Skip navigation
2011

At last week’s 4G World conference in Chicago, I had a couple of very interesting (and very similar) conversations with Telecom Equipment Manufacturers (TEMs) who visited our booth. In both cases, the architects from these TEMs told me that their next-generation equipment (gateways) would be based exclusively on commodity hardware platforms. And they were referring not just to telecom-oriented ATCA blades, but rather to standard, off-the-shelf server blades and COTS compute platforms. As one of them explained, “We need to think of ourselves as a software company, that’s where our true value lies”.

 

In the case of these two companies, a number of factors have contributed to this strategy:

 

First, standard hardware platforms based on the latest multicore processors are capable of delivering performance at least as high as that achievable on custom hardware, especially when those standard platforms run software such as 6WIND’s in order to maximize networking performance (which is why these architects were visiting our booth).

 

Second, the time-to-market advantages of using standard hardware are overwhelming, and this is especially important in 4G networks where new infrastructure must be deployed on aggressive schedules in order to meet the requirements for exponential growth in network capacity.

 

Third, standard platforms (at least from the right set of suppliers) now deliver the high levels of reliability that are required for “five-nines” or “zero-downtime” telecom equipment.

 

Fourth, the opportunities for TEMs to differentiate core networking products at the hardware level are extremely limited. Their real differentiation is in the upper levels of the system software (layer 7 and above), where their innovations can contribute directly either to providing a better end-user experience or to delivering improved ROI for operators. The open software APIs supported on standard platforms encourage the rapid development of value-added applications within industry-standard software environments.

 

Finally, and maybe most importantly, the use of standard hardware platforms enables TEMs to provide infrastructure that is designed to scale in line with ever-increasing requirements for network capacity and new, advanced services.

 

This last point was emphasized by Rose Schooler, General Manager of Intel’s Communications Infrastructure Division, during a well-attended keynote address at the conference, who referred to a recent statement from Verizon that their next-generation network will be deployed on COTS platforms with open APIs.

 

As Ms. Schooler discussed in her speech, open platforms enable operators to design their networks for the scale required to address future requirements that are not predictable today, whether in terms of capacity, connectivity or threats. She explained the parallels between the transitions happening in the telecom industry today with the evolutions in computing during the ‘90s, when data centers transitioned from multiple proprietary hardware platforms, with closed APIs and divergent form-factors, to a model based on standardized hardware platforms and open APIs. Telecom operators have a great opportunity to leverage successes and learnings from the compute sector, in areas such as power management (OPEX), virtualization (CAPEX), security and others, as the industry moves towards the Software-Defined Network concept.

 

 

Were you at 4G World last week? What were the main trends that you observed at the conference?

As the number of intelligent embedded devices climbs into the billions, designers are refining the next generation of machine-to-machine (M2M) communications techniques to provide continuous, autonomous connectivity and effective remote management. M2M has been called the “next big thing” in embedded design and promises to revolutionize and standardize connectivity across many market segments including industrial, transportation, healthcare, utilities, retail, and consumer electronics. In the past, these markets employed a hodgepodge of communications techniques to interconnect devices and the enterprise however, as the M2M community expands developers are taking the next step and integrating the Internet into a cloud-computing arrangement to not only provide traditional data reporting, but to expand into additional functions such as smart services. For example, with cloud-based M2M communications, industrial device manufacturers can increase revenues by offering a variety of after sale services including remote product updates, failure detection, and even on-site repairs to reduce customer support costs and personnel.

 

To spark the M2M revolution, Intel has joined forces with a number of hardware and software firms to help standardize the connectivity options and build a supply of off-the-shelf products to simplify the transition to cloud computing.  In the remote terminal application area, the Intel® Atom™E6xx architecture is very popular and provides a number of performance improvements to simplify M2M, small-form-factor device designs including a 7-year life cycle commitment. The E6xx series combines the 45 nm processor core plus memory and display controller into one package to reduce the component count and lower overall power requirements. Also, the front side bus used in previous generations has been replaced with a four-lane PCI Express interface giving designers the option of replacing the companion chipset with custom or third-party circuitry to create specialized I/O functions as needed. To deal with the potentially rugged environments found in many M2M applications the Intel® Atom™E6xx series processors are available in the -40 to 85 °C extended temperature range.

 

Several Intel® Embedded Alliance members offer service-ready platforms and start up kits designed specifically for M2M applications. For example, Kontron has developed the M2M Smart Services Developer Kit in conjunction with Intel to enable designers to develop and test an application’s connectivity and performance to shorten the product deployment process (See figure 1). The kit is based on the 1 GHz Intel® Atom™ E640T processor-powered COM Express module that is supported by an M2M-specific base board plus an audio/visual board for applications with a local graphical interface. Built-in wireless networking capabilities include 802.11a/b/g/n Wi-Fi and 802.15.4 wireless personal area network. In addition, cellular connectivity is available as a pre-installed option, or by adding a 3G/4G mini PCI Express miniCard. Storage space for M2M smart service applications, middleware, and an operating system are provided by internal and external microSD cards. With a built-in accelerometer, dual HDMI interfaces, and HD audio support, the M2M Smart Services Developer Kit enables both movement tracking as well as audio/visual intensive smart services applications.

 

Kontron M2M short.jpg

Targeting remote terminal applications in mining, industrial, transportation, and emergency crews, Eurotech recently introduced the Zypad BR2000, a rugged, small form factor wearable and vehicle mountable computer designed for extreme environmental conditions where reliable wired and wireless connectivity are required (See figure 2). The Zypad BR2000 weighs less than 2lbs, operates from 4 to 6 hours in typical applications, and interfaces with a remote display and/or helmet monocle. Based on the 1.3 GHz Intel® Atom™ E660T processor architecture together with high-speed wired and wireless network, the Zypad BR2000 offers WiFi, Bluetooth, and GPS  communications capabilities. Standard I/O includes gigabit Ethernet, USB 2.0, RS232/422, on-board Flash, audio, and 2D/3D video output. Operating system support includes Windows/Windows Embedded Standard 7, Wind River Linux, and several real-time operating systems. The Zypad BR2000 can run Eurotech’s Everyware™ Software Framework designed to enable M2M communications in a variety of environments. The system is also compatible with Eurotech’s Everyware™ Device Cloud with building blocks to provide device-to-cloud data exchange between distributed devices and business applications. You can find out more about the Everyware™ Device Cloud technology in part two of this series.

 

zypad_br2000.jpg

 

Low power operation and high performance are key technologies in the development of the next generation of ubiquitously connected devices. The Intel® Atom™E6xx series architecture provides these features through a flexible I/O architecture that simplifies M2M designs and shortens the time to market. If you are starting or have completed a cloud-based M2M embedded design, please offer your suggestions and share your experience or questions via comments with fellow followers of the Intel® Embedded Community.  You can keep up with the latest technical articles and product announcements at the Embedded Computing Design archives on both M2M communications and connectivity.

 

To view other community content on connectivity, see “Connectivity - Top Picks

connectivity.jpg

Warren Webb
OpenSystems Media®, by special arrangement with Intel® Embedded Alliance

 

Kontron is a Premier member of the by Intel® Embedded Alliance. Eurotech is an Associate member of the Alliance.

At last week’s Linley Tech Processor Conference in San Jose, there was a very interesting discussion about Software-Defined Networking.

 

In a keynote address, Christos Kolias, from the Strategy and Development Group at Orange’s Silicon Valley subsidiary, described the benefits of Software-Defined Networking (SDN), its impact on multicore processor platforms and some specific use cases.

 

As Mr. Kolias explained, existing broadband networks have become rigid, static and inefficient. Over many years, they have evolved to include a wide range of overlays (VPN, MPLS, VoIP etc) and a diverse range of discrete appliances (load-balancers, firewalls, monitoring devices etc). At the same time, routers are overloaded with many features that are used infrequently if at all, while many routing protocols are old and not stable. Finally, system interfaces tend to be closed and proprietary.

 

Mr. Kolias outlined some of the key challenges for future networks (such as video, virtualization, multi-tenancy and security) and stressed the importance for operators of being able to define their own routing and security policies. He pointed to trends in openness (open compute, OpenFlow), hardware commoditization and unified access. The inescapable conclusion is that today’s networks are too difficult to build operate and manage, given these emerging trends.

 

The key principle of SDN is that users can define traffic flows and decide what paths they take in the network. This concept brings a number of important advantages such as: the remote control of network hardware by software in a dynamic, programmatic fashion; the separation of control and data; the use of standard, open hardware interfaces; a full network virtualization plane. Operators benefit from the ability to run new services and applications, from added flexibility and control, and from opportunities for advanced optimization and customization. The concept of a network-wide OS is also supported.

 

The SDN concept is a good fit with multicore processor platforms. Mr. Kolias discussed how embedded multicore processors deliver quicker time-to-market than ASICs, provide transparent acceleration of networking functions through on-chip offload engines and lead to simple, open system architectures.

 

What’s your opinion on SDN? What are the key challenges that need to be addressed to ensure a viable transition from legacy network architectures to an SDN model? What are the implications for suppliers of processors, OSs and networking equipment? Is it reasonable to expect a shift to SDN over the next few years?

With faster Intel® Architecture (IA) processors that integrate four to six cores such as members of the Intel® Xeon™ 5500 and 5600 series, telecom equipment manufacturers have more flexibility than ever in developing powerful application servers using commercial-of-the-shelf (COTS) hardware and software. Such servers provide the equipment manufacturers with a tremendous time-to-market advantage and with a product that they can sell into many different applications due to open and interoperable software platforms. Moreover, the servers allow the network service providers to easily add support for new applications ranging from multimedia streaming to transaction processing in a manner that can scale over time as traffic demands continues to escalate. Using the standard COTS platform also allows the service provider to simplify their supply chain by stocking a single server platform that can be deployed into multiple applications.

 

There are several key technology elements that are intertwined in the concept of COTS carrier-grade servers including ATCA (Advanced Telecommunications Computing Architecture), the IA processors, Intel® Virtualization Technology (Intel® VT), Intel® QuickPath Interconnect (QPI) that can accelerate packet processing, readily-available carrier-grade software platforms, and multiple manufactures of modular server products. Together the elements result in powerful servers that can host software from many vendors and that ultimately lower service provider cost of ownership via code re-use, supply-chain advantages, and synergies with the IT server space that yields the IA processors. Let's discuss these elements individually.

 

ATCA is an open standard that was conceived specifically for telecom space. Systems based on ATCA can achieve five nines reliability – less than five minutes of unscheduled downtime per year. While ATCA products cost more than products based on other modular computing standards, the cost is justified by the application. Moreover, ATCA as an open standard has enabled competition among many suppliers and delivers much lower-cost systems that a closed, proprietary architecture can deliver.

 

If you would like to dig deeper into the ATCA topic, there is quite a lot of good content on the Intel® Embedded Community web site. An article from late 2009 details the ATCA characteristics that enable high-availability systems. ATCA is also proving popular in military & aerospace applications where high-reliability is a requisite. There is an excellent article by my colleague Kenton Williston comparing ATCA with other standards such as VPX and MicroTCA for military & aerospace applications. Back in the communications area, ATCA seems to be the prime choice for service providers in the wireless space where the bulk of the growth in the communications segment is happening, as I covered in a recent article on 40-Gbps network equipment.

 

Of course a carrier-grade hardware platform is only half of the story. Telecom service providers have to have equal reliability on the software side to ensure reliable networks and services. Wind River1 is one software vendor that has focused on the opportunity in the telecom segment. The company offers what it calls carrier-grade versions of Linux and its real-time VxWorks operating systems. Wind River's Carrier Grade Linux 4.0 product complies with the Carrier Grade Linux Specification developed by the Linux Foundation.

 

Wind River has published a whitepaper entitled "Carrier grade open platforms are the key to success for next-generation network elements." You have to register to download the PDF, but the paper includes good background on the concept of next-generation networks, lays a foundation for a platform approach, and describes the Wind River solution.

 

WR_CGOP.jpg

 

The above figure summarizes the elements of the open platform. As you can see, the operating system is a small part of the overall platform. The carrier grade elements depicted on the right of the figure are applied across all of the layers of the platform that are stacked on the left side of the figure. My colleague Henry Davis wrote a dedicated article on the topic of multi-core processors and next-generation networks that covers this area in more detail.

 

The Wind River carrier-grade operating systems include support for a number of features that are critical in enabling network applications on IA processors with optimal performance. For example, the operating systems support both traditional symmetrical multiprocessing (SMP) and asymmetric multiprocessing. In the SMP case, tasks are spread across cores in a processor scheduled by the operating system. In AMP, cores can be dedicated to specific tasks allowing more resources to be dedicated to real-time packet or stream processing. I covered a Wind River evaluation of packet processing performance in an earlier article in which AMT delivered a significant performance advantage. The implementation relies on Intel VT technology to partition the real-time tasks.

 

Wind River also offers a Network Packet Acceleration stack for use with its standard and carrier-grade operating systems that utilizes AMP. The link I provided above addresses the acceleration product and my colleague Warren Webb covered it recently as well in an article on breaking network performance barriers.

 

The Wind River operating systems and the Network Acceleration Platform are compatible with ATCA products from a number of vendors including Advantech2, RadiSys3 (Continuous Computing), Emerson Network Power Embedded Computing4, and Kontron5.

 

Kontron, for example, offers the AT8050 ATCA server blade with a choice of a 6-core Xeon 5600 processor or a 4-core 5500 processor. In either case the blade can host as much as 48 Gbytes of memory, although in high-reliability applications system designers may turn to registered ECC memory and the blade can integrate only 24 Gbytes of the ECC memory. There is also an Advanced Mezzanine Card (AMC) slot that among other potential uses could add an additional IA processor.

 

Advantech_MIC-5322.jpg

 

Advantech also offers ATCA blades based on the Xeon 5500 or 5600 series, and in fact the MIC-5322 pictured above integrates dual processors. In the case of the 5600 version, the blade includes 12 cores that can process 24 execution threads. The blade also features the Intel® 82599 Gbit Ethernet controller and the Intel® 5520 server-class chipset.

 

The readily-available ATCA hardware and software such as Wind River's carrier class products allow a telecom equipment maker to easily and quickly bring high-availability servers to market. Moreover the systems allow interoperable support for software from Wind River and third parties.

 

Do you have experience integrating ATCA-based technology into high-availability systems? Other followers of the Intel® Embedded Community would welcome the chance to benefit from your experience via comments. Have you used carrier-grade software in a system? What were the roadblocks you faced and tell us about the benefits that you achieved in the project.

 

To view other community content focused on interoperability, see "Interoperability – Top Picks."

interoperability.jpg

 

Maury Wright

Roving Reporter (Intel Contractor)

Intel® Embedded Alliance

 

1 Wind River is an Associate member of the Intel® Embedded Alliance

2 Advantech is a Premier member of the Alliance

3 Radisys is a Premier member of the Alliance

4 Emerson Network Power Embedded Computing is a Premier member of the Alliance

5 Kontron is a Premier member of the Alliance

Cloud computing isn’t just for the enterprise – telecom equipment manufacturers (TEMS) and service providers can also benefit from cloud computing. To learn how use can use this technology – and how to overcome the attendant challenges – check out the latest Embedded Innovator webinar from Emerson and RadiSys.  The webinar explains:

 

  • Benefits and use cases for cloud computing
  • Fundamentals of Intel® Virtualization Technology (Intel® VT)
  • Design options using AdvancedTCA* (ATCA)

 

You can access more articles like this by subscribing to the Embedded Innovator.  Subscribers receive a quarterly newsletter and an annual magazine that bring you the latest industry trends and design ideas.  Get your subscription today!

 

workload_consolidation.pngFor more on building flexible networking solutions, see intel.com/go/embedded-consolidation

 

Emerson Network Power and RadiSys are Premier Members of the Intel® Embedded Alliance

 

 

Kenton Williston

Roving Reporter (Intel Contractor), Intel® Embedded Alliance

Editor-In-Chief, Embedded Innovator magazine

 

Follow me on Twitter: @kentonwilliston

By Charlie Ashton, VP of Marketing for 6WIND

 

 

Last week I attended Broadband World Forum in Paris (where 6WIND, Qosmos, RadiSys and Tilera exhibited in the Network Intelligence Alliance booth).

 

It was fascinating to talk with a range of people from both service providers and equipment manufacturers. A couple of topics from those conversations may be of interest to readers of this blog:

 

  • The majority of our discussions with attendees were not about network performance and capacity issues, but rather about the challenge of maximizing network infrastructure Return on Investment (ROI). Service providers now view this as a key area of differentiation, seeing enormous competitive advantage in how they exploit real-time knowledge about the traffic characteristics and demand in their network, segmented by application, by user and by time-of-day. As discussed in this blog last week, this information is key to the delivery of customized services that provide high value to specific customers. In addition, real-time application-level analysis of network traffic, including advanced heuristics or pattern recognition, allows operators to defend their networks against ever more sophisticated security threats and attacks.

 

  • Because I was at our booth most of the time, I missed the keynote presentations. Apparently, though, one very interesting presentation was given by Aparna Khurjekar, executive director of business solutions group at Verizon. Discussing the market opportunity presented by “connected cars”, Khurjekar discussed a wide range of in-car connectivity services that Verizon could deliver as part of extending their reach beyond cellular. These services could include the creation of “intelligent transportation” that could help to dramatically reduce the rate of road crashes, which currently occur every five seconds in the US, with a fatality every five minutes. 25% of driving time in the US is spent in traffic and connected technology could help to reduce congestion by 70-80%, helping to improve the economy through increased productivity time. Other examples of connected in-car uses would be Vehicle-to-Device, with embedded LTE enabling new forms of in-car entertainment such as Netflix streamed directly to screens in the back seats.

 

Were you at Broadband World Forum last week? What were your main impressions from the event? What were some of the highlights that I missed because I was “booth-bound”?

When we use our ubiquitous smartphone, most of us aren't individually straining data networks. The problem is that there are a lot of us that are taking advantage of faster wireless services. Indeed the move to 4G wireless network technology including WiMax and LTE will drive the industry to deploy 40-Gbps-capable switches and packet-processing systems. The first systems and blades for the 40-Gbps transition have come to market this year and last, and the telecom service providers have stated a preference for systems based on standard technologies such as ATCA (Advanced Telecommunications Computing Architecture) and broadly available ICs. Let's have a look at the requirements for such systems and how Intel® Architecture processors might be deployed in such systems starting with system management today and data-plane applications down the road.

 

The fact is that bandwidth has been on a steady rise for years with an increase in services deployed over IP networks. Music and photo sharing have been followed by video consumption. And now we even consume such content over wireless networks and use our smartphones as impromptu Wi-Fi hot spots to stream multimedia data to our notebooks and tablets.

 

Emerson Network Power Embedded Computing* has developed an excellent downloadable ebook on the topic entitled "Get ready for 40G ATCA." Emerson notes that what began as video chats, web cam access, and YouTube viewing on smartphones has evolved to consumers viewing feature-length content.

And it's not just consumers on smartphones driving bandwidth requirements. Machine-to-machine (M2M) communications will account for an increasing share of the bandwidth load going forward. You have automotive systems sharing routing information in real time with some cars having multiple wireless links. The industrial, commercial, and transportation markets are consuming bandwidth tracking goods, managing buildings, and deploying wireless sensors. Many of the applications don't consume much bandwidth individually. But there are going to be far more embedded connections to the internet and M2M links in the future than there will be smartphones, tablets or PCs.

 

Today we are in the midst of the 4G-mobile-network deployment stage with per-subscriber maximum data rates going to 100 Mbps. Emerson summarizes the trend in the chart below. As you can see, we are only a year or two from the first deployment of networks that will support 1-Gbps subscriber downlink speeds. The message is that even if you aren't planning on supporting 40-Gbps speeds in a system design today, design your system in a way that it can evolve to the faster speeds.

 

emerson_4G_table.jpg

 

Emerson notes that IC technologies capable of 40-Gbps switching and packet-processing are emerging. Specialty ICs with tens of communications-centric cores will handle some of the real-time tasks, especially early in the 40-Gbps era. But expect IA processors to play a big role as well just as they have evolved into usage beyond the control plane in prior-generation systems. The Emerson document explains that you can deploy dual 6-core Intel® Xeon® processors on a single ATCA blade today with that platform able to handle 10-Gbps rates. You can fully expect next-generation IA processors to include architectural enhancements and additional cores that will enable usage in 40-Gbps data planes.

 

The ATCA standard has also evolved in preparation for the 40-Gbps transition. The PICMG (PCI Industrial Computer Manufacturers Group) added support for 40-Gbps transmission on ATCA backplanes in PICMG 3.1 Option 9. The standard allows for 4 10-Gbps links or a single 40-Gbps link.

 

Let's have a look at how Emerson is supporting 40-Gbps designs in its Centellis 4440 ATCA platform. The backplane and thermal design is fully capable of 40-Gbps deployment even if an initial deployment only supports 10-Gbps speeds. The company offers the ATCA-9405 packet-processing blade and the ATCA-F140 switch blade – both capable of 40-Gbps speeds.

 

While the ATCA-F140 includes dedicated 40-Gbps switching, it also offers an AMC (Advanced Mezzanine Card) site that can be populated with numerous IA-based mezzanine cards. The IA processor can handle control-plane tasks. Moreover, the IA processors support Intel® Active Management Technology (AMT). AMT is one of several IA technologies that support the mission-critical reliability required in applications such as telecom systems, and also allows for remote management of a system.

 

Emerson also offers numerous server blades based on IA processors that can be deployed in a Centellis-based system. For example, the ATCA-7367 integrates a 6-core Intel Xeon L5638 processor and can host a second IA processor via an AMC site. The ATCA-7365 integrates dual L5638 processors.

 

Emerson isn't alone in targeting the 40-Gbps transition. For example, Radisys**, Kontron***, and Advantech*** have all introduced some level of 40-Gbps systems and modular products.

 

Let's consider the Radisys offering, part of which came courtesy of the company's acquisition of Continuous Computing. The company's FlexTCA 40G Platform (pictured) starts with a 40-Gbps backplane and chassis and includes numerous configuration options via blades. The offering includes both switch and packet-processing blades that operate at full line speed.

 

Radisys_40G.jpg

 

The platform also supports IA-based computing blades. For example, the FlexCompute ATCA-XE80 is available in single- and dual-processor versions based on the 6-core Intel Xeon L5645 processor.

 

Radisys has posted a number of other information resources you might find interesting when considering the 40-Gbps future. For example, there are also white papers entitled "40G: Is this race necessary?" and "High-performance ATCA: 80 Gbps."

 

There are also some other 40-Gbps centric posts from my Roving Reporter colleagues that you might review:

 

How do you prepare for major transitions in technology such as the migration to 40-Gbps networks? How are you utilizing processors such as the 6-core Intel Xeon processors mentioned in this post? And how do you implement remote management for mission-critical systems. Please share you experiences with fellow followers of the Intel® Embedded Community via comments.

 

To view other community content focused on manageability, see "Manageability – Top Picks."

manageability.jpg

 

Maury Wright

Roving Reporter (Intel Contractor)

Intel® Embedded Alliance

 

* Emerson Network Power Embedded Computing is a Premier member of the Intel® Embedded Alliance

**Radisys is a Premier member of the Alliance

***Kontron is a Premier member of the Alliance

****Advantech is a Premier member of the Alliance

Filter Blog

By date: By tag: