Feature Articles
Product News
Spotlight Events
Subscribe | Update Profile | Past Issues | Reader Survey | PDF Format | RSS
By Rich Pellegrini, Product Marketing, NMS Communications
The Advanced Telecom Computing Architecture (AdvancedTCA or ATCA) was defined by the PCI Industrial Computers Manufacturing Group (PICMG) to address the need for commercial off-the-shelf (COTS) components capable of supporting highly available (five-nines) solutions. ATCA’s predecessor, CompactPCI, although an improvement over earlier standard hardware components, failed to meet the needs of system developers, primarily in the areas of bandwidth and availability. ATCA addresses these shortcomings and allows system developers to purchase COTS components that respond to carriers’ requirements for value-added services with a shorter time-to-market.
ATCA System Components and Form Factors
Some of the key features of ATCA include:
- Support for up to 16 slots, for high-density solutions
- Larger board size, providing more room for increased DSP and packet processing power
- Cooling for 200W per slot, allowing the use of state-of-the-art processors
- Hot-swappable chassis components, for reduced MTTR
- An Intelligent Platform Management Interface (IPMI) for monitoring the physical health of all components
- 48 Vdc power for continual operation
The foundation of the ATCA family of specifications is the PICMG® 3.0 base specification, which defines the base shelf board and chassis mechanicals, including power distribution, rear I/O, power dissipation, and data transport (for system management and switching fabric interconnect). Additional sub-specifications (3.x) specify the various data plane fabric options/technologies, including Gigibit Ethernet (3.1), Infiniband™ (3.2), StarFabric (3.3), PCI Express™ (3.4), Serial RapidIO® (3.5), and Packet Routing Switch (PRS) (3.6). The data plane (Fabric Switch Interface) on the backplane is wired in one of three redundant topologies—Dual Star, Dual-Dual Star, and Full Mesh. Refer to the NMS white paper, AdvancedTCA: Next-Generation Architecture for Communications Applications, for more details on ATCA.
Additional form factor: Advanced Mezzanine Cards (AMCs)
One of the design intents of ATCA was to offer high levels of modularity and configurability. ATCA achieves this with Advanced Mezzanine Cards (AMCs), mezzanine daughter boards that mount on standard ATCA boards, both processing boards and “dumb” carrier boards. Standard ATCA boards can be purpose-built (e.g., Ethernet switch or DSP) or general-purpose (e.g., single board computer (SBC) or dumb carrier), each of which can provide one to four AMC sites for providing processing, I/O, DSP, disk storage, memory, etc.
Just like the full-size ATCA boards, AMCs are hot swappable, allowing for removal and replacement in a live running system to maintain availability. They can also be used directly in MicroTCA chassis (see Figure 1).
Figure 1: AMCs in ATCA and µTCA chassis
MicroTCA (µTCA)
The MicroTCA chassis standard, now ratified, is a small chassis form factor which allows AMCs to be plugged directly into a chassis backplane without the need for a carrier board. The standard provides most of the features of a full-size ATCA chassis. MicroTCA supports hot-swap, IPMI management, and a switched backplane fabric (AMC.1 PCI Express/Advanced Switching, AMC.2 Gigabit Ethernet (not yet ratified), AMC.3 Storage Interfaces, AMC.4 Rapid I/O), meeting carrier-grade telecom equipment requirements.
Use of AMCs in MicroTCA chassis is lagging about one year behind ATCA for deployed chassis sales. The first MicroTCA-compliant chassis hit the market in early 2007 for evaluation, with initial deployment sales expected in early 2008.
The two primary components of a MicroTCA system are:
- Advanced Mezzanine Cards (AMC)
- MicroTCA carrier hub (MCH)
The AMCs provide processing, storage, memory, and system I/O. The MCH provides system management, including platform management via the IPMI interface, the switch fabric, and performs functions similar to a carrier board in ATCA systems. As many as two MCH cards and 12 AMCs can operate together as a subsystem. MicroTCA provides scalable bandwidth up to 40 Gbit/sec and supports star, dual-star, and full-mesh topologies and provides availability from three-nines to five-nines. MicroTCA shelf densities range from 4 to 48 on average, with the largest shelf offering room for 192 AMCs as a composition of MicroTCA subsystems.
Differences in the MicroTCA chassis versus that of a full size ATCA chassis are as follows:
- Minimal subset of IPMI commands
- +12 Vdc payload power and +3.3 Vdc Management Power to AdvancedMC Modules versus -48 Vdc (-48 Vdc is converted by the chassis)
Form Factors and Market Suitability
While ATCA and MicroTCA are similar in some ways, each is more suitable for particular markets. Full-sized ATCA boards and chassis are best suited for implementing communications equipment that resides in the core and edge of converged telecom networks (e.g., media gateways, Multimedia Resource Function Processors (MRFPs), and the Call Session Control Function (CSCF)), where carrier-grade features and five-nines availability are required.
MicroTCA targets the access part of the network which consists of telecom edge applications (e.g., WiMAX access points, VoIP access gateways, cellular base stations) and customer premise equipment (e.g., enterprise IP PBXs and IP gateways). MicroTCA is also applicable to the medical, digital imaging, industrial, and federal market segments, including defense and aerospace.
NMS and ATCA
NMS has a long history in the development of next-generation standards in industrial computing. NMS was one of the first companies to introduce CompactPCI-based (PICMG 2.x) products in 1998 and is now actively involved in ATCA efforts. We are a constant participant at ATCA interoperability events and other activities. Our ATCA media processing blade, the MG 7000A, offers many high availability features, such as hot swap support, dual PPC controllers, dual Ethernet switches, a craft port, status lights, and front I/O . This special-purpose media processing blade is targeted at the value-added services (VAS) market, where feature-rich multimedia processing is required—such as a media resource function (MRF) supporting a next-generation converged network application.
Main features of the MG 7000A include:
- Up to 480 “universal” ports, allowing one blade to support a variety of applications on either IP or TDM networks
- 16 trunks of rear I/O
- ATCA 3.1 Gigabit Ethernet switch fabric support
- Full support for IPMB chassis management bus
- Fully hot swappable
Figure 2 shows the major components of the MG 7000A.
Figure 2: MG 7000A
Developer Considerations
When developing highly available, carrier-grade, core application-based systems with the MG 7000A, the ATCA chassis must support the following switch fabric options:
- Fabric Option: PICMG 3.1 Gigabit Ethernet
- Fabric Topology: Dual Star, Dual-Dual Star, or Full Mesh as a Node board (and Dual Star as a Hub board)
A unique feature of the MG 7000A media processing blade is that it includes on-board Gigabit Ethernet base and fabric switches and can act as an ATCA hub board in chassis with up to 5 slots, providing a Dual Star switch fabric topology with 1+1 switch redundancy as shown in Figure 3. That is, for a 5-slot media processing chassis, only MG 7000As and an ATCA single board computer (SBC) are required to deploy an ATCA media processing system.
Figure 3: MG 7000A and 1+1 Redundancy
More…
Rich Pellegrini may be reached at +1 508 271 1293 or rich_pellegrini@nmss.com.
By Brough Turner, SVP & CTO, NMS Communications
The following article was featured in the February 2007 issue of Internet Telephony Magazine (www.tmcnet.com).
Fixed/mobile convergence (FMC) has been a hot topic for several years. Korea Telecom launched its “One Phone” service in 2004. British Telecom launched “BT Fusion” in 2005. Then the flood gates opened in 2006, with numerous WiFi-enabled mobile phones and a plethora of companies offering converged voice service platforms. Certainly, the hype is substantial. But is there any adoption? Would you want to use any of these services?
So far, the answer is ‘not much’—not much adoption and no overwhelming desire to adopt. Early service offers haven’t been compelling and user experiences have been cumbersome. There’s no doubt many of the pieces are in place. We have proven VoIP telephony infrastructure from multiple vendors. Now we have dual mode phones from major suppliers including the market leader, Nokia. But the associated hype cycle would have us believe VoIP over WiFi is about to sweep the market, destroying traditional mobile business models. In fact, the revolution will take time. Practical user problems remain, like complexity in configuring handsets and service profiles, limited talk time (WiFi power efficiency is still playing catch-up), and difficulty in accessing any but the home WiFi hotspot. There are two classes of customer and, thus, two paths forward.
Corporate FMC
The first is FMC for corporations. In this market the user experience must be acceptable, but isn’t the deciding factor, if the value proposition is compelling and the support burden is not too great. Many companies are pursuing this path, from small startups to British Telecom, which has launched its Corporate Fusion service in the UK and Spain and is investigating service launches in other countries, including the United States.
Given the high costs incurred when employees use mobile phone minutes at work instead of free PBX services, workplace FMC has the potential to provide significant savings. Corporate IT departments can pre-configure handsets and roaming can be limited to WiFi hotspots on the corporate network, so an acceptable user experience is within reach. Thus, we can expect to see growth in enterprise FMC during 2007 and 2008.
Consumer FMC
The second opportunity is consumer FMC. That’s a lot harder. Consumer products and services can afford fewer glitches. A 1% problem rate swamps any support budget and an even mildly cumbersome startup procedure blocks the majority from ever adopting a service.
Consider the success of Skype, or rising Web 2.0 services like YouTube. In each case, the user experience is simple—it just works! Okay, it doesn’t hurt that the service is compelling and the cost is low or free, but, in each case, the critical difference is an absolute minimum of fuss. It just works. There are no support issues.
For a consumer FMC service to “just work” means trivial software installation or pre-configured handsets, a compelling user interface, but also seamless roaming across unaffiliated WiFi hotspots. Preconfigured handsets are feasible today, but that’s not what’s being delivered. User interface software is an art, but one that’s been mastered in the past, so it will appear sooner or later. However, seamless roaming will take some doing. Many hotspots, even if free, require a browser-based login. And once connected, there’s a question of capacity. Public WiFi hotspots provide 11 Mbps or 54 Mbps connections over the air, but frequently use a DSL line with perhaps only a few hundred Kbps of upstream capacity for Internet connectivity. So, multiple VoIP calls in one hotspot simultaneously are problematic. Further, while my residential wireless router provides upstream priority for VoIP’s UDP packets, that hasn’t been an issue for commercial hotspots, at least until now, so an outgoing email from the next table is likely to glitch any VoIP conversations.
Finally, there’s an extra problem in the U.S. Mobile handsets are pre-configured by our mobile operators to support just the services that mobile operator specifies, locking out features the mobile operator doesn’t wish to support. This blocks 3rd party services, likely stalling consumer FMC in the U.S.
The situation in Europe is much better. Handsets are open. Startups are introducing new services. I have yet to see handset- based software that automatically detects unaffiliated hotspots, figures out how to connect, and does so, but I fully expect such software to emerge. Similarly, today’s software download and installation procedures seem to require dozens of clicks, but someone will figure out how to automate this (or will offer pre-configured handsets).
In short, there is quite a bit of innovation still required before VoIP over WiFi is a done deal, but the window is now open. It may take a year or two, but someone—probably in Europe—is going to figure out how to deliver the needed user experience: launch a service that just works.
Brough Turner may be reached at +1 508 271 1312 or rbt@nmss.com. Also be sure to check out Brough’s blog at blogs.nmss.com/communications .
Product News
In late December, NMS notified affected customers that the CG 6500C, AG 4040, and AG 4040C will be discontinued by the end of this year, according to the following schedule:
|
Event |
Date |
|
Last Order Date |
28 September 2007 |
|
Last Ship Date |
28 December 2007 |
Table 1: Product Discontinuation for the CG 6500C, AG 4040, and AG 4040C
The CG 6500C is being replaced by the CG 6565C, and the AG 4040 and AG 4040C are being replaced by the CG 6060 and CG 6060C, respectively.
CG 6500C
The suggested replacement boards for the CG 6500C are shown in Table 2. The CG 6565C boards have higher performance than their CG 6500C counterparts and are offered at similar prices.
|
Part No. |
Existing Model |
Part No. |
Replacement Model |
|
Existing board |
CG 6500C/ Fusion CG6500C |
Replacement board |
CG 6565C |
|
82047 |
CG 6500C-2L/8TE |
82404 |
CG 6565C/32-2L/8TE |
|
82043 |
Fusion/CG 6500C-2L/8TE |
82404 |
CG 6565C/32-2L/8TE |
|
82048 |
CG 6500C/64-2L/8TE |
82405 |
CG6565C/64-2L/8TE |
|
82044 |
Fusion/CG 6500C/64-2L/8TE |
82405 |
CG6565C/64-2L/8TE |
|
82049 |
CG 6500C/96-2L/8TE |
82406 |
CG 6565C/128-2L/8TE |
|
82045 |
Fusion/CG 6500C/96-2L/8TE |
82406 |
CG 6565C/128-2L/8TE |
|
82244 |
CG 6500C/20-2L/16TE |
82407 |
CG 6565C/32-2L/16TE |
|
82164 |
CG 6500C-2L/16TE |
82407 |
CG 6565C/32-2L/16TE |
|
82088 |
Fusion/CG 6500C-2L/16TE |
82407 |
CG 6565C/32-2L/16TE |
|
82052 |
CG 6500C/64-2L/16TE |
82408 |
CG 6565C/64-2L/16TE |
|
82053 |
CG 6500C/96-2L/16TE |
82409 |
CG 6565C/128-2L/ 16TE |
|
82054 |
Fusion/CG 6500C/96-2L/16TE |
82409 |
CG 6565C/128-2L/ 16TE |
|
82240 |
Fusion/CG 6500C-2L/0TE |
82410 |
CG 5656C/32-2L |
|
82084 |
Fusion/CG 6500C/64-2L/0TE |
82411 |
CG 6565C/64-2L |
|
82085 |
Fusion/CG 6500C/96-2L/0TE |
82412 |
CG 6565C/128-2L |
|
Existing board |
CG 6500C PICMG 2.16/ Fusion |
Replacement board |
CG 6565C |
|
82058 |
CG 6500C/64 (no RTM) |
82420 |
CG 6565C/64-2.16L |
|
82059 |
CG 6500C/96 (no RTM) |
82421 |
CG 6565C/128-2.16L |
|
82072 |
CG 6500C-0L/8TE |
82413 |
CG 6565C/32-2.16L/8TE |
|
82068 |
Fusion/CG 6500C-0L/8TE |
82413 |
CG 6565C/32-2.16L/8TE |
|
82073 |
CG 6500C/64-0L/8TE |
82414 |
CG 6565C/64-2.16L/16TE |
|
82069 |
Fusion/CG 6500C/64-0L/8TE |
82414 |
CG 6565C/64-2.16L/16TE |
|
82074 |
CG 6500C/96-0L/8TE |
82415 |
CG 6565C/128-2.16L/8TE |
|
82070 |
Fusion/CG 6500C/96-0L/8TE |
82415 |
CG 6565C/128-2.16L/8TE |
|
82165 |
CG 6500C-0L/16TE |
82416 |
CG 6565C/32-2.16L/16TE |
|
82089 |
Fusion/CG 6500C-0L/16TE |
82416 |
CG 6565C/32-2.16L/16TE |
|
82075 |
CG 6500C/64-0L/16TE |
82417 |
CG 6565C/64-2.16L/16TE |
|
82076 |
CG 6500C/96-0L/16TE |
82418 |
CG 6565C/128-2.16L/16TE |
|
82071 |
Fusion/CG 6500C/96-0L/16TE |
82418 |
CG 6565C/128-2.16L/16TE |
Table 2: Discontinued CG 6500C Products and Suggested Replacements
AG 4040(C)
The suggested replacement boards for the AG 4040 and AG 4040C are shown in Table 3. The replacement CG 6060(C) boards have much more power than their AG 4040(C) equivalents and include the ability to support VoIP.
|
Part No. |
Existing Model |
Part No. |
Replacement Model |
|
Existing board |
AG 4040 |
Replacement board |
CG 6060 |
|
82100 |
AG 4040/4-1TE |
82926 |
CG 6060/11-2L/1TE (new product) |
|
82101 |
AG 4040/8-2TE |
82383 |
CG 6060/11-2L/2TE |
|
82102 |
AG 4040/16-2TE |
82911 |
CG 6060/16-2L/2TE (new product) |
|
82103 |
AG 4040/16-4TE |
82384 |
CG 6060/16-2L/4TE |
|
82104 |
AG 4040/32-4TE |
82385 |
CG 6060/32-2L/4TE |
|
82114 |
AG 4040/40- 4TE |
82387 |
CG 6060/42-2L/4TE |
|
82235 |
Kit 2 AG 4040/4-1TE |
|
|
|
Existing board |
AG 4040C |
Replacement board |
CG 6060C |
|
82107 |
AG 4040C/16-4TE |
82388 |
CG 6060C/16-2L/4TE |
|
82108 |
AG 4040C/32-4TE |
82389 |
CG 6060C/32-2L/4TE |
Table 3: Discontinued AG 4040(C) Products and Suggested Replacements
The new CG 6060/16-2L/2TE is available for shipment now; the CG 6060/11-2L/1TE is available beginning April 1, 2007.
Contact your NMS sales representative if you have any questions or require additional information regarding these product discontinuations.
Spotlight Events
NMS will be exhibiting at this year’s Spring 2007 VON, being held in San Jose, California from March 20–22. Drop by booth 1019 to learn more about our carrier-grade media platform products, including our Vision VoiceXML Server, Vision Signaling Server & Video Gateway, and ATCA blade.
FREE Spring VON Expo pass—Get yours today!
Click here to register for your free Expo access. Enter NMSC in the priority code box.
Exhibit hours:
Tuesday, March 20—10am-5pm
Wednesday, March 21—10am-5pm
Thursday, March 22—10am-2pm
Meeting request
If you are interested in arranging a meeting with a NMS representative at the show, please email janice_manning@nmss.com, indicating your preferred meeting date and area of interest.
Connect 2007, hosted by NMS, is a series of global conferences that target key business and technical decision makers who are interested in learning about the vital drivers shaping the communications ecosystem—the business issues, the market drivers, the technology challenges, the compelling applications, and the end user demands. This year events take place in Boston (USA), Guilin (China) and Madrid (Spain).
Connect Americas Conference
October 2–3 Boston, MA, USA
Connect Asia Conference
October 17–18 Guilin, China
Connect EMEA Conference
November 7–8 Madrid, Spain
To sign-up for event updates, please click here to fill out a brief form. If you are interested in speaking or exhibiting, please contact Christine Krajewski at Christine_Krajewski@nmss.com or +1 508 271 1129.
On March 27, 2007, NMS will present the web seminar
“Development Tools for Converged IVR Systems”
For over 20 years, companies have relied on proprietary interactive voice response (IVR) systems. As networks converge, speech automation technology and vendors evolve. A more standards-based development environment is needed to simplify the creation of new voice services. The emergence of VoiceXML— the open standard language tool that allows spoken dialog between people and machines—has made it easier to develop speech–based customer self–service applications.
Who Should Attend
Product managers, systems architects and developers interested in IVR systems that can be deployed on fixed or wireless communications networks.
Register now for this web seminar on March 27 at 11:00 am Eastern / 8:00 am Pacific. |