 |
|
One of the most intransigent barriers to full convergence between data networks and the telecom network has been the continuing presence of a large amount of legacy TDM (time-division multiplexing) traffic. The willy-nilly, arbitrated nature of most early computer networks did not play well with the incredibly strict timing requirements of TDM. Even today, with the increasing acceptance of VOIP, there is still a massive amount of TDM traffic which will not be going away soon.
Because of the ongoing need to accommodate TDM, most makers of high-capacity computer-based telecom equipment were ultimately forced to add an auxiliary telecom bus such as the H.1x0 bus in the case of PCI and CompactPCI. Since its introduction, an entire class of devices has sprung up in support of this bus, including framers, TSIs and DSPs. The existence of these devices was an asset that the designers of AdvancedTCA could not ignore.
The issue of TDM traffic gained a completely new level of complexity with the creation of the AdvancedTCA architecture. Now, instead of a shared parallel bus, AdvancedTCA introduced a switched fabric backplane, and so the H.1x0 bus was hardly an option. A great deal of work went into solving this issue, and in March 2005 PICMG ratified a new specification—PICMG SPF.1 ‘Internal TDM’—designed to incorporate TDM traffic into the AdvancedTCA ecosystem. (Note: It is fair to assume that MicroTCA, as a subsidiary specification to AdvancedTCA, will offer these same capabilities.)
Virtual H.1x0
Right from the start, one of the goals of Internal TDM (I-TDM) was to accommodate the large number of existing telecom related devices such as framers and DSPs that were already designed to work with the H.1x0 bus. Since these devices were already in existence and had a successful track record, there was no need to re-invent the wheel.
In the end, the solution chosen by the PICMG specification writers was to design a protocol that would replace the physical H.1x0 bus with a virtual H.1x0 bus. From the perspective of the existing framers and DSPs, the H.1x0 bus exists, and they continue to function as before. They forward their flows to an I-TDM device that encapsulates them for transport across the packet backplane to the target device on the other end of the link. There, an analogous I-TDM device converts the IP packets and forwards them to the TDM bus device and the legacy network devices are none the wiser.
Layer 2 ½ Shim
All this sleight of hand is based on a new protocol specification, PICMG SFP.0, which is the System Fabric Plane Format. The System Fabric Plane is considered a Layer 2 ½ protocol because it is inserted between Layer 2 and layer 3 by the system. Protocols such as the SFP have been around for quite some time and are common in most routers and firewalls, etc. They are commonly called “shims” because they are inserted between the two layers. In the past, these shims have been proprietary protocols, but the PICMG SFP.0 is an open, standardized shim.
SFP.0 is a generic shim that can be used to encapsulate several different flavors of packet or cell-based traffic, such as TDM and ATM, in order to allow it to be carried on a switched fabric. SFP.1 (I-TDM) is a specification that has been optimized for TDM traffic that's carried over Ethernet or another high-speed switched fabric. Other specifications will likely be released over time for other types of traffic, and they will carry similar identifiers such as SFP.2, SFP.3, etc.
Compliant Products Available
The SFP.0 and SFP.1 specifications have been ratified and available for nearly a year now and SBS has already released compliant products. For example, we have had an 8-port T1/E1/J1 AdvancedMC on our product list since SuperComm of 2005.
This article is intended to give only the most cursory overview of the general ideas behind I-TDM. For a much more thorough discussion of the details of Internal TDM please download the SBS white paper entitled “Internal Time Division Multiplexing (I-TDM) for AdvancedTCA.”
Download I-TDM White Paper here
|
|
| |
 |
 |
Rubin Dhillon, V.P. Communications & Enterprise
I braved the wrath of my wife for being away from home on Valentine’s Day to attend 3GSM World Congress 2006 (in romantic Barcelona of all places). Thank goodness for high-speed Internet service, new nifty Internet phones, and of course an Interflora bouquet to help with the situation.
|
3GSM World Congress is one of the biggest mobile events of the year, and you couldn’t help but be impressed by the size and enthusiasm of this event when you approached the 3GSM show grounds. Approximately fifty thousand attendees were there to attest to the fact that the 3G wireless market is one of the hottest segments in the Convergent Communications market. Money is being made--and spent--on all of this stuff, and those of us in the open-architecture, modular computing camp (i.e. AMC & ATCA), were very happy to participate.
So what was the buzz at 3GSM? It was mostly Digital Broadcast Video and IMS technology and their real world applications. For example, Intel was demonstrating near-NTSC resolution digital broadcast video over DVB-H streaming technology using only 200Kbytes/sec bandwidth speed. This is a significant application breakthrough and it will enable the next generation of mobile TV broadcasting. In fact, DVB-H technology is capable of delivering quality transmissions to mobile devices that are traveling at over 75 mph.
Why would people watch TV on a cell phone anyway? Well, according to the large number of content providers at the show, the biggest audiences for mobile IP/TV are people who wish to shoot at aliens (video games), and those looking for adult entertainment. Given the significant number of content providers with, shall we say, very interesting adult booth displays, it is the latter audience that vendors are counting on to drive revenue growth in the IP/TV space.
IP/TV and IMS equipment require significant processing power for digital video broadcast servers and voice/data/media signal interworking. Telecom equipment manufacturers (TEMs) are clearly looking to Intel, IBM, HP and Sun for the hardware to build their next generation IP/TV and IMS servers. Using open architecture, modular building blocks such as ATCA, AdvancedMC and MicroTCA is the quickest route for vendors to deliver next generation equipment.
Roaming the aisles at 3GSM, I saw ATCA systems prominently displayed in the Alcatel, Continuous Computing, HP, Lucent, Motorola, NetCentrex and Sun booths, and MicroTCA platforms were displayed at the Motorola, Artesyn and SBS Technologies booths. SBS also announced its BCT4-AMC1 carrier blade at 3GSM.
The BCT4-AMC1 is an IBM carrier blade that incorporates AdvancedMC system building blocks into IBM's BladeCenterT platform. A live working demo in the IBM booth of a BladeCenterT server using SBS AdvancedMC modules to transfer video over an ATM network proved that AdvancedMCs are real and that the BladeCenterT server is a viable platform for today’s telecommunications infrastructure. Based on the questions I received from the press, and reports from our friends in the IBM booth, there was a tremendous amount of interest for this platform.
Visitors to the SBS booth saw the largest selection of working and available AdvancedMC products on the market. Many visitors said they were pleased to see that our Telum AdvancedMC cards were real and available already. As one visitor to our booth commented, “This is the first time I have seen an actual working AMC – not just a block diagram in PowerPoint”.
Our recently announced Telum 1204, the industry’s first Network Processor-based protocol interworking AMC module, had a lot of people searching for our booth. This was quite amazing for a show of this size and excitement--you wouldn’t expect people to go hunting just to see a little mezzanine card. But the module is a testament to the performance and flexibility of the AdvancedMC standard and it is clear that we will be connecting legacy ATM networks to IP networks for many years to come.
After attending 3GSM, and seeing the excitement gathering around 3G networks based on clear consumer demand and technological and economic drivers, it is very difficult for those of us in the ATCA, BladeCenter and MicroTCA business to not be overly optimistic. These are the open platforms that will host the next generation of telecom applications and bring all those exciting 3G services to the public.
I encourage everyone to visit 3GSM 2007 in Barcelona, if not for the leading edge technology, then for the tapas, sangria and La Rambla--the most famous street in Barcelona. That is if you are not too busy watching movies on your phone.
|
|
| |
 |
 |
By Jeff Marden, Technologist - Communications
Thanks to its close family ties to AdvancedTCA®, MicroTCA was endowed with robust management capabilities right from the start. These capabilities include temperature and voltage monitoring, FRU insertion and removal, and Electronic Keying. MicroTCA adds certain options for management that AdvancedTCA does not offer, which should be exciting news for anyone designing MicroTCA systems. Dedicated system configuration storage on the shelf, direct cooling and power module management are examples of these features.
|
It is fair to say that a designer familiar with IPMI v1.5 or v2.0 will be comfortable with the MicroTCA management structure. The same basic constructs apply, including the ideas of a System Manager, Shelf Manager, Carrier Manager and Module Manager. There are obvious differences considering that MicroTCA doesn’t incorporate carrier cards, but in general the structure holds, particularly if you think of the MicroTCA backplane as a “carrier.”
Management in the MicroTCA world is further explained in the following diagram and description.
The System Manager
The System Manager is highest level of management in a system and is integrated into the MicroTCA system environment in a fashion similar to ATCA; an IP interface and management software running at a remote location. Protocols for communication to the shelves are directly comparable to ATCA.
The Shelf Manager
The MicroTCA Shelf Manager has responsibility for the management of one or more MicroTCA virtual carriers in a shelf, similar to shelf management in ATCA. The main difference, and this is a point which may make MicroTCA attractive at the edge of the network, is that the MicroTCA Shelf Manager can reside in the system’s MicroTCA Carrier Hub (MCH) on an ordinary line card along with the AMC Module Management Controller (MMC), or on another, dedicated card in the shelf. This means that an entire MicroTCA system can consist of only one or two cards and this tiny system size, while it wouldn’t be an advantage for AdvancedTCA, is a feature that should make MicroTCA a powerfully disruptive technology.
The Carrier Manager
For MicroTCA, the Carrier Manager is located on the MicroTCA Carrier Hub (MCH) and includes the MicroTCA Carrier Management Controller (MCMC) and its connections to the AMC modules, Cooling Units , Power Modules and system FRU storage devices on the carrier. The Carrier Manager function is comparable to that of the Carrier IPMC on an ATCA card.
The explosive growth of wireless networks increasingly depends on IP-friendly equipment which also offers something that data networks are not well known for: powerful system management functions, remote diagnostics and high availability characteristics. MicroTCA delivers all these features using the same AdvancedMC modules that were originally designed to be mezzanines in an AdvancedTCA system.
Small Size. Big Capabilities.
MicroTCA appears to be one of those unusual technologies that has come along at the right time with the right features. Much of the action for wireless providers will most likely be happening at the edge of the network with new services like video. And a high availability, high performance, small form factor like MicroTCA is just what these providers are looking for. Of almost equal importance is that MicroTCA is based on pre-existing modules, so the cost for MicroTCA systems is likely to be extremely reasonable.
Although the first MicroTCA systems will likely be designed as full-sized, high capacity telecom units that will reside in air conditioned comfort in existing cabinets in central offices, they may quickly escape the confines of these buildings. Pole-mounted MicroTCA is probably a bit of a stretch at the moment, but other applications seem quite reasonable. For example, this technology may find rapid adoption with the deployment of municipal Wi-Fi and WiMax networks within densely built up urban environments. Management of these complex networks will require remote diagnostics, field replaceable capabilities, affordable equipment, small size and massive bandwidth. That sounds a lot like MicroTCA, doesn’t it?
|
|
 |
|
|
|
