements, redundancy requirements, and management requirements. How you effectively outline your requirements for these three areas will largely dictate what type of Asterisk based IP PBX a vendor will quote for you.

Before I get into specific details about an RFP, I want to make sure that you understand a few important conceptual differences between a proprietary IP PBX and an Open Source IP PBX that will help you understand what you’re getting into. I might bring up these conceptual differences now and again… and I’ll start them with “TIME TO THINK DIFFERENT” just for fun…

TIME TO THINK DIFFERENT- It is an undeniable truth that most Open Source companies get a D for marketing and sales material in comparison to proprietary vendors. The simple reason is that proprietary PBX vendors have the cash $$$ to blow on marketing and most Open Source firms don’t (guess where those licensing fees go?). What results from this dynamic is that most proprietary vendors can show up with sexy clear cut marketing material touting all the bells and whistles of their IP PBX systems. This “loud” marketing material gets customers all riled up about the cool, new, and interesting things a pbx can do. Of course this makes sense, people prefer to learn visually and that’s what marketing material is for.

But anyone who knows this industry will tell you that there’s often a big difference between how marketing departments price and sell telecom solutions and how those telecom solutions are actually engineered. For example, proprietary PBX vendors will convince a company to buy a $10K magic box to expand their exsiting PBX’s voice mail capabilities when in technical reality that box is usually 80% empty and is nothing more than a couple $100 RAID1 hard drives. Imagine if proprietary vendors actually charged what things truly cost (plus a reasonable margin)? Now on the flip side, Open Source IP PBX vendors, the ones who really understand the technology that is, will sell their solution based on the cost to build it… hardware+software+development.

Ok, back to marketing material…. the thing about an Open Source IP PBX like one built with Asterisk is that you are literally faced with an UNLIMITED number of options for what you can do with that system. So rather than being presented with a list of capabilities which is what proprietary vendors do, many Open Source vendors prefer not to put that box around their customers by limiting capabilities to a simple sheet of paper or product brochure. If I were to write marketing material for all the things you could do with Asterisk, and trust me I’ve tried, the resulting product would result in a compendium of work no man or woman would ever care to read. Instead, companies looking for an Open Source IP PBX need to think a lot harder about what THEY want and what THEY need versus going the easy route of just picking a bunch of features off a page. And, if experience serves me right, too few companies actually address their own telephony needs because they’re so accustomed to waiting on a vendor to simply tell them that “these are the features you’re going to get” & “this is how its done”… hence why I’m writing this document – KNOW YOUR REQUIREMENTS BEFORE ANYTHING ELSE…. then put them into an RFP…. makes so much sense doesn’t it….

It’s undoubtedly a daunting task to be told “your IP PBX can do anything” (and it literally can) and then being asked, “now what do you want it to do”… but that is the case so instead of someone giving you parameters for functionality, you need to set your own when looking at Asterisk. DISCLAIMER – Yes systems like Switchvox and Trixbox have a definable set of features packaged into different software tiers much like proprietary systems, they even carry per-extension licensing fees like proprietary systems, but they’re also not your only Asterisk based option which is why you NEED to outline requirements because there might be a better Asterisk solution which is more appropriate for your company. Plus, I want to focus on large enterprises and unfortunately the bundled Trixbox and Switchvox options don’t satisfy organizational requirements that demand more than 150 simultaneous calls whereas Asterisk alone can handle many times that in the right deployment. I’ve worked with Asterisk systems (often built using additinoal complementary Open Source VoIP software) capable of supporting over 20,000 simultaneous calls. So anyone who questions Asterisk’s ability to reliably support large call loads either doesn’t know what they’re talking about or are scared shitless that their proprietary ways are in serious jeopardy so they’re just in denial.

As a side note, Open Source routers and session border controllers are also extremely stable. We’ve worked with software such as OpenSIPS/SER and I’ve seen these systems route well over 80 million minutes/mth through carrier networks. For big enterprises, OpenSIPS might be part of your Open Source IP PBX solution….ya never know.

Where to start:

Ok so you’ve been assigned the responsiblity to source a new communication system for your firm. What do you do? “Oh, that’s easy” you say, I should start contacting vendors and see what my options are for replacing my janky ass key system…. WRONG! The very first thing you should do is determine what your users need, this is the RIGHT APPROACH. Consider you have a clean slate, anything goes, it’s Christmas, and anything you could ever want in a communications system is possible.

TIME TO THINK DIFFERENT -If you’re used to the proprietary PBX world, you’re probably thinking “well there’s always a big different between what we want and what we can afford”. And you would be correct, some features and functionality cost more than others. But, compared to proprietary systems where you’re accustomed to every single extra non-out-of-the-box feature costing money, this will likely not be the case for an Open Source IP PBX.

There’s a big difference between how most proprietary vendors like Cisco and Avaya price their IP PBX systems, and how Open Source systems are priced. The cost of most proprietary systems are usually a reflection of the market and what a company can afford to get away with yet still remain competitive. Usually this pricing is in the form of licensing fees, sometimes hardware costs, and usually if you want more features you’ll be paying [license fee]x[number of users] for a set of particular features. Open Source systems are quite different, and again it depends on what type of OS PBX you’re looking at, but in my experience the cost of an Open Source IP PBX is a direct result of the “engineering time involved in getting the thing configured + hardware + software”. And I guarantee, when talking about a big phone system, it will always cost less to custom develop a complex telephony feature using Asterisk than it would cost to purchase that same feature from a proprietary vendor.

Requirements Gathering:

So in writing an RFP, there’s always a few standard procedures all of which start with “Requirements gathering”. This is the process where you go to all your departments and listen to them either bitch about features they wish they had, or highligh the features they can’t live without.

Some companies prefer to gather requirements by forming an adhoc committee made up of individuals appointed from each department or division. This may make sense for a larger company where department heads can filter their groups requirements into a larger committee pool, but for smaller companies it might be just as effective to notify managers about the impending technology purchase and have them gather some comments/suggestions from their employees.

Information Technology
Operations
Sales/Marketing
Accounting/Finance
Executives
And whoever else I forgot….

Often the above departments might have their own opinions about how a new communications system can improve productivity or provide a competitive edge over your competitors. Let your employees be creative in listing new features which might make a difference in your operations. I’m going to list some department specific out-of-the-box features in Part2 of this article so hold tight.

And here’s where I stop and tell you to wait for the next installment of this article. Hope you enjoyed it and stay tuned for more tips about writing an RFP for an Open Source IP PBX.

Premise based Open Source IP PBX or hosted VoIP?

Aaron Rosenthal — Tue, 08 Jul 2008 21:54:37 +0000

This is a question I hear constantly and the answer typically relies on three things – cost, business size, and required functionality. Most small-mid sized businesses exploring the move to a new phone system will consider the ups and downs between purchasing a premise based IP PBX or a hosted VoIP solution. These days I think purchasing an old TDM phone system instead of an IP based system is about as smart as buying a VHS player so lets assume that most businesses won’t go this route. Cost, reliability, scalability, and functionality each vary depending on which solution your business chooses.

In true OS-VoIP fashion, I’ll make the comparison between hosted VoIP and an Open Source IP PBX because there are some unique advantages an OS IP PBX has over hosted VoIP. I’ve seen plenty of companies decided that hosted VoIP made more sense than a proprietary premise based IP PBX but once you throw an OS IP PBX into the mix, the metrics can change quite dramatically.

Cost:

A customer premise OS IP PBX will require an upfront capital expense plus the additional cost of yearly maintenance which can vary in price depending on the level of maintenance required. Although an OS IP PBX does not hold the same costly licensing fees associated with proprietary systems, the capital expense is still not one to overlook. Depending on the type of OS IP PBX, you’ll probably discover a savings of 30%-50% compared to its proprietary counterpart.

By using a hosted VoIP service you’ll avoid this capital investment but may soon find that the operating expense usually exceeds the capital cost of a customer premise OS IP PBX in as little as 2 years. Because an OS IP PBX can be 50% lower in cost, your ROI time frame is half that of a proprietary system. It’s obviously much easier for a business to stomach a 2yr ROI than a 4-5yr ROI compared to what you would have paid for hosted services. The same reason why you wouldn’t rent a car for 2 years is exactly why you wouldn’t “rent” your IP PBX. There is still the cost of telecom services which are required with a premise based system but at least you have the flexibility of using whichever telco service you happen to find as the most cost effective and reliable- this may include POTS, T1, SIP, or some other internet based VoIP service (which I recommend against for a businesses).

I will argue that under most circumstances, a business under 25 seats will probably find more cost advantages for hosted VoIP than in purchasing an OS IP PBX simply because the investment required for a premise based IP PBX is usually more than a small business can afford. When it comes to larger organizations who focus largely on the TCO, a premise based OS IP PBX will win hands down in cost when you spread that cost out over a few years.

Hosted VoIP solutions are usually based on flat fees costing anywhere from $39 to $69 per month per station depending on the features required. Most hosted VoIP providers also require that you purchase your own IP phones which typically start at $100 for a decent device.

Hosted VoIP providers typically have a few pricing tiers which are based on features plus you may find an a la carte selection of advanced features you can purchase on a per/mth basis. Sometimes a business will be forced to purchase a higher tiered hosted VoIP product simply because they needed just 1 advanced feature which is not included in the lower tier. Some smaller hosted VoIP providers may be flexible enough to work out special pricing but this is not the norm.

OS IP PBX system on the other hand come standard with a full set of features that require no additional fee beyond the cost of the system itself. More advanced features and functionality may be an additional cost, but this cost need only be made once because there aren’t any ongoing licensing fees associated with an Open Source system. When you add a feature to a proprietary system, you may be required to pay a fee for every extension on that system, this is not the case for an OS IP PBX and therefore advanced features a far more affordable for larger systems.

Reliability:

Hosted VoIP providers treat reliability as a number 1 priority since they are responsible for the voice services of countless customers. The infrastructure on which your VoIP service operates are located in data centers with backup power, systems monitoring, high security, and much more. Since each hosted VoIP provider delivers VoIP service to thousands of customers on a single system, they can hardly afford an outage. That being said, with all the levels of redundancy in place at a hosting provider, the most common outage is typically not with the hosted VoIP infrastructure itself but with the services delivering voice calls from the hosting provider to your business location. VoIP delivery methods range from your existing internet connection, to dedicated T1’s or MetroE circuits. VoIP services delivered from most hosting providers run over a single circuit with zero redundancy. If for example, your T1 or cable goes down, so does all your voice service. This is one of the biggest problem hosting providers face when trying to deliver a reliable voice product to their clients. And because many hosting providers still rely on other Local Exchange Carriers such as Verizon for delivering these circuits, your business is now at the mercy of multiple telecommunication companies.

There is one vital flaw with hosted VoIP services and this flaw is that the majority of hosted VoIP providers rely on the internet to transmit an IP voice call. Unlike traditional phone service which utilizes a carrier’s network to send and receive calls over the PSTN, the quality of a hosted VoIP call can be largely dependent on the “weather” of the internet. Sometimes packets are sent from A to B without any loss but sometimes if the Internet is having a bad day, you’ll lose some precious data packets which will result in a choppy voice call or even worse, a dropped call. This is why most hosted VoIP providers don’t offer SLA’s since there are too many factors and external parties involved in delivering a stable voice call. This isn’t to say that VoIP over the internet doesn’t work, because it does, and it works well much of the time. It’s just not possible to guarantee the same results to every customer, and if your business needs its phone service to work 100% of the time then using an internet based hosted VoIP provider is not my recommendation.

Instead, the most reliable way to go hosted VoIP is from a carrier who owns their own network. Companies like M5 Networks will sell you a dedicated T1 over which your VoIP service is delivered. This T1 does not hit the Internet and instead carries a call from your desktop, over the T1, then directly into the carriers network, their PBX (which gives you features), and out through to the PSTN. There’s no need to hit the internet which dramatically reduces the chances of packet loss and poor voice quality.

It would be hypocritical to say that a premise based IP PBX isn’t susceptible to such service outages since often telecom services to an IP PBX are delivered over the very same T1’s, PRI’s, and MPLS circuits as hosted VoIP solutions. The reason why a premise based IP PBX can be more reliable than hosted VoIP is because a premise system can utilize redundant services such as the 100 year old technology called POTS (copper lines) which to this date is still one of the most reliable. “Should” a primary voice circuit fail, the PBX will automatically route calls over POTS thus maintaining full system functionality during a circuit outage. The cost of maintaining a few POTS for redundancy is something most small businesses can afford and certainly worth while should a primary voice circuit fail.

One great thing about hosted VoIP is that you are never responsible for its overall health and availability. Just set it and forget it! A premise based IP PBX is a piece of expensive equipment that you’re responsible for. The unique advantage of many Open Source IP PBX systems is that they’re built on the very same hardware that your IT staff are already familiar with. I would recommend that every IP PBX be supported by a vendors maintenance plan, but simple things like replacing a hardrive can be quickly done by any average IT employee.

Scalability:

Hosting providers are in the business of scaling a system infinitely. If they couldn’t, they wouldn’t be able to add additional customers. As your business grows, so can your hosted solution. What one must not forget is that for every user added to a system, there is a direct linear cost associated with the number of users added. Most OS IP PBX systems have the ability to scale significantly but instead of sharing system resources with thousands of hosted customers, a premise based IP PBX is all yours.

Hosted VoIP scales easily- just place the order and you’re good to go. This does mean that you’re entirely reliant on the schedule of your hosted VoIP provider. If you have purchased a support plan with your OS IP PBX, typically you’ll find that most vendors have a 30-60 minute response time to MAC requests which can be done remotely and usually accomplished quicker than the amount of time it would have taken with a hosted VoIP provider. There are obviously hosted VoIP providers who can be quick and IP PBX vendors that take forever so just make sure that each company has an SLA that meets your needs.

Functionality:

Functionality and features vary from hosting provider to hosting provider, but it’s safe to say that most will
deliver many of the standard features businesses require such as voice-mail, call transfer, call forward,
auto-attendant, and more. In some cases, hosted VoIP may come with very similar features to those included standard with an OS IP PBX.

The downside is that the features available with hosted VoIP solutions are limited to a finite a la carte menu. Some may come standard, where others may have an additional monthly fee, and some advanced features aren’t even an option. One of the most popular uses of VoIP for mid sized businesses is having the ability to integrate with other 3rd party internal applications for more efficient business processes. Hosted VoIP rarely has this level of versatility simply because the customer does not have the necessary access to their communications system in order to achieve integrations of this type.

Whether you need it today or not, having the ability to implement 3rd party integrations, unified communications, collaboration, call center applications, and many others is the reason why many organizations choose a premise based OS IP PBX. The Open Source nature of an IP PBX gives an organization a lot more flexibility in how quickly and easily an integration between the IP PBX and 3rd party application can be achieved.

This may change in the near future as I expect many hosted VoIP providers will begin partnering with 3rd party application developers. Already companies like Ribbit have integrated their VoIP services with Salesforce.com’s CRM package.

Tips:

The advent of VoIP as a mainstream business product has prompted thousands of hosted VoIP companies and Open Source IP PBX vendors to open up shop. I think it is almost too easy to get into the hosted VoIP business and companies like Fonality will recruit anyone as a re-seller of their phone systems. My recommendation to any small business owner is to use well established companies with existing customers and a good track record. If you’re going with a hosted VoIP provider, make sure you research whether they’ve had many network outages and for how long. Network outages happen, but they should never be frequent nor should they be for long durations….obviously. I would advise against using an internet based VoIP service if call quality is of high importance but maybe I’m just being a perfectionist since this is what most people do.

If looking at an OS IP PBX, make sure you use a company who fully understands the inner workings of the product they’re selling. Open Source systems require a more in-depth technological understanding than a simple plug and play proprietary system. Many OS IP PBX systems have easy to use admin interfaces but If you ever run into a complication with your IP PBX, you’ll be very happy to have a vendor who has the right engineering talent to solve the problem quickly.