A study from the Uptime Institute reveals that most data centers would max out electrical capacity and cooling capabilities during the next 12-60 months. According to the Uptime Institute, 1U server space costs $1,600 per year in facilities costs, and $700 of that cost is just for electricity.
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.
I always wanted to create a headline without words. This is my chance. You have heard of Multi-Protocol Label Switching (MPLS), but Virtual Private LAN Service (VPLS) may be new to you. MPLS has been available for a few years. VPLS was announced by Verizon in March 2007. AT&T and Qwest are also beginning to offer VPLS, which is a new service for both the service provider and enterprise, so the experience level is low. It is now possible to consider either or both of these services for VoIP calls among sites.
One area where IP telephony has the potential to make a clear impact is in the hospitality industry; most vendors have a solution or set of solutions for this market. One example is Avaya’s new SIP iPhones from Teledex, which let hotel guests experience visuals, which hoteliers should appreciate. Adding the visual to the phone isn’t just cool, but it provides interlinks to the hotel’s services, moving hotels beyond selling rooms at rack rates, beyond just staying competitive, and to potentially greater profitability by adding all the trimmings offered by the hotel and surrounding area of interest to their guests.
First the disclaimers. Hosted Asterisk, IMHO is something to consider only if you are SOHO or very small businesses that don’t have capital for either a TDM key-system, small Hybrid or low end IP-PBX and even then, you’d better know exactly what you are getting into. First impressions and attitude are everything. Never, regardless of how small or big you are, ever, put your dial tone on the line and get into that “we are committed” way of thinking if too many indicators or red flags sound off. You will likely have a bad experience that will stick in the users’ minds for a long time afterwards even if you are successful, and usually you’re just more accepting because you bought off on the concept.
The cost for VoIP/IP Telephony service calls is 150% more than every other category of IT technology--higher than cabling, software, servers, PCs, you name it. This translates into $449 for the average service professional work order as compared to the next highest work order cost of $282 for wiring and cabling followed by software at $255. This is according the report just published by OnForce covering the first quarter of 2008. The report, “State of the IT Industry Report”, covers a wide range of analysis, by locations and IT categories.
I’ve been seeing and hearing about one significant change in the delivery of telephony systems to customers. Distribution is providing the installing company (Interconnect / VAR/Dealer) with staging services to save time and money, and to avoid costly mistakes on deploying IPT systems. Distribution, for a small fee (I’m not exaggerating) will pre-configure the IPT system with the necessary licensing, firmware updates and other initial configuration requirements that are necessary to get the system initially up and running for each customer.
It’s common knowledge that Siemens AG, the Munich based conglomerate that spun off its carrier systems division into a joint venture with Nokia, has been seeking suitors for its enterprise division as well. Nortel and private equity firm Cerberus Capital are the companies whose names most often come up these days and June has been the month when the wedding bells were widely expected to ring. Plans have apparently changed at bit.
In several posts I’ve made mention of the Power-Save capacitor that we have been installing on sites including our own office. We’ve logged our site data including a unit installed at one of our technicians’ home for three months.
What’s really interesting is my site is consistently down 18%, and my tech’s home is down 17% on power consumption.
Via Forbes.com, Charlie Giancarlo told Reuters today that he won't consider the Avaya president/CEO job on anything but an interim basis.
The Lou D’Ambrosio news from Avaya certainly drew a lot of attention. With that attention, new insights emerge, as is often the case.
Judging by the Comments to my original story on Lou D'Ambrosio's resignation from Avaya, and from some email I've gotten, there seems to be a feeling that I was casting doubt on the stated explanation for the resignation. I certainly wasn't doing that.
A commenter on my Unified Communications blog at Information Week (where I moonlight) made an interesting argument, drawing a parallel between Avaya now and Seagate Technologies in 2000, when it too was acquired by Silver Lake Partners. The basic argument is that if things go for Avaya the way they went for Seagate, Avaya's got a bright future ahead of it.
I agree with Eric Krapf’s earlier post and with the Avaya press release, that in roughly two years as CEO, Lou D’Ambrosio has, “shaped a compelling strategy, built a strong team, and led the company through important technology transitions…”
My first thought is to pause for a moment and support Lou in a rapid return of his health.
Then, what about Avaya?
As Eric pointed out, leadership changes are afoot at Avaya. Lou D’Ambrosio has stepped down for medical reasons, Charlie Giancarlo has stepped in as interim CEO, and the company has stepped up its search for a permanent leader.
Here's the letter that Lou D'Ambrosio sent out to Avaya employees today as he stepped down as the company's CEO and president:
Avaya today announced that Lou D'Ambrosio (left) is stepping down as CEO for unspecified "medical reasons," and that Charlie Giancarlo (right), who left Cisco late last year, will step in as interim CEO and President while a search is conducted for D'Ambrosio's successor. Giancarlo left Cisco last December to move to Silver Lake Partners, the private equity firm that acquired Avaya in mid-2007.
My previous blog, “Is Your Network Counterfeit – Part 1?” discussed the discovery by the FBI of considerable counterfeit network equipment. The security of the counterfeit equipment does not seem to be an issue, YET. I worked in military intelligence R&D. The best intrusion is one that is very difficult to detect.
Hre are some comments on Eric's IPT deployment piece ‘What’s Your Biggest Concern?’ My law firm, Levine, Blaszak & Boothby, and our consulting affiliate, TechCaliber Consulting, are doing a bunch of these deals now for Fortune 100’s, so I have both interest and experience.
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.
Last year, I noted in Costco Wants In Too that vendors are doing anything to gain market share in the IPT space. I can’t fault Costco for that, but what I can’t help is even though the gear is good- in these spaces, it almost always is installed poorly. Sure they get a deal, but is it really a deal after it’s been installed- maybe one year later?
The N.Y. Times headline on May 9, 2008 read, “F.B.I. Says the Military Had Bogus Computer Gear”. This headline did not make me feel comfortable. The idea that the equipment, mainly Cisco knockoffs, have been employed in government networks should alarm not only the network operations staff, but the security people as well.
Converged WAN links carry a combination of data, voice and video traffic intermingled. Because these applications use the bandwidth in very different ways, they can interfere with each other when bandwidth is limited, as is often the case in the WAN access link. In my last post we looked at how these traffic types are different. Now I want to talk about how they interfere with each other and how we can manage that interference, especially in the WAN.
Via the VAR Guy, VOIP blogger Garret Smith talks about the apparent slowing of growth in the VOIP market. Smith discusses residential and business, which are very different animals, but appears to see slowing growth in both markets. This notion of a softening market, at least in North America, also emerges from a just-released study by the market researchers at Infonetics.
I just got a demo of some of the new aspects of ShoreTel's recently-announced ShoreTel 8.0 release (announcement here). What's cool about the cool stuff in this product is the regard it has for the communications environment that customers actually operate in today.
No one really likes to discuss what-if scenarios unless of course you work for some underground agency or security firm or are one of those earning a buck to see the dark side of IT and peer into the inner workings of everything vulnerable.
A Review of Webster’s New World Telecom Dictionary by Ray Horak
Many folks who entered the telecom trade during the 1980s and the go-go years of the 1990s came to rely on Newton’s Telecom Dictionary, from long-time BCR columnist Harry Newton, as their guide to the wild and wooly world of telecom.
Some direct responses to my reply about the status of TDM generated some important points made by readers and another publisher. After recently attending a boot-camp training session, I collected a few comments from that group comprising 16 members from large-enterprise including two others from SMB/E.
Understanding how to calculate bandwidth requirements for a converged WAN link requires an understanding of the traffic we are trying to converge. If we were just calculating water flow through a pipe we could add the demands of each appliance that needs water and determine the size of the pipe. But the two types of network traffic we are converging have very different characteristics, and this needs to be taken into account if we want to deliver good performance for both types of traffic. IT teams understand data traffic well because they have supported it for a long time. But real-time traffic is different.
I’ve been spending a little time coming up to speed with Aastra lately. You may recall that the company just closed on its acquisition of Ericsson’s enterprise business. It was the latest in a string of M&A activity that brought DeTeWe, the PBX businesses of Ascom and EADS, and some of Nortel’s European operations into the fold. Also in the mix is Intecom, the PBX developer that continues to market itself as Aastra Intecom; Intecom had been acquired by EADS and was part of its PBX division.
Via the Newark Star-Ledger, Avaya is laying off 400 people, after also eliminating 600 positions via attrition since the start of the year. Earlier this year, Nortel and Siemens had both announced layoffs of more than 2,000 workers each.
We’ve had several No Jitter posts and also VoiceCon webinars that touched on the issue of managed and hosted services for IP telephony and Unified Communications, and the upshot seems to be that enterprises will consider some level of managed service, but probably aren’t yet at the point that they’ll dive into a service provider-hosted solution as a fully-outsourced way to deliver real-time communications.
In my adventures to find more about power generation, storage and transfer, the folks over at DOE gave me some great resources in addition to their Energy Star site:
There are a few network situations where QoS appears to be set up correctly, and yet packet loss still occurs. These situations can be mystifying for those who don’t understand the behavior of the priority queuing mechanisms. Let’s take a look at some of these situations and the right approaches to ensure queuing works properly to give us low packet loss for voice and video.
There doesn’t seem to be much doubt that the concepts of “software as a service” and “cloud computing” and network applications are all on a roll. There also seems to be at least some chance that the economic angst we’re experiencing could make any form of outsourcing seem very attractive. All of this could add up to a kind of perfect application storm, and this would of course create a pretty significant impact on the whole enterprise IT organization and its mission. What has to be done to prepare?
Vampire Loads are also known as Phantom loads that are caused by equipment that while turned off, still draw current that you are paying for. A few months ago we put to task measuring and documenting our internal phantom loads- an inventory of our gear to determine what energy savings we could identify and then achieve.
We've written a fair amount this year around the topic of operational expenses (opex) of IP Telephony. I say written "around" the topic because we've mostly discussed whether the potential for opex savings could be what's driving the market to continue investing in IPT despite the overall economic slowdown. But we haven't really taken a systematic look at the opex picture. That's why I was so glad to get Robin Gareiss of Nemertes Research on a VoiceCon webinar on this topic (go here to get the replay, and here for the archive of recent webinars).
Some large enterprise executives, staff, and economists are among those who doubt that implementing energy efficient network and telephony gear today will have a positive benefit. A theory that came alive during the oil crisis of the 1970s known as the Khazzoom-Brookes Postulate that states, “reductions in energy intensity of output that are not damaging to the economy are associated with increases, not decreases, in energy demand.” This theory goes on to state that “improvements in energy efficiency lead to ever and ever-greater levels of energy usage.”
We had a fantastic webinar last week in which we basically turned the hour over to Robin Gareiss of Nemertes Research, who delivered a really useful talk on opex in IP telephony (watch it here). I'll post this week's VoiceCon eNews here tomorrow, in which I discuss Robin's main findings. In the meantime, some side points were noteworthy.
Back in the early 80’s before adoption of any Category-X standards for wiring, we were faced with a decision that couldn’t wait. At the time, we spoke with and visited AT&T (Atlanta Works Wire Division) Dupont, Belden, and Mowhawk Cable companies. The decision to begin the cabling project was pending as was the effort to begin selection of a product for the cable plant. We were about to wire a national landmark and everyone agreed it needed to last.
In my postings I have addressed a series of the design issues associated with deploying QoS and getting clean voice or video traffic across the IP network. I lump those topics into the ‘design’ category, meaning that you structure the network according to those principals (classification, forwarding behavior, bandwith demand and management, etc). But many of us have a set of problems in the network I call ‘implementation problems’ which are basically bugs that go unnoticed until we introduce real-time traffic.
One of our favorite network troubleshooting gurus, Terry Slattery, has put together a very cool-looking network diagram showing the 25 Biggest Network Problems. Not surprisingly, virtually all of them are either directly or indirectly relevant to real-time/voice traffic. I talked with Terry about some of the high- (or low-) lights.
At Interop last week, I had a chance to sit down with Manfred Arndt, who's Distinguished Technologist with HP ProCurve Networking, which has been aggressively going after market share in the switch/routing business. Manfred is co-author of a standard that's going to be increasingly important as enterprises deploy IP telephony and unified communications: Link Layer Discovery Protocol-Media Endpoint Discovery, or LLDP-MED, which is standardized as ANSI/TIA-1057-2006.
If you're in a conference session, and an Ethernet switch vendor tells you to use 10/100 instead of Gigabit wherever you can, you must be in a session on Green IT.
The WAN optimizer is hardware designed to reduce bandwidth consumption. WAN optimizers are designed for TCP traffic, which dominates the IP network. TCP traffic has a lot of redundancy and can be compressed; it does not have the network performance requirements of VoIP traffic.
Here's a WSJ blog that says, among other things, that IT datacenters are responsible for half as much pollution as the airline industry. Green was a big topic at Interop last week, and I'll have more on it in tomorrow's VoiceCon eNews, which will be posted here as well. But for now, some random facts and comments.
Header compression is always mentioned in the same breath as QoS when we discuss supporting voice on an IP network. But it is not about QoS, it is about reducing bandwidth consumption. Header compression is most important on the WAN because that is where bandwidth is constrained and expensive. Let’s take a look at why header compression helps for voice.
Whenever economic issues threaten budgets, management looks at outsourcing to cover budget shortfalls. Networking in the US has long been dependent on in-house technology while in Europe, managed services have dominated. Given this, US executives are already looking harder at managed services, and network operators here are expanding their programs. The question is whether managed services are a good idea, and if so, where optimum value could be obtained.
What's another word for "features," as in "PBX features" (the infamous list of 500-800)? Well, the word that one of the audience questioners used in my SIP session this morning was: Value. As in, "People I talk to are concerned about the danger of losing value in the system" if they migrate to a SIP-based system that provides fewer functions.
Of course: "Functions" = "Value"
Edgar Masri is out as CEO of 3Com, but the real news is that his replacement, Robert Mao, will be based in China. I think that tells you all you need to know about where 3Com sees much of its future, as I discussed in this post from VoiceCon.
I'll have some longer, more in-depth posts on some cool stuff I heard today at Interop, but for now I've got time for just this quick observation, from Al Baker, VP at Siemens, on the subject of SIP trunk availability.
The three key network specifications for real-time (voice and video) traffic are packet loss, jitter and latency. Whenever I talk to folks about these, they understand the packet loss issue and they worry about the latency issue, but there is often little discussion about the jitter. So I wanted to tackle the jitter topic here and lay out both why jitter is a problem and some of the issues around how we measure it in the network.
There continues to be a lot of discussion of the technology of VoIP and IP Telephony (IPT). There is nothing wrong about technology, but buying or leasing an IP PBX is as much a business decision as a technology decision. Since the IP PBX is not a yet a commodity product, there are lots of hidden gotchas, difficult to understand SLAs, odd pricing models and maintenance considerations.
Word’s already out that the OCGD rates are rising worldwide. Panic sets in with folks and cut-and-run are what some consider or do, when they get their next power bill.
Sound like a story?
In the western Lonesome Dove, Cap’n Call said it best: “There ain’t no excuse for bad behavior”--and reminding the IT guys that this was the moment after the character Captain Woodrow F. Call delivered a severe beating to an offending bad guy.
Sign up for our weekly newsletters and receive the latest news, blogs, event information covering the convergence market.
VoiceCon makes its European debut 14-16 October, 2008, with Johan Krebbers of Shell keynoting. More information HERE
Small-medium business VOIP adoption lags the large enterprise, so SMB is a target-rich environment for vendors. READ MORE »
Mobile UC isn't just a subset of Unified Communications. And although it's emerging quicker in Europe, it will soon affect enterprises everywhere. READ MORE »
Advertisement
 |
||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||
