DataComm I: Understanding IP Networks for Data and VOIP
DataComm II: Switching & Routing Technologies for Converged Networks
Voice over IP and IP Telephony
VOIP II: Deploying and Best Practices in the Enterprise
Planning and Implementing VOIP Unified Communications
Managing and Controlling the IT Power Bill
Cost Control of Wired and Wireless Networks
Plus more than 30 programs that can be brought to your site: bcrtraining.com
Fred Knight and I had the opportunity to travel down to Huntsville, AL, last week to visit with the folks at Digium, the company founded by Mark Spencer, creator of the Asterisk open source PBX. What we found was a company that appears to be making the familiar tech industry passage from a young startup focused on breaking new ground, to VC-funded company on the IPO track, focused on execution and building its market.
Digium has always had lots of buzz, largely thanks to Mark Spencer’s high profile, enthusiasm and energy in advocating for open source in the communications world. In classic startup form, the company has brought in a team of industry veterans to push Digium to the next level. Many of these leaders, including CEO Danny Windham and worldwide sales VP Steve Harvey, moved over from Adtran, whose HQ is just across Huntsville’s industrial park from Digium. Another exec, product marketing VP Bill Miller, is a veteran of Fujitsu, 3Com and other venerable networking and telecom companies.
What I wasn’t really aware of was how solid a footing Digium has built: It’s on its 26th consecutive quarter of profitable growth, driven primarily by analog and digital gateway boards, which account for 70% of the company’s revenue. In fact, up until August 2006, all of its revenues were tied to boards, according to Bill Miller, who admitted, “The challenge has been for us to monetize Asterisk.”
Things picked up in a big way last year. The company launched its first Asterisk appliance; concluded three OEM deals (3Com plus two companies not publicly named); and acquired Switchvox, a San Diego-based firm that had built a 400-station-capable PBX on Asterisk. According to Bill Miller, the Switchvox acquisition and product is what will boost Digium out of the low-end SMB market into larger deployments. Digium also launched 24×7 worldwide subscription support services last year.
The company touts its growing customer base, recently announcing that its 4 millionth port of Asterisk was shipped to Wisconsin-based Ashley Furniture, which claims to be the #1 selling brand of home furniture in North America. Among its other installations is a 1,200-seat contact center for Aheeva (a call center/CTI software vendor) which has 4 sites, 28 T1s, 650 lines and 10 Asterisk servers, handling a peak of 160K calls in one day, 55K calls on average. Digium is working on scaling its Business Edition Software’s capabilities, and in the next (4.0) release will start to build some of the clustering capabilities needed to support growing deployment sizes, Miller said.
Cost is still the major appeal of Asterisk, according to Bill Miller, who said ROI for Digium’s AA50 Asterisk appliance or its Business Edition software package can be “months.” SMBs, he said, “don’t care about Asterisk”—that is, they’re not open source advocates—they just want an appliance that works.
The other big advantage is flexibility. Customers can work with their VARs (Digium sells only through the channel) to customize the software in ways that would be much more expensive and time-consuming with legacy vendors’ products, Miller said.
Digium/Asterisk may not be a PBX for the large enterprise today, and if Unified Communications had never come along, it might have continued to show up as nothing more than a science project on the part of a few motivated hobbyists within enterprise IT/telecom shops. But UC is beginning to cause enterprises to think differently (to coin a phrase) about phone systems. I recently spoke with Blake Baxter of Dimension Data, who said that in some client enterprises, UC is “pulling” IP telephony instead of vice versa. This tracks with the general perception that enterprises might want to first find the communications applications that will add value, then build IP telephony out where it enables these apps. In a context like that, call control can become something of a wild card. And so could Digium/Asterisk.
The venture capital firm Matrix Partners invested $13 million in Digium in 2006, and Adtran also holds a 13% stake in Digium. Now isn’t exactly the time for any startup to be bandying about specific plans for an IPO, but that’s clearly where Digium will be positioned in the medium term—which might just coincide with the time when the market rebounds.
What do you think? Drop me a note at ekrapf@techweb.com
Eric H. Krapf
Editor & Lead Blogger, NoJitter.com
VoiceCon Program Chair
This issue of VoiceCon Enews is sponsored by VoiceCon Amsterdam:
The enterprise telephony market in Europe continues its strong growth, and European enterprises are actively exploring their choices for migrating to IP Telephony and Unified Communications, so now is the right time for enterprise decision makers to plan on attending VoiceCon Amsterdam: 14–16 October 2008 at the Amsterdam RAI. VoiceCon is the best place to see the innovators and hear from experts and pioneering end users. We’re proud of our reputation for accuracy, objectivity and technical depth.
Register now and save up to €250.
Over on NoJitter.com, Marty Parker and Fred Knight have had an interesting back-and-forth on the issue of whether users are confused about the value of Unified Communications, whether they’re deploying UC, and generally how big a deal UC is today. Marty is bullish; Fred more skeptical.
Much of their debate centers on a Forrester Research report on enterprise decision-makers’ attitudes about a range of communications technologies, from WAN services to VOIP to UC. I’ve dug through the Forrester report and could easily pull out data to support either Fred’s or Marty’s position. So I’ll leave them to battle it out and call attention to one other question from Forrester’s survey of enterprise decision-makers.
Forrester asked, “How important to your organization is each of the following features of Unified Communications? (very important/somewhat important/not important/don’t know)”. And here are the percentage results, in the same descending order of importance as in the parenthesis from the previous sentence:
Integrated voice, email and IM: 58/38/3/0
Audio, Web and videoconferencing capabilities: 62/33/5/0
Integration with collaboration software: 42/49/8/2
Presence: 38/48/14/0
Business telephone features on mobile devices: 37/43/18/2
Call control from the desktop: 25/51/23/1
Calendars that can be accessed and updated by voice commands: 21/40/38/1
I’m not at all surprised by the position that conferencing took in the survey. Audio conferencing—i.e., getting rid of expensive conferencing services–is the killer app for VOIP/UC cost justification, and videoconferencing is surging in the enterprise. I am a little surprised by the high importance that the respondents placed on integrating all communications channels, primarily because the payback on such integration is extremely soft.
And yet, I was talking to Allan Sulkin not long ago, and Allan asked me, “When was the last time somebody said: I need a new voice mail system; I’ve got to go do an ROI?” It looks to me like this survey is telling us that integrated UC portals like Microsoft Communicator and Lotus Sametime are reaching the same level of acceptance, at least in the minds of IT decision makers. Which means that the UC story has been very good news for Microsoft and IBM, less so for others.
At the same time, the survey respondents seem to have a somewhat constrained—or maybe less grandiose—vision of the world than do the vendors. There’s decidedly less interest in Presence, even though vendors seem to agree that Presence is the heart of UC. All you ever hear about is how you’ve got to know what your co-workers are up to so that you can contact them the right way on the first try.
This survey seems to suggest otherwise. The respondents clearly grasp the idea that unifying a worker’s communications into a single interface will make that worker more productive. They believe that collaboration tools and integration will likewise help end users do their jobs. But fewer of them see the importance of dictating too carefully exactly how those employees collaborate.
Another thing may be that there’s more resistance out there than we realize to the notion that people should be findable at all times—even if the individual gets to configure his or her own findability. Communications professionals and power users may love the idea of Presence, but maybe the same IT folks see less evidence of demand for it among their broader workforce.
In closing, I was also kind of surprised by the relatively low position given to business-feature-enabling of mobile devices. Fixed-mobile convergence clearly is a hot topic in the industry, but if it isn’t so important to port business telephone features onto mobile devices, does this mean that mobility isn’t as critical as we thought? Or does it mean that, for now, Blackberry email and plain-vanilla cellular voice are good enough for the mobile worker?
What do you think? Drop me a note at ekrapf@techweb.com
Eric H. Krapf
Editor & Lead Blogger, NoJitter.com
VoiceCon Program Chair
The Challenge of Unifying Communications and Collaboration
Your challenge: provide simple, effective ways for your organization to communicate and collaborate. IBM understands. You have telephony systems from multiple vendors; you can’t afford to rip and replace; you need to extend your existing investments and shield your users from these complexities. IBM solutions work with the industry leaders; IBM partners are your partners. IBM Unified Communications and Collaboration.
The issue of security for IP telephony is, if not well understood, at least satisfactorily grasped by professionals in the IT/telecom and security organizations today. There’s the gamut of potential problems, which will be serious challenge if and when they actually materialize—like spam over IP telephony (SPIT), eavesdropping, voice phishing and the like. And then there are the problems we see in the wild today, which mostly involve exploits against IP “data” networks that affect the voice traffic running on those networks; basically, when a denial of service or other attack brings down the IP network, it now takes voice traffic with it, or at least it can. Experts like Mark Collier of SecureLogix and the VOIP Security Alliance say such exploits are the real danger for now.
But what about Unified Communications? What unique security challenges could we see when UC starts moving into enterprises?
Ted Ritter of Nemertes Research, who’s a CISSP (Certified Information Systems Security Professional), suggests that UC will pose new security and compliance risks because of several factors. Ted will be discussing some of these challenges in a VoiceCon webinar tomorrow (sign up here).
One of the issues he’ll touch on is compliance. Since Unified Communications leverages more corporate data, there is greater risk of data leakages that violate corporate privacy policies, Ted points out. Also, expanded rules governing legal discovery mean that voice mails—including those embedded in voice mail as part of unified messaging—may be discoverable in litigation.
Ted’s presentation will also touch on the gap between the relatively low incidence of security breaches in early generations of IP telephony, in contrast to the higher risk that’s likely to exist in the world of UC. Several factors account for this; I’ll start with one that Ted doesn’t mention in his slides, but that I think is a legitimate concern: Microsoft will be a much bigger player in the UC future than they’ve been in IP telephony. Microsoft is the target that hackers most relish going after. You can’t ignore this reality.
As Ted Ritter notes, some emerging issues relate to the way that IP telephony has been implemented so far. To date, IPT systems have been deployed largely as islands, connected via dedicated IP pipes or to the legacy PSTN via gateways. In other words, they haven’t been Internet-connected.
One of the key assumptions about UC is that the boundaries of the enterprise will be much more fluid, with users’ need for mobility and remote connectivity driving several trends that can only jack up the security threat level. Those trends include more connection via the Internet, and more use of softphones.
Ted makes an analogy that I’m interested to hear him flesh out in the webinar. He draws a parallel between UC and Service Oriented Architectures (SOA), the technology with which—it’s widely believed—UC will combine to create Communications-Enabled Business Processes (CEBP), which integrates communications into business process apps. Ted’s not trying to sketch out an all-encompassing security architecture for CEBP; rather, he’s pointing out similarities between UC (or UCC, as Nemertes calls it) and SOA. His bullet points:
Like UCC, SOA benefits are increased business agility and flexibility
Like UCC, SOA security must be pervasive, with centralized management
Like UCC, SOA security is very sensitive to performance and must be performance-based to meet SLAs
Like UCC, SOA developers are not security experts
These points address the perimeter-less nature of the communications (perhaps reaching beyond the enterprise to partners and customers); the absolute requirement that performance not be sacrificed either to application behavior or to the security measures used to protect the app; and the organizational need for an even higher level of coordination among the various teams in the IT structure.
What do you think? Drop me a note at ekrapf@techweb.com
Eric H. Krapf
Editor & Lead Blogger, NoJitter.com
VoiceCon Program Chair
This webinar presents a case study from Del Monte Foods, with insights from a leading analyst and technology producer. The session provides enterprise customers with a strategy that enables them to deliver productivity enhancing Unified Communications applications today with minimum impact on their employees, customers, suppliers, telephony and data infrastructure, resulting in a strong ROI on your investments.
Newcomers to the enterprise real-time communications market—Cisco in the late 1990s, Microsoft in the past couple of years—have had to make the case that the change they were bringing to the market was urgent. Perhaps paradoxically, this was because the real-time communications market moves so slowly. As our old friend Hank Levine wrote on No Jitter last week (see “Thoughts on IPT Deployment Concerns”): “The historic tendency in the voice world—unlike, say, PCs—is to ride the gear until it drops, which is more like a decade than the five years it takes to depreciate it fully.”
Thus, the voice market’s historically slow movement put a premium on market-overhang strategies by newcomers. If major enterprises only go out to bid once every 10 years, the new vendors have to make sure they don’t miss that window of opportunity because their technology is perceived as either immature or not adding value vis-a-vis the old technology. They do that by convincing the customer that the risk of buying “old” technology in the replacement cycle is higher than the risk of holding off on the procurement altogether for a couple of more years. Enterprises tend to be more than happy to oblige by delaying purchases as long as possible, as Hank noted.
None of the foregoing is new, but I’m restating it as a lead-in to an interesting perspective that I heard last week from Sonu Aggarwal. Sonu left Microsoft in April after 10 years, and is now CEO of UnifySquare, a systems integration and consulting company that specializes in Microsoft Office Communications Server (OCS) installations. When I chatted with Sonu recently, he said UnifySquare already has three contracts in hand for OCS implementations with global enterprises of 60,000 to 200,000 employees—two proof-of-concept-scale deployments, and one global rollout.
Sonu’s take is that the migration toward the complete Microsoft/unified communications vision is a long-term process: “This is a long transformation,” he told me. “In some senses, it’s a 20-year transformation,” running from 1995 to 2015.
Partly this comes from Sonu’s own perspective, which is as a founder (in 1996) of one of the early instant messaging companies, Flash Communications. Flash built an enterprise-focused IM system, and was acquired by Microsoft in 1998, which is what brought Sonu to Microsoft. So his timeline begins just when enterprises were beginning to see the business value in instant messaging. But the real grabber is his conception that the fulfillment of the UC vision won’t come until 2015.
Or maybe it’s not such a remarkably long time. We’re talking about six and a half years from now. I looked back six and a half years ago, to January 2002, where I found this Miercom test article in that month’s edition of Business Communications Review: “IP-PBXs: Ready and Waiting.” That optimistic title was appropriate for its time, as Miercom found near-perfect call completion rates and increasing feature/functionality. Yet there were still various gaps that you wouldn’t find today: Cisco was the most scalable at just 10,000 stations per system; only Avaya and Alcatel supported 1,000 or more IP stations per system. Shoretel (then called Shoreline) was still selling a system that supported only analog phones.
Meanwhile, Allan Sulkin’s January 2002 market review was titled “IP: Only Bright Spot in a Down PBX Market.” And the first factor that Allan cited for the PBXs’ down year was, “Market overhang from IP telephony.” In that 2002 article, Allan reported that in the previous year (2001), IP station shipments had grown year over year but still represented less than 8% of the total market. By 2005, IP was out-shipping TDM. You could be a TDM bigot in 2002 without looking completely ridiculous. By 2005, you couldn’t.
Of course, there’s no guarantee that UC as envisioned by Microsoft, IBM and their ilk will follow the same adoption curve—but on the other hand, there’s no reason to imagine it’ll be wildly different, either. The controlling factor in voice system adoption is replacement of old systems. What those systems are replaced with depends on the maturity and value of the new technology.
Maybe UC won’t be ready to take center stage by 2015. But I wouldn’t bet against it.
is a global leader in IP communications products including IP-PBX systems, standards based telephones, unified communications and contact center applications. With 29 years experience delivering integrated voice and data communications services to some of the most demanding companies in the world, Aastra Intecom has developed deep industry expertise in solving the telecommunications challenges faced by large enterprises. Our unique perspective allows us to deliver highly reliable, scalable systems, deployed on time and professionally supported.
As IP telephony continues to roll out and begins to scale in many enterprises, what are the real challenges? According to the Yankee Group, some are what you’d expect, some maybe not.
In a VoiceCon Webinar a couple of weeks ago, Sandra Palumbo of the Yankee Group presented some survey data they’d collected from enterprises about technical concerns when it came to IP telephony and Unified Communications. (Go here for the replay.)
Here’s a bit of a surprise: The conventional wisdom about IP telephony voice quality—at least from the vendors—is that it’s rock-solid, absolutely at parity with TDM systems. For the most part, they’re probably right, but there are a few caveats. Yankee Group found that almost 40% of planners in enterprise IT cited “uncertain voice quality” as a main technical concern about IPT and UC; the figure was just over 25% for the deployers in the enterprise—who presumably have a more intimate acquaintance with the technology at work than do the planners.
If I were a planner, I think I’d automatically be concerned about voice quality, if only because voice quality is just about the most important thing, so you’re kind of being paid to worry about it. But with even a quarter of the deployers saying it’s a main concern, you clearly can’t dismiss this issue out of hand. And here’s an especially telling figure: 55% of financial services firms in the Yankee survey called voice quality their number one issue.
My gut reaction is that this isn’t about the IP-PBXs, it’s about the underlying IP network. John Bartlett’s ongoing blog posts at No Jitter, which focus on QOS and data network performance, are consistently among the best-read on the site. A host of problems can bedevil even well-designed networks, and my guess is that enough implementers and planners realize this.
Nevertheless, voice quality was actually a little way down Yankee Group’s list of main technical concerns. Number one, for both planners (50%) and deployers (40%), was “high upfront cost for IPT equipment.” Kind of self-explanatory, I guess.
The next tier, at just over 40% each for planners, was “single point of failure,” “network security” and “high upfront network costs.” Interestingly, the percentage of deployers concerned about these issues was relatively close for network costs (about 35%) and security (about 32%), but significantly lower for “single point of failure” (20%).
Throughout this survey, planners showed consistently higher levels of concerns about IPT and UC technology than deployers did. There was only one exception: On the issue “Lack of skills” twice as many deployers (just over 20%) expressed concern, versus just over 10% of planners.
So the large majority of enterprise IT decision-makers really didn’t see a lack of skills, but it’s noteworthy that the deployers—those who are out there, hands-on—are much more likely to identify this as a problem than are the planners. Sounds like this is a factor whose status the planner would do well to understand within his or her enterprise before proceeding. After all, most of those other concerns—voice quality, security, availability—will likely play out as either more or less of a real-life problem depending on the skill level of the folks deploying and operating the systems.
As for those high costs: Certainly the upfront costs of IPT aren’t any lower than traditional systems, and the concern over network costs indicates that ongoing savings aren’t expected there either. If you’re going to control communications costs via IPT and UC, you’ll have to do it on opex (which is possible; see this No Jitter post).
But that, then, gets us right back to skill levels.
What do you think? Drop me a note at ekrapf@techweb.com
Eric H. Krapf
Editor & Lead Blogger, NoJitter.com
VoiceCon Program Chair
