IP Convergence: Beyond VoIP, Beyond Cost Savings - GLOBALCOMM

Virtualization – Part 3 - The Abstraction of Applications

Fri, 03 Oct 2008 13:53:58 +0000

Sorry for the long delay between postings. Within the last 10 months I have moved into more of a customer facing role and find myself – always with a customer and no time to blog. I recommend you start the series by reading “Virtualization – Part 1 - The Abstraction of the Internet” and “Virtualization – Part 2 - The Abstraction of a Computer”.

This is the third of a five part series on the evolution of virtualization.

Let’s start off by reviewing the definition of an “Application Programmers Interface (API)” from http://www.wikipedia.org/ :

An Application Programming Interface (API) is a set of functions, procedures, methods or classes that an operating system, library or service provides to support requests made by computer programs.

Language-dependent APIs are available only in a particular programming language. They utilize the syntax and elements of the programming language to make the API convenient to use in this particular context.

Language-independent APIs are written in a way that means they can be called from several programming languages. This is a desired feature for a service-style API which is not bound to a particular process or system and is available as a remote procedure call.

You find APIs within your desktop operating system (Windows, Linux etc…) . Applications that you use (i.e. computer programs) utilize the operating system’s API to read data from your CD, your keyboard, mouse – In fact any hardware device.

Traditionally, applications using APIs are in your operating system and in a format called “machine code”. “Machine code” is the actual 1’s and 0’s that your Intel or AMD processor can understand . Humans need to translate the 1’s and 0’s to something more readable which is called “Source Code” (e.g. abstracted programming commands in languages other than machine code like “C”, “C++”, assembler , etc…).

Having the processor read and execute machine code is much faster than having a run time interpreter execute using an abstracted language – or at least it was ;-)

You also find APIs between two (2) applications (computer programs) where one application needs the other application to perform a task or get some data.

In both of these examples – Applications using APIs to use an Operating System or Applications using APIs to use another Application – traditionally were developed by programmers creating source code , then using another tool called a complier to create machine code allowing the system to work faster.

The problem is when one system gets upgraded , some of the APIs may break while supporting previous revisions of software or some may not work at all.

The concept of complied APIs was useful as it allowed for problems to broken down to smaller pieces and compartmentalized, however it created a sustaining engineering problem.

The next evolution of APIs were APIs that were not complied or dynamically linked at run time (e.g. windows .dll files) , they were interpreted.

Interpreted languages are just what they sound like, in real-time an interpreter is being executed by the processor , and interprets a more readable and sustainable command into something that the operating system and processor can understand.

Early examples of interpreters are LISP, API which were not designed to be machine independent.

Today we all use a machine and operating system independent interpreter with JAVA.

These solutions addressed the issue of sustainability while providing a method of breaking down a problem.

While all of these software engineering problems were being addressed, the hardware and network engineers were making processors’ faster, memory cheaper and networks faster and faster. The internet provided a learning platform for abstractaction of text based languages like HTML, XML.

And that is where the world of APIs, XML and web services all come together.

The advancement of hardware, networks and programming methodologies have brought us to the point of having sustainable systems that are distributed across many computers and networks.

We have machines sending HTML pages between a server and your browser. Your experience on web server can change very quickly by having the web master change HTML text and make it available immediately. We have machines creating HTML pages automatically and changing your experience in real-time.

We have machines sending XML pages between two (2) different servers on different networks to create a trouble ticket or check the status.

The idea of having machines exchange information is not new, but it is a relatively new idea of using extensible abstracted languages’ that take advantage of performance gains in computing and networking,

Together they form the concept of web service either available within a private enterprise network (Intranet) or available on the Internet.

Well take this a bit further as well explore the next topic:

Virtualization – Part 4 – Virtualization of Voice Communications
Telephony basics in the circuit switched voice network and the evolution of the packet switched voice network , eg. VoIP.

Google Phone + TellMe Services = Google 2.0

Tue, 20 Mar 2007 14:19:30 +0000

Just posted a comment on a fellow blogger's post on the Google mobile phone.

In that Google's announcement is not a position against the iphone, but more importantly against Microsoft with their latest announcement of their acquisition of Tell Me.

Do you know what Tell Me does?

Well, Tell Me provides a highly scalable Speech Recognition Interactive Voice Response platform for the Fortune 100. For example, if you call United Airlines you'll experience Tell Me and can tell (no pun intended) that it's Tell me when you hear the gears between you speaking and that platform deciding what you may have said and the next menu prompt that may be associated with your input.

Back to google, Google is google not only to their brilliant search engine , but more importantly how it makes money. Google has created a totally automated advertising market between marketers and their markets

Totally automated in the sense that a marketer selects words that are important to their offer and bid against click thoroughs (not purchases). Have a hot word in a hot market, pay a lot

They limit their bad collections by requiring marketers to pay upfront where google can deduct the click thorough in real-time. Don't have enough money in your account, that's ok, google will offer the next higher bidder to the search request.

Brilliant technology working in a brilliant business model. No people , no bad collections and all to the highest bidder.

Back to Google 2.0, if you take the Google cell phone that is to be mass marketed and to a market that doesn't use a PC and add the best of google and tell me and you have their next market disrupter.

Google 2.0 = Google Cell Phone + Tell Me like services + Google Adwords will result in game changing play as to how traditional telephone directory services (yellow pages and directory assistance) make their money.

I expect to blog more of this reality in 18 months

Will 2007 be the year of significant VoIP Adoption?

Mon, 26 Feb 2007 22:04:32 +0000

Seems that the financial community has started to understand how companies offering VoIP services need to be reevaluated.

Today Credit Suisse reiterating their position on Sonus Networks, resulting in an almost 6% increase in stock price - Sonus' 52 week high.

Look at Vonage and they have over 2.2 million subscribers, an almost 50% increase from the 1.4 million subscribers in 2005.

For Global Crossing, the future looks brighter everyday. And it's just not me saying it, it's the analysts and media as well.

INTERNET TELEPHONY magazine has recognized Global Crossing VoIP On-Net Plus Service™ with its 2006 Product of the Year Award.

VoIP On-Net Plus is a service that recognizes that transformation into the VoIP world is not a flip of a switch, it's a transition over time. VoIP On-Net Plus allows an Enterprise the ability to share a private dial plan across their TDM and VoIP sites.

We use it internally and it's great to not have to remember 10 plus digits to contact any resource across the globe.

Check it out!

A Call to Action by the VON Coalition

Mon, 18 Dec 2006 20:18:22 +0000

Although I support and wish them luck, a call to action to remove VoIP barriers internationally is a daunting task.

In Jeff Pulver's blog dated December 15th, 2006, I quote my comment dated December 18th, 2006:
"Jeff [Pulver], I wish you and the VON Coalition good luck.

However it may be an up hill battle that is based on each countries political and bureaucratic process. Let’s not forget that the incumbents grew out of the PT&Ts (Post Telephone & Telegraph) and they would act in the best interest of a country and not the market.

History has many examples of how this played out, remember the proprietary signaling protocols across international boundaries, remember that some PT&Ts were successful in blocking entire NPAs in an effort to stop international toll bypass.

Countries like India have embraced VoIP as a technology enabler to grow their share of the outsource call center market, but stop short of allowing VoIP to originating or terminating within their PSTN.

Other countries like China have adopted a policy of licensing and setting the cost at (un)reasonable level – estimates as high as $2M US.

You may want to take the India policy as an example of allowing VoIP to enable a market. And replay this to countries like Malaysia which are up and coming outsourcing countries – as a start."

Take a look at a Global view VoIP regulations as developed by the Global IP Alliance.

Typically, countries have taken an open market approach to spur innovation, a middle ground approach to control innovation and a “not in my back yard” approach to stifle innovation.

A PT&T approach of stifling innovation is a near sighted perspective to protect tax revenue and the incumbent’s control – far a market perspective the market will move those calls via VoIP to countries that are either in the open market or middle ground approach.

For example, US and Europe are open market countries, India is a middle ground country, and China is “not in my back yard” country.

So how can you move China into the middle ground?

I think we should use India’s logic, in providing cost effective communications (VoIP inbound/outbound) to support outsourcing while leveraging low cost labor.

Global Crossing announces VoIP Peering

Sat, 10 Jun 2006 17:01:22 +0000

Global Crossing announced the re-release of our VoIP services portfolio at GlobalComm this week, which essentially means that we added a variety of new VoIP related features, both to the VoIP platform (VoIP Peering, Voice VPN) as well as the IP-VPN network (cRTP and FRF.12).

The most exciting piece, hands down, is the VoIP Peering. Andy Abramson picked up on it in his blog early in the week, and he says :

Add Global Crossing to the growing list of Peering Fabric type of companies. Ipeerex, Xconnect, Stealth Communications and others.

One has to wonder if Level3, AT&T and Verizon will also all start offering this too.

One things for sure, when calls stay on net, the costs drop, and when they peer over the same carrier, the costs go to almost nothing. As a result if everyone can cross peer with one another costs go way down, so why will anyone need Skype?

I agree!

But, we've got some pieces that the peering fabric companies and Skype don't have. For example, the peering fabric companies have no services to offer that I am aware of. If you want DID's, local service, Outbound/LD, dial-plan management and 800 number service you'll have to get them from somewhere else. While you can get many of those features with Skype, try hooking Skype up to your enterprise PBX!

We are hoping that this will be a truly disruptive feature for enterprise class service providers. Although right now we don't have any big carrier partners as bilateral peers, we think that it won't be too long before we get some. In the mean time, every customer calls every other customer for free. Not bad, eh?

edit: See this Alec Saunders article on a Skype gateway for the enterprise.

IMS - Fixed Mobile Convergence

Mon, 31 Jul 2006 14:40:28 +0000

I sat on my 2nd GlobalComm panel on Wednesday, again at Telecom U Summit, titled IMS: Fixed Mobile Convergence. This is my first post that mentions IMS and FMC together, and hopefully you'll see why by the end of this article.

The attendance was maybe only half of what the talk from the day before on IMS and Security was, probably due to it being the last day. The show floor was fairly deserted, as you can see in this picture taken by Brough Turner.

Among my co-panelists was Ben Vos of Sprint Nextel, their VP of Core Network Architecture and Strategic Planning group. Sprint Nextel is apparently very optimistic about IMS, to the point where they really seem to be driving the development of IMS, which is in contrast to Global Crossing where it is viewed with what I think it is a healthy skepticism by most of us, except my colleague Gary Miller. :-)

All the panelists seem to agree on several things:
1) IMS is not required for many if not most fixed mobile convergence features
2) IMS is very complex, perhaps too complex
3) IMS is a reference architecture, and it is unlikely that anyone will implement all elements of the architecture. More likely they will pick elements that are important to the services they want to provide and go with that.

I was asked by the moderator what elements the carrier should implement, and I responded based on a tip I got from the moderator the day before. "It depends." It depends on who your target customer is and what solution you want to provide to them. Every different market segment has a different set of needs, and every service provider is at a slightly different place in the supply chain, many (MVNO's for example) not even having a mobile or fixed network. These differences will drive you to evaluate each element of the IMS architecture individually for it's role in your service architecture.

IMS, in my opinion, gives you two main things that you can't get easily otherwise. The first and possibly the most important is that your view of the customer goes from a fairly wide view as defined by where the call originates into your network (we call them trunk groups, which are groups of T1 trunks plugged into a PSTN Gateway, or an dedicated IP address on a session border controller) down to a very specific view, the individual subscriber. Where as previously a subscriber might be only identified by ANI, which is unreliable at best, IMS enables you to identify a subscriber across one OR MORE trunk groups, sometimes without even knowing ahead of time what path the subscriber is coming in on (e.g. the Internet). The second big thing you get is call state, which even current VoIP networks do not keep track of. Call state enables certain billing functions to be tracked better, but also enables presence. When combined with subscriber information and the subscriber profile, you can create services where the subscriber is in the drivers seat in definining who can interupt them when and who gets instant voice mail.

I also had the pleasure of spending some time with a variety of folks from Hammerhead Systems including one of their senior developers, Ping Pan. Hammerhead has an edge aggregation box that we have been evaluating in our labs for some time now, and so you can imagine my surprise when I discovered that Ping knew quite a bit about IMS. He has been preparing for the fact that some day their switch/router will take provisioning information from an IMS system, providing either positive or negative feedback about the success of that provisioning request so that it can be fed back into the session control aspect of the IMS core. I had been thinking that IMS might play a role in actual network policy and provisioning, and this was confirmation that I was on the right track.

On the flip slide of the IMS coin, I visited Personeta's booth and saw a demo of their sim(ultaneous)-ring and mobile-to-fixed call transfer application running on their Application Server TappS. They demonstrated a PC calling with SIP into an SBC, the call then rung on a blackberry and an IP phone. He picked up the IP phone and answered the call, pressed ## and continued speaking into the mobile phone while the IP phone began ringing. A moment later he brought the handset of the IP phone to his hear and continued his talk on that while his mobile phone disconnected.

All running on a fairly typical application server, and no IMS in sight.

Tags: GlobalComm, IMS, FMC

Pulver party at Globalcomm 2006

Fri, 09 Jun 2006 22:00:19 +0000

Jeff Pulver put on a kicking party Tuesday night at the House of Blues in Chicago during the GLOBALCOMM2006 show.

Good food and an open bar until 10pm weren't the only benefits, but the stage was held by a cover band known as Herding Cats (and occasionally a few women who were given the opportunity to dance on stage during the set).

Apart of the fact that Jeff saying so a few times, you could tell how much he liked them. I have to wonder if the name of the band has anything to do with it, a common metaphor often used to describe activities in project management as well as the technology-heavy industries. And of course, we can't forget our favorite herding cats commercial from EDS years ago.

The only thing I wish I had more of is an opportunity to network during the party. My colleagues and I chose a side-stage box office that was a bit quieter and at least allowed us to have a bit of a conversation during the show, but in the main space it was impossible to try and hold a conversation with someone. One other benefit of our location was not only a great view of the band but also of the people right in front of the stage. I think I spotted Brough Turner and Jeff along with about a dozen others that danced during most of the show, while the other 900 of us stood around like geeks at the back of a middle school dance.

Tags: GlobalComm