This is Part 1 of a 3-Part Series click here for part 2 published on April 23, 2009 or part 3 published on April 29, 2009 

Recently, I had the pleasure to hear Gurdeep Singh Pall, the Vice President of Unified Communications at Microsoft, present his view of reality at the Voicecon Orlando Keynote. It was an excellent presentation. Gurdeep was entertaining, informative, and provocative. I am a big fan of Office Communications Server (OCS), and was eager to hear him. At my company we adopted OCS very early – back when it was called LCS (Live Communications Server). We interconnected it to our Mitel 3300 phone system and found that the solution had a significant positive impact on our business. Over the years, OCS has matured and grown – and can now be integrated to a large number of phone systems or completely replace a phone system (Gurdeep’s vision).

Keynote speeches are very similar to late night infomercials. Gurdeep’s presentation was an hour long, and made several points and shots. The goal of a keynote (and an infomercial) is to sell the audience on a particular vision or product – in this case both. I found his presentation perplexing as I agreed with many of his points, but I vehemently disagreed with many of his proofs.  OCS is a very intriguing product from Microsoft – one that will test many of its partnerships. It is clear the product will experience significant growth over the next few years, but I need to take exception to some of the points used in making its case.

The title of the keynote presentation was “Smart Decisions for Tough Times”, with an overriding theme purporting that during recessionary periods, it is time to think outside the PBX and apply transformational solutions as a means to survive and thrive. During the presentation, Gurdeep made five key points that I examine below:

1. OCS is a software-only solution and traditional hardware solutions are obsolete.
2. OCS is a viable product today
3. OCS can be a complete PBX replacement
4. OCS offers unique cost-effective capabilities
5. Organizations must take a radical transformational approach to telecommunications and UC to survive

OCS is a Software Solution

Is OCS a software-only product? Unequivocally yes. Is OCS a software-only solution? Unequivocally no. A complete OCS implementation requires about the same amount of hardware as UC solutions from the traditional switch players. The actual number of boxes and devices in a given solution varies with factors such as scalability and number of locations.

For a complete OCS implementation, the core “backend server” should run on dedicated hardware, and Microsoft even recommends running OCS on multiple servers with high availability and load balancing. The company also recommends an OCS mediation server to manage the PSTN and other gateway functions. An additional OCS server may be needed in the DMZ for remote access and federation services – key parts of the solution. For voice mail, Exchange is recommended on yet another server or cluster. SIP trunks can be connected logically, but PSTN or analog services must go through separate stand-alone third party hardware devices. Then there are the phones, optional as Microsoft points out, but from yet another vendor and likely requiring an additional TFTP server for configuration. Note that Microsoft recommends SIP phones optimized specifically for OCS. Polycom, with the ever-popular Soundpoint IP SIP phones, created an entire new phone line (the CX Series) for OCS. 

Comparing OCS to the Avaya Call Manager solution offers surprisingly similar results. Avaya’s software (sold as a solution, not as software only) runs on a dedicated processor – typically an HP or Dell server class device. It utilizes gateways for the DSP resources that separate the server from the network. Voice mail can be a separate server or completely integrated into the gateway. SIP trunks don’t require specialized Avaya hardware, but will require some network gear for voice prioritization and SIP-aware firewalls (same for OCS), otherwise PSTN hardware is typically directly integrated into the gateways. For IP connections, phones or soft phones connect to the gateways via separate or integrated LAN switches.  The gateways can directly support analog ports for analog phones or devices. The phones don’t require additional LAN ports as the phone consumes one connection and the desktop can connect to the phone (one drop per desktop). For UC-type applications, additional servers for presence/IM, for example, can be implemented, or the solution can interconnect to OCS or Sametime Servers.

Regardless where you are coming from- whether a PBX with UC applications or a presence engine with telephony capabilities - the end solutions overlap significantly. The amount of hardware required is fairly comparable. While I used an Avaya example, Cisco, Mitel and the others are utilizing similar amounts of hardware in their solutions. Every enterprise class PBX offers soft phones, and the decision to go with a “soft phone only” implementation is independent of the UC solution selected. The fact is, “soft phone only” implementations are not very common for a variety of reasons, and I do not expect that to change with the increasing popularity of OCS. Now I have to point out that during the anti-desktop phone part of the presentation, the slide said two words: “Think Differently”. Was he proposing replacing IP phones with Macintoshes?

It is also worth noting that every “PBX” maker is very likely spending much more money on software development than on hardware development. That crossover took place years ago. I agree telecommunications, or specifically UC, is a software-dominant solution, but I don’t believe this is a differentiator among the offerings.