June 11, 2008

Out-of-the-Box Compatibility? - Excerpts from a TMC Article on IMS Interoperability

Richard Grigonis of of TMC (IMS Magazine, Internet Telephony Magazine) published an article on "Interoperability and Other Testing in IMS". A portion of the article was based on an interview I did with him several weeks ago.

See the whole article here.

Out-of-the-Box Compatibility?

At Empirix, one of the most respected names in the testing hardware and software business, Chad Hart, Product Marketing Manager, says, “Interoperability is still a challenge. It has come along way but it still has a ways to go. As a result of my conversations with network operators who are actually deploying IMS, most of them are choosing to go with a single vendor as opposed to many. It’s not the best time for achieving interoperability, but it’s good for IMS in that operators are actually deploying it. Interoperability is one of the bigger challenges faced by IMS adoption. The way the IMS/NGN Plugfests are set up, they are usually done in stages. The first stage is just establishing a basic call or doing a basic registration between different vendors, which is what practically the whole first Plugfest was about. At the most recent Plugfest, vendors achieved the first step very quickly. It’s a sign of real progress for interoperability. How quickly can vendors’ equipment actually work together practically right out of the box?”

“Are operators now so comfortable with multivendor operability that they’re willing to make deployments in their live networks?” asks Hart. “I don’t think that’s true on a large scale yet. Certainly some of the larger operators have pretty sophisticated interoperability labs where they’re demonstrating multivendor interoperability internally, but they haven’t deployed to the networks yet, though of course they have plans to do so. Another trend is that many vendors are aware of the interoperability issues and have partnered with some of the best-of-breed component vendors out there to make sure they can interoperate with at least specific vendors or devices.”

“There are of course economic considerations,” says Hart. “While it might be theoretically possible to mix-and-match all of these various IMS functions from different vendors, that doesn’t necessarily imply that such a practice is economically viable. If you buy a lot of equipment from one vendor, they’re going to give you a better price than if you just buy one component from one vendor and one from another vendor. The reality is that interoperability is probably going to be done in ‘layers’. Certainly there are many different kinds of application servers out there, along with many different services. One main goal appears to be to achieve pretty broad interoperability with the service layer. Similarly, there’s a lot of different kinds of network edge access equipment. So you need to have pretty board interoperability with that equipment. Interestingly, in the network core it doesn’t seem to be as important for most operators to have a variety of different vendors represented there in their infrastructure. A typical operator is quite happy dealing with one vendor who controls the core network, as long as they’re compliant with the IMS specification.”

“Also, there are vendors out there such as BEA that are working on service delivery platforms or the app server layer, and companies like Broadsoft, that has made excellent traction in this space with their IP Centrex-type business VoIP tech services,” says Hart. “But those are application server vendors who plug pretty neatly into the IMS architecture, at least in its simplest form today. There are things on the horizon that may muddy the picture a bit. Today, many vendors just plug into an IMS core via an ISC interface. [IMS architecture identifies the interface between a SIP applications server and the S-CSCF as an ISC or ‘IMS Service Control’ interface, to perform multimedia session control.] There are other interfaces specified in the IMS architecture and those interfaces are getting to be much more complex and there are things that allow application servers to talk to other application servers or to have app servers ‘share’ common elements amongst them, such as presence. The interoperability of devices with those interfaces is probably a few years off, so it won’t be commercially viable for a while yet.’

“There’s also data sharing among these applications servers,” says Hart. “The IMS architecture allows applications to store user profile information and application-specific information on the HSS [Home Subscriber Server] user database in the IMS architecture. But it doesn’t appear that very many vendors are supporting that capability today, and the specifications for actually putting and sharing that user data in a common location is not a mature practice yet. More standardization work needs to be done.”

“The MSF, or MultiService Forum (News - Alert) holds an event every two years called the GMI, or Global MSF Interoperability event,” says Hart. “The MSF is backed by the major service providers. The most recent one was in 2006 by BT, France Telecom, NTT and Verizon here in the U.S.

At the GMI, they test multivendor operability, such as aNortel ( News - Alert) device talking to one from Ericsson. But they also test network-to-network interoperability. There will ultimately be multiple IMS networks out there, and the networks themselves will talk to each other, as opposed to just the vendors’ devices within those networks. That’s another layer of complexity, since every IMS-based operator is faced with hundreds of configuration options. If some of those configurations are a bit ‘off’ they can cause interoperability issues between the networks when you try to pass a call from one network to another.”

One should note that the MSF has a permanent test bed at the University of  New Hampshire InterOperability Lab (UNH-IOL) with the aim of testing NGN (Next Generation Networks (News - Alert)) components and interfaces. The MSF NGN Test Bed supports testing of all MSF Implementation Agreements (IAs) and will grow to encompass future IAs on topics includingIPTV, Quality of Service, Location Management and support for Operational Support Systems. Aside from in-depth testing based on key GMI interoperability scenarios such as the optimal media routing of IMS interconnection between subscribers in the MSF R3 domain and legacy networks, the MSF NGN Test Bed will provide demonstration and test facilities supporting MSF members' in-house laboratories.

“Our test and monitoring equipment at Empirix continues to be refined, and we help quite a bit with the challenges of IMS interoperability testing, such as complexity,” says Hart. “We help simplify test scenarios and help make the automated tests easier and able to handle more tests. There are so many tests you can run today, simply because there are so many more things to test than previously. In other areas, you could be dealing with a standards ambiguity of IMS, since there are always changes to specifications. Fortunately, it looks like the Release 7 specifications are pretty locked down. Most operators, however, are still implementing Release 6, but all vendors are working with Release 8 specs, which are largely still to be defined or are otherwise in progress. Adapting to all of these changes is important along with being able to adapt quickly, which is another key capability that we’ve built into our products.”

“Our monitoring system, Empirix Hammer XMS, has also become an important piece of equipment,” says Hart. “Basically we took the concept of our Call Analyzer and asked ourselves, ‘What if we wanted to do analysis and diagnostics on a live operational network?’ Instead of the Call Analyzer being a kind of a spot diagnostics tool, we wanted to create a network-wide monitoring system that could call upon multiple distributed probes and scale up to millions of subscribers. That’s our monitoring system.”

June 05, 2008

Border Gateway Function Confusion

BGFs have me confused.

I am doing some research on SBC's in IMS for and came across the Core Border Gateway Function (C-BGF). We did not identify this element in the book or poster and now it seems to be referenced fairly frequently.

After my colleagues (Gordon Beith and Darius Minai-Azary) and I did some research we have yet to find a consistent definition of the C-BGF's various interfaces. In some references the C-BGF appears to be the same as the A-BGF. In others the C-BGF talks to the A-BGF.

Some of access-network to core call flows I found include:

  • A-BGF->P-CSCF
  • C-BGF->P-CSCF
  • A-BGF->RACF->P-CSCF
  • C-BGF->SPDF->P-CSCF
  • A-BGF->C-BGF->P-CSCF
  • PDN/WAG->C-BGF->P-CSCF

In the book and poster we showed the A-BGF talking to both the PDF and P-CSCF. It was assumed the PDF was part of a larger RACS and therefore we did not show a RACF.

The MSF R3 architecture defines 7 different Session Border Gateway variants:

  • SBG - Session Border Gateway (combined S-SBG and D-SBG)
  • D-SBG - Data Path Session Border Gateway
  • D-SBG-NC - Data Path Session Border Gateway, Network Core (i.e., facing a peer network)
  • D-SBG-NE - Data Path Session Border Gateway, Network Edge (i.e., customer-facing)
  • S-SBG - Signaling Path Session Border Gateway
  • S-SBG-NC - Signaling Path Session Border Gateway, Network Core (i.e., facing a peer
  • network)
  • S-SBG-NE - Signaling Path Session Border Gateway, Network Edge (i.e., customer-facing)

2 Border Gateway Functions:

  • C-BGF - Core Border Gateway Function
  • I-BGF - Interconnection Border Gateway Function

A several policy and resource control functions:

  • PDF - Policy Decision Function
  • SPDF - Service Policy Decision Function
  • BM - Bandwidth Manager

I was considering building a big table to map and compare the various architectural differences, but I am not sure if it is worth it at this stage given most BGF vendors could probably fit their products to any of the architectures if needed. Most IMS implementations include a single SBC. A few include a distributed SBC, but nothing as distributed as what is called out in the specs. So I guess I will go on being confused by all this until I find a good unifying reference or make one myself.

May 14, 2008

Some of my commentary on the state of IMS for IMS Vision

I recently did an interview with Daniele Tricarico of IMS Vision. Here are some excerpts from the article:

While interoperability is a core value, this situation makes it exceptionally hard to get vendors from diverse backgrounds and with conflicting commercial interests to interwork.

Chad Hart, Product Marketing Manager at Empirix, a US based testing, monitoring and performance solutions provider, believes that complexity and standards ambiguity are the greatest challenges for the true realisation of IMS. "The complexity of the IMS architecture is the main obstacle to real deployment together with standards ambiguity, even though the industry has come a long way in the last 2 to 3 years. On one hand we see consolidation with the operational convergence between 3GPP/3GPP2 and ETSI TISPAN but, on the other hand, standards are not always clear. For example, if we look at the status with VCC for fixed mobile convergence, it is clear that we don't really have a standardised blueprint on how to build an FMC network."

A considerable amount of testing activity has been recently conducted and initiatives such as the IMS Forum Plugfest and the Global MSF (MultiService Forum) Interoperability events (GMI) have been particularly successful. GMI for instance is focusing  on providing a technical validation of the MSF Release 4 Architecture and supporting implementation agreements as the basis for interoperability between vendors within practical network deployments.   

If testing and monitoring was not much of a concern in the early days, the tough reality of real world IMS implementation has put a renewed focus on equipment interoperability testing in recent years and as more operators move to live IMS networks, either partially or completely, network monitoring is the next challenge. 

You can view the whole article here: http://www.informatm.com/newt/l/imsvision/viewarticle.html?artid=20017529071

May 12, 2008

I-CSCF does have an ISC interface (it is called Ma)?

In case you didn't see it, A couple of weeks ago Aayush posted a great comment on my post correction on Figure 3.13:

I-CSCF has to have an ISC interface. How else will it route incoming INVITE requests addressed to pre-configured sub-domain PSIs to the SIP-AS? Sometimes the I-CSCF sends an incoming INVITE directly to the AS, and the S-CSCF is not required in the signaling path for iFc processing. Similarly, the AS can also send requests to the I-CSCF directly.

I-CSCF and AS interaction is given in TS 23.228, Rel-8..Section 5.4.12.3 point c and Section 5.4.12.5 point c.

Gordon Beith (the Hammer G5 Product Manager and book co-author) and I did some research on this. Gordon found the following:

Actually, it [the AS to I-CSCF interface] was assigned an interface name, Ma, in R7.

In 23.002-710, it lists the following…

This interface between Interrogating-CSCF and the Application Servers (i.e. SIP Application Server, OSA Service Capability Server, or CAMEL IM-SSF) is used to forward SIP requests destined to a Public Service Identity hosted by an Application Server directly to the Application Server.

Details are described in TS 23.228 [34], subclause 5.4.12.

23.228-760 clause 5.4.12.2 states…

  1. The HSS maintains the address information of the AS hosting the PSI for the "PSI user". In this case, the AS address information for the PSI is returned to the I‑CSCF in the location query response, in which case the I‑CSCF will forward the request directly to the AS hosting the PSI.

That’s about as much as it says in R8 as well.

We checked our notes and could not find any customer off hand who was implementing the "ISC" on the I-CSCF to check (this does not mean they aren't though). We wrote the book based on R7 which further adds to the confusion. We did make reference to the Ma interface from the VCC AS to the I-CSCF in Section 5.5.2 in the book and Figure 5.9, but not in the Section 3.3.2 that covers the I-CSCF. We will see if we can dig into this a bit more with some customers.

Anyone else out there implementing this spec?

May 01, 2008

More on SBC's

I came across this article on Acme Packet today where Seamus Horihan was defending the SBC as a stand-alone network element commenting: "If you look at the history of communications equipment, while many people have pursued [multi-function, all-in-one] 'God' boxes, none were ever bought or delivered," he said.

I completely agree with Seamus - every IMS operator I have spoken with, and most of the major vendors, are using independant, integrated SBC's in their network. In fact almost all of them are using Acme - independant of who their core provider was (NSN, Ericsson, ALU, etc.). So it seems the SBC's have made a strong foothold in these arly IMS.

However, I found the 'God' box comment interesting - if anything SBC's are becomming 'God' boxes of IMS, encompassing many functions including the A-BGF, C-BGF, I-BGF, IBCF, IWF, P-CSCF, I-CSCF, SEG, PDF/SPDF, and or RACF in many cases. Most SBC vendors offer a distributed SBC architecure that breaks these functions out, but it seems most operators are choosing an integrated approach in these early stages of real deployments.

From a quality perspective this makes implementation easy - fewer components and interfaces means there is less you have to test (see my post on Why Test if I have a SBC from last week for more info on SBC testing). Verifying the configuration of one device is easier than verifying the configurations of 5. But is also means less flexibility for later and more trust in your SBC vendor and system integrator. It is difficult to peer inside the inner workings of the SBC to see what it is really doing, which is fine until something goes wrong. Time will tell how scalable this approach is.

April 24, 2008

IMS MARKET NEARS $1.2 BILLION

I was doing some research and came across this article from 2005 IMS MARKET NEARS $1.2 BILLION. 2.6 Billion by 2008? No way! Then I clicked the link and remembered this article was based on the IMS market research study I produced back in 2005.

So why did I predict the IMS market would be so huge back in 2005 and where is this $1.2 Billion in IMS equipment today?

First off I should note that my original definition of "IMS Infrastructure" was pretty liberal. It included big-ticket devices like Trunking Gateways and Gateway Controllers that fit cleanly into the IMS architecture but were not really anything new. It also included many SIP-based App Servers and SBC's that plug nicely into IMS and already existed. These pre-existing devices made up most of the forecast. The newly-designed CSCF's and HSS's were very a small part.

Today I would say CSCF's and HSS's are still a small part of the overall IMS/NGN market in terms of revenue. However, we see many service providers using these devices in serious lab tests and field trials. In most cases they are supporting pretty limited configurations - integrated CSCF's as opposed to distributed, and HSS's with low subscriber licenses.

I do think this will start to change in 2008 - especially after many of the recent conversations I have had with carriers in the past few months. You can see my What Happened to IMS? post for more details. Still, we are a long way off from $2.6 Billion. $50 Million in true-CSCF and HSS sales in 2008 would be a good start...

April 21, 2008

Why do I need to test if I have an SBC?

Nearly all the operators I have talked to who are deploying IMS networks are using SBC's (mainly from Acme Packet). I have noticed a trend where many of these operators, especially the smaller ones, are not doing much testing of the situations described in Chapter 4 because they have an SBC and "that's what the SBC does". As a reminder, the book covers the following topics in Chapter 4:

  • Outage Recovery,
  • Peer network floods, and
  • Malicious attacks like DoS, Registration hijacking, and CallID spoofing

Yes, SBC's are designed to deal with many of these problems and they typically do a good job of it. But is your core network designed to deal with the SBC? SBCs are complex devices that require a lot of configuration. The goal of testing these scenarios is not to break the SBC, but to ensure the SBC is properly configured and optimized for your environment.

Using the Power Outage Restoration example, the SBC can throttle messages sent to the Core, but how do you know what your core can handle? What about the customized application you are running? Your system integrator will likely be very conservative and go with a low number, but this could lead to unnecessary capacity upgrades when they are not needed. A few relatively simple tests can help to answer these questions and keep your integrators/vendors honest.

April 09, 2008

First Poster Correction - Gq/Gq' Interface

I actually noticed this as soon as I got my first printed copy, but it was too late to correct in the first print run.

On the "IMS Core Testing Reference Guide" section, in the P-CSCF Test section, we show the P-CSCF to PDF interface as Gm. This should be Gq or Gq' as we describe in the call-out text. This corresponds to Figure 3.14 in the book where we illustrate it correctly. Here is the book graphic:

F314_pcscf_dut

Gq is usually used to refer to the original 3GPP-wireless oriented interface to the PDF. Gq' usually refers to the ETSI TISPAN implemention and connection to the RACS. Since the ETSI spec transfer I have noticed more references to Gq' in the new and in-progress R8 specs.

3GPP TS 23.517 version 8.0.0 Release 8, section 12 has some high-level explanation.

April 01, 2008

IMS Poster is now available!

Dan Teichman, Empirix's Monitoring Product Marketing Manager, and I put together a new poster covering IMS Testing and Monitoring. You can sign up for the poster here http://www.empirix.com/imsposter.

One side of this poster focuses on Testing- highlighting an example IMS implementation and showing individual DUT diagrams for each individual IMS function and showing key test metrics for each.The other side shows a monitoring perspective – this section highlights key protocols used in a complex IMS implementation and shows key monitoring metrics for several IMS services. The content is based on the IMS book with some updates.

Revisions to the poster will be posted here. 

Here are a few low-resolution examples of what the poster looks like.

Ims_poster_monitoring

Ims_poster_testing

March 27, 2008

A few corrections to the first revision

Here are a few corrections/clarifications I have been meaning to post. Most of these are due to inconsistencies between different R7 specifications, so they are not necessarily wrong, just outdated or not yet synchronized:

Chapter 3

  • Section 3.2.6 and Section 3.3.3 - there are some inconsistencies on communications between the A-BGF, the PDF, and the P-CSCF. This is very ambiguous in the R7 specs and has been somewhat cleared up in R8. I will post more on this at a later date.
  • Figure 3.20 - Mj should be added to the circular interface showing the Gateway controller connection to itself (indicating an connection to other Gateway controllers). Mj is used for BGCF<->MGCF communication
  • Table 3.18 - the MGCF<->BGCF interface should be Mj, not Mi as indicated in the table

Chapter 5

  • Figure 5.3 - the "Isolation" diagram - the green line from the Device Emulation Box to the IP Centrex AS should be labled SIP, and the purple line should be labeled "Diameter"