OFC/NFOEC 2010: Announcements round-up
The Infinera Express: Infinera's 80-foot-long truck-based mobile demo unit, is at OFC/NFOEC. Infinera is part of a demo of live 100 GigE data traffic with Juniper Networks, Finisar, Opnext and Reflex Photonics. The truck also contains Infinera’s ATN metro edge platform.
Demonstrations and displays
The Optical Internetworking Forum (OIF) is displaying components and hardware as part of its integrated transmitter and receiver initiative for 100Gbps transponders. OIF member companies taking part include Fujitsu Optical Components, NEC, NeoPhotonics, Opnext, Picometrix, Sumitomo Osaka Cement, TriQuint Semiconductor, u2t Photonics and Vitesse Semiconductor.
Opnext is demonstrating a real-time 100Gbps DWDM link using a single wavelength coherent receiver. This follows Opnext’s recent participation in AT&T’s 100Gbps trial. Opnext has already detailed its 100Gbps silicon germanium multiplexer IC and last week it announced it had partnered with A/D converter (ADC) specialist Mobius Semiconductor to develop a CMOS-based polarisation multiplexing quadrature phase-shift keying (PM-QPSK) receiver chip. The integrated circuit (IC) includes ADCs, a digital signal processor and forward error correction (FEC).
40 and 100Gbps line side transmission
Until recently, Nortel (acquired by Ciena) was the sole vendor with 40Gbps coherent technology. Last week Fujitsu announced that it has added 40Gbps coherent technology to its Flashwave 7500 platform. Now at OFC, CoreOptics has announced a 40Gbps coherent module and is demonstrating the technology with a Nokia Siemens Networks' system. CoreOptics claims its 40Gbps 300-pin MSA delivers a 2,000km reach without requiring dispersion compensation modules.
NeoPhotonics has announced integrated coherent receivers for 40 and 100Gbps. The receiver combines an integrated dual 90° hybrid coherent mixer with four balanced photodiodes and linear amplifiers in a package (see photonic integration feature).
There are also more 40Gbps DQPSK products being announced at OFC.
Opnext is showing its 40Gbps DQPSK 7x5inch transponder while u2t Photonics has unveiled its integrated DQPSK receiver which is now sampling. The company claims that integrating two balanced receivers in a single package saves up to 70% board space.
Oclaro last week announced that its 40Gbps DQPSK transponder had been qualified to Telcordia standards.
Lastly, Infinera has announced it has recruited John McNicol, a senior engineer involved in the development of Nortel’s coherent technology. “He will be critically involved in the development of Infinera’s next-generation optical networking systems,” says the firm, indicating Infinera’s intention to development a coherent-based system.
40 and 100 Gigabit Ethernet (GbE)
Transceiver firm ColorChip can claim a first by implementing the 40GBASE-LR4 standard in a QSFP module. The industry is working to decrease the size of the 40GbE modules. The first implementations use the CFP module while system vendors want smaller modules to increase the interfaces that can be placed on a card.
40GBASE-LR4 CFPs were first announced six months ago around ECOC 2009 and since then Hitachi Cable has developed a 40GBASE-LR4 X2 module. Now ColorChip has used its system-on-glass technology to squeeze the design into a QSFP. ColorChip has said it will deliver samples later this year with volume manufacturing beginning in 2011.
There is also an interoperability demonstration of 100GbE CFP modules involving Infinera, Juniper, Finisar and Opnext. The demonstration includes Juniper’s T1600 core router and Infinera’s DTN optical system and 100GBASE-LR4 modules from Finisar and Opnext. Reflex Photonics’ 100GBASE-SR10 CFP modules are also used as part of the demonstration.
Passive Optical Networks
NeoPhotonics has announced 10G PON transceivers for the GPON and EPON standards. The GPON pluggable transceivers supports 10Gbps in the downstream and 2.5Gbps burst mode transmission in the upstream with link budgets of 29 dB and 31 dB. The EPON transceivers support 10Gbps in the downstream and 1Gbps in the upstream direction and have a link budget up to 30.5 dB.
The GPON and EPON optical line terminal (OLT) transceivers used at the central office are implemented using the XFP while the optical networking unit (ONU) transceivers at a PON's end points are implemented using an SFP+ form factor.
Parallel optics
Avago Technologies has announced a miniature 12-channel MicroPOD parallel optics transmitter and receiver modules that it is promoting as a follow-on to SNAP 12. The modules, measuring 7.8mm (L) by 8.2 mm (W) by 3.9mm (H), support lane rates of up to 12.5 Gbit per second for an aggregate bandwidth of 150Gbps.
Avago claims the MicroPOD modules are compliant with the IBTA 12xQDR Infiniband and IEEE 802.3ba 100GBASE-SR10 specifications. Avago also announced it has developed with IBM new miniature low power 120Gbps 12-channel modules for IBM’s upcoming POWER7 supercomputing systems.
Luxtera is using OFC to demonstrate its OptoPHY board-mountable optical transceivers. The company’s CMOS-based optical engine is being shown supporting 40GbE.
Reflex Photonics has launched two surface-mount 12-channel optical engines. The LightABLE devices take up 2.3 cm2 of board space and consume 42 mW of power per channel, transmitting or receiving at 120Gbps.
OFC tweets
To follow OFC/NFOEC on Twitter search on #ofcnfoec.
You can also follow analysts Andrew Schmitt of Infonetics (@aschmitt), Eve Griliches of ACG Research (@EveGr) and journalists Craig Matsumoto of Light Reading (@craigmatsumoto) and Stephen Hardy of Lightwave (@lightwaveonline).
Stephen Hardy also has an OFC show blog
Verizon plans coherent-optimised routes
"Next-gen lines will be coherent only"
Glenn Wellbrock, Verizon Business
Muxponders at 40Gbps
Given the expense of OC-768 very short reach transponders, Verizon is a keen proponent of 4x10Gbps muxponders. Instead of using the OC-768 client side interface, Verizon uses 4x10Gbps pluggables which are multiplexed into the 40Gbps line-side interface. The muxponder approach is even more attractive with compared to 40Gbps IP core router interfaces which are considerable more expensive than 4x10Gbps pluggables.
DQPSK will be deployed this year
Verizon has been selective in its use of differential phase-shift keying (DPSK) based 40Gbps transmission within its network. It must measure the polarisation mode dispersion (PMD) on a proposed 40Gbps route and its variable nature means that impairment issues can arise over time. For this reason Verizon favours differential quadrature phase-shift keying (DQPSK) modulation.
According to Wellbrock, DPSK has a typical PMD tolerance of 4 ps while DQPSK is closer to 8 ps. In contrast, 10Gbps DWDM systems have around 12 ps. “That [8 ps of DQPSK] is the right ballpark figure,” he says, pointing out that a measuring a route's PMD must still be done.
Verizon is testing the technology in its labs and Wellbrock says Verizon will deploy 40Gbps DQPSK technology this year.
Cost of 100Gbps
Verizon Business has already deployed Nortel’s 100Gbps dual- polarization quadrature phase-shift keying (DP-QSPK) coherent system in Europe, connecting Frankfurt and Paris. However, given 100Gbps is at the very early stages of development it will take time to meet the goal of costing 2x 40Gbps.
That said, Verizon expects at least one other system vendor to have a 100Gbps system available for deployment this year. And around mid-2011, at least three 300-pin module makers will likely have products. It will be the advent of 100Gbps modules and the additional 100Gbps systems they will enable that will reduce the price of 100Gbps. This has already happened with 40Gbps line side transponders; with 100Gbps the advent of 300-pin MSAs will happen far much quickly, says Wellbrock.
Next-gen routes coherent only
When Verizon starts deploying its next-generation fibre routes they will be optimised for 100Gbps coherent systems. This means that there will be no dispersion compensation fibre used on the links, depending on the 100Gbps receiver’s electronics to execute the dispersion compensation instead.
The routes will accommodate 40Gbps transmission but only if the systems use coherent detection. Moreover, much care will be needed in how these links are architected since they will need to comply with future higher-speed optical transmission schemes.
Verizon expects to start such routes in 2011 and “certainly” in 2012.
OFC/NFOEC 2010: Technical paper highlights
Here is a sample of some of the noteworthy papers.
Optical transmission
Nortel’s Next Generation Transmission Fiber for Coherent Systems details how various fibre parameters impact coherent system performance. This is important for existing 40 and 100Gbps systems and for future ones based on even higher data rates.
In 40G and 100G Deployment on 10G Infrastructure: Market Overview and Trends, Coherent Versus Conventional Technology, Alcatel-Lucent discusses 40G and 100G deployment strategies over 10G infrastructures based on a trial using live commercial traffic.
Two papers demonstrate possible future optical modulation steps.
In Ultra-High Spectral Efficiency Transmission, Bell Labs Alcatel-Lucent details the generation, transmission and coherent detection of 14-Gbaud polarization-division multiplexed, 16-ary quadrature-amplitude-modulation (16-QAM) signals achieving spectral efficiencies as high as 6.2 b/s/Hz.
Meanwhile, NEC Labs America and AT&T Labs address 112.8-Gb/s PM-RZ-64QAM Optical Signal Generation and Transmission on a 12.5GHz WDM Grid. The optical signal was sent over 2x40km using an 8-channel WDM using 12.5GHz grid spacing.
Photonic integration
In High Performance Photonic Integrated Circuits for Coherent Fiber Communication, Chris Doerr of Bell Labs, Alcatel-Lucent presents how photonic integration can benefit high-speed transmission. In particular, how optical integration can be used to tackle the complex circuitry needed for coherent systems to reduce the area, cost, and power consumption of optical coherent transceivers.
Another photonic integration development is the CMOS-Integrated Low-Noise Germanium Waveguide Avalanche Photodetector Operating at 40Gbps from IBM T.J. Watson Research Center. The avalanche photodiode has a gain-bandwidth-product above 350GHz operating at 3V. The avalanche photodetector is monolithically integrated into CMOS.
Optical access
An update will be given on the EU’s Seventh Framework programme for WDM-PON, dubbed Sardana - Scalable Advanced Ring-based passive Dense Access Network Architecture. The paper, Results from EU Project SARDANA on 10G Extended Reach WDM PONs, details the integration of WDM metro and PON access technologies to implement ring protection, 100km reach and up to 1024 users served at 10Gbps using a passive infrastructure.
In 44-Gb/s/λ Upstream OFDMA-PON Transmission with Polarization-Insensitive Source-Free ONUs, NEC Labs America details its work on colourless 44-Gb/s/λ upstream OFDMA-PON transmission using polarization-insensitive, source-free ONUs.
Green telecom and datacom
There are other, more subtle developments at OFC/NFOEC. Two papers from Japan have ‘Green’ in the title, highlighting how power consumption is increasingly a concern.
High Performance “Green” VCSELs for Data Centers from Furukawa Electric Co. Ltd details how careful design can achieve a 62% power conversion efficiency in the 1060nm VCSEL.
The second paper tackles power consumption in access networks. Key Roles of Green Technology for Access Network Systems from NTT Labs in Japan addresses the ITU-T’s standardisation activities. Optics for flow and interconnect
In Optical Flow Switching, Vincent Chan of MIT will discuss 'optical flow switching' that promises significant growth, power-efficiency and cost-effective scalability of next-generation networks.
Meanwhile Bell Labs, Lucent Technologies has a paper entitled Photonic Terabit Routers: The IRIS Project, detailing the results of the DARPA-MTO funded program to develop a router with an all-optical data plane and a total capacity of more than 100 Tbps.
Another important topic is optical interconnect. Low Power and High Density Optical Interconnects for Future Supercomputers from IBM Research reviews the status and prospects of technologies required to build low power, high density board and chip level interconnects needed to meet future supercomputers requirements.
NFOEC papers
There are also some noteworthy NFOEC papers bound to stir interest:
- Google reviews the optical communication technologies required to support data center operations and warehouse-scale computing.
- Verizon shares lessons learned during the five years of Verizon’s FiOS and the need to continually evolve product and service offerings.
- AT&T details the key decisions required in defining its new 100G backbone.
There is a comprehensive OFC/NFOEC preview in the February issue of IEEE Communications magazine, click on the "conference preview" tab.
Framing the information age
When writing features for FibreSystems Europe, I repeatedly asked for high-resolution striking images. The magazine's editors always wanted photos that included people, like Maurice Broomfield's photos. Getting hold of such images did happen but not often.
Inspired by the Financial Times’ interview and Maurice Broomfield's beautiful images, some of the better images sent are presented here.
IBM data centre
I’m on the look-out for more. So if you are the media relations for an operator, equipment maker, optical transceiver or component (optical or IC) vendor, can I please request some inspiring photos - ideally with people - and I'll create a photo gallery of the best.
Network Operations Centre (NOC) Source: AT&T
Source: Cisco Systems
An Intel silicon photonics device
And here is an image of Tokyo's data centre on Flickr
UNIC silicon modulator
This is the silicon photonic start-up’s first announced modulator. The design has been developed in conjunction with Sun Microsystems as part of the DAPRA Ultraperformance Nanophotonic Intrachip Communications (UNIC) programme.
An image of the modulator and a cross-section diagram of the ring waveguide. Source: Kotura
Why is it important?
Optical components use a range of specialist, expensive materials. Silicon is one material that could transform the economics for optics. But for this to happen, the main optical functions – light generation, transmission and detection – need to be supported in silicon. To date, all the required functions except the laser itself - waveguides, modulators and photo-detectors - have been mastered and implemented in silicon.
However, the use of silicon photonics in commercial products has till now been limited. For example, Luxtera makes active optical cable that uses silicon photonics-based transceivers while Kotura has been producing silicon photonics-based VOAs for several years. Its VOA is used within reconfigurable optical add/drop multiplexers (ROADMs) and as a dimmer switch to protect optical receivers from network transients.
Kotura is also supplying its silicon-based Echelle gratings product for 40 and 100 Gigabit Ethernet (GbE) transceiver designs that require the multiplexing and demultiplexing of 4 and 10 wavelengths. The company’s gratings are also being used in Santur’s 100Gbit/s (10x10Gbit/s) transceiver design.
Kotura is in volume production of its VOAs and sampling its Ethernet gratings products, says Arlon Martin, vice-president of marketing and sales at Kotura: “The biggest interest is in 40 Gigabit Ethernet.” Given the small size of the gratings, Kotura is also seeing interest from vendors developing 40GbE transceivers in smaller form factors than the CFP module, such as the QSFP.
This will enable 1Tbit/s data rates over a single fibre to connect high-speed multi-core processor computing elements.
Arlon Martin of Kotura.
But the true potential for silicon photonics, one that promises huge volumes, is very short reach optical interconnects for use in high performance computing and within data centres. Having a low power silicon modulator means it can be integrated with other circuitry in CMOS rather than as a discrete design. Such an integrated approach ensures interconnect reliability.
Method used
There are several ways to modulate a laser. Direct modulation uses electronics to switch the laser on and off at the required rate to imprint the data onto the light. An electro-absorption modulated laser, in contrast, adds an element in front of an always-on laser that either passes or absorbs the light. Kotura’s modulator uses a third approach based on a micro-ring resonator and an adjacent waveguide.
The dimension of the ring – its circumference – dictates when optical resonance occurs. And by carefully matching the power coupling of the micro-ring and waveguide to that of the ring loss, signal attenuation– the light-off condition – is improved. The wavelength at which resonance occurs can be changed by playing with the optical properties of the ring waveguide.
Kotura and Sun have demonstrated the silicon modulator working at up to 11GHz, requiring a peak-to-peak voltage of 2V only. The modulator’s insertion loss is also an attractive 2dB though its working spectrum width is only 0.1nm.
“Our power number – 0.5mW at 10GHz - does not include the driver. But if you want to integrate a number of these on one chip, the low power consumption would enable this,” says Martin. Kotura claims the power consumption achieved is the lowest yet reported.
What next?
The modulator is one of the milestones of the DARPA UNIC programme now into the second of its five-year duration. “This [modulator] is prototype work, not a product,” says Martin, adding that Kotura has not fixed a date as to when the modulator will be commercially used.
As for how the device will ultimately be used, Kotura talks of interfaces operating between 100Gbit/s and 1 Tbit/s. Kotura is already working on an independent programme with CyOptics - the NIST Advanced Technology Programme - developing up to 1Tbit/s links using wavelength division multiplexing (WDM). Such designs use separate laser arrays - each laser at a specific wavelength – as well as gratings and photo-detectors.
In the future inexpensive light sources could generate up to 80 separate modulated lightpaths, Martin says. This will enable 1Tbit/s data rates over a single fibre to connect high-speed multi-core processor computing elements.
Is the idea similar to a broadband light source as proposed for WDM-PON? The UNIC partners have yet to reveal the programme’s detail. “Potentially on the right path,” is all Martin would say.
References:
[1] “Low Vpp, ultralow-energy, compact, high-speed silicon electro-optic modulator.” To read Kotura’s technical paper, click here.
[2] "PHOTONICS APPLIED: INTEGRATED PHOTONICS: Can optical integration solve the computational bottleneck?" OptoIQ, March 1, 2009, click here.
Service providers' network planning in need of an overhaul
These are the findings of an operator study conducted by Analysys Mason on behalf of Amdocs, the business and operational support systems (BSS/ OSS) vendor.
Columns (left to right): 1) Stove-pipe solutions and legacy systems with no time-lined consolidated view 2) Too much time spent on manual processes 3) Too much time (or too little time) and investment on integration efforts with different OSS 4) Lack of consistent processes or tools to roll-out same resources/ technologies 5) Competition difficulties 6) Delays in launching new services. Source: Analysys MasonClick here to view full chart.
What is network planning?
Every service provider has a network planning organisation, connected to engineering but a separate unit. According to Mark Mortensen, senior analyst at Analysys Mason and co-author of the study, the unit typically numbers fewer than 100 staff although BT’s, for example, has 600.
"They are highly technical; you will have a ROADM specialist, radio frequency experts, someone knowledgeable on Juniper and Cisco routers," says Mortensen. "Their job is to figure out how to augment the network using the available budget."
In particular, the unit's tasks include strategic planning, doing ‘what-if’ analyses two years ahead to assess likely demand on the network. Technical planning, meanwhile, includes assessing what needs to be bought in the coming year assuming the budget comes in.
The network planners must also address immediate issues such as when an operator wins a contract and must connect an enterprise’s facilities in locations where the operator has no network presence.
“What operators did in two years of planning five years ago they are now doing in a quarter.”
Mark Mortensen, Analysys Mason.
Network planning issues
- Operators have less time to plan. “What operators did in two years of planning five years ago they are now doing in a quarter,” says Mortensen. “BT wants to be able to run a new plan overnight.”
- Automated and sophisticated planning tools do not exist. The small size of the network planning group has meant OSS vendors’ attention has been focused elsewhere.
- If operators could plan forward orders and traffic with greater confidence, they could reduce the amount of extra-capacity they currently have in place. This, according to Mortensen, could save operators 5% of their capital budget.
Key study findings
- Changes in budgets and networks are happening faster than ever before.
- Network planning is becoming more complex requiring the processing of many data inputs. These include how fast network resources are being consumed, by what services and how quickly the services are growing.
- As a result network planning takes longer than the very changes it needs to accommodate. “It [network planning] is a very manual process,” says Mortensen.
- Marketing people now control the budgets. This makes the network planners’ task more complex and requires interaction between the two groups. “This is not a known art and requires compromise,” he says. Mortensen admits that he was surprised by the degree to which the marketing people now control budgets.
In summary
Even if OSS vendors develop sophisticated network planning tools, it is unlikely that end users will notice a difference, says Mortensen. However, it will impact significantly operators’ efficiencies and competitiveness.
Users will also not be as frustrated when new service are launched, such as the poor network performance that resulted due to the huge increases in data generated by the introduction of the latest smartphones. This change may not be evident to users but will be welcome nonetheless.
Study details
Analysys Mason interviewed 24 operators including (40%) mobile, (50%) fixed and (10%) cable. A dozen were Tier One operators while two were Tier Three. The rest - Tier Two operators - are classed as having yearly revenues ranging from US$1bn and 10bn. Lastly, half the operators surveyed were European while the rest were split between Asia Pacific and North America. One Latin American operator was also included.
Optical core switching tops 4 Terabit-per-second.
Event:
Alcatel-Lucent has launched its 1870 Transport Tera Switch (TTS) that has a switch capacity of 4 Terabits-per-second (Tbps). The platform switches and grooms traffic at 1Gbps granularity while supporting lightpaths up to 100Gbps.
“It is designed to address the explosion of traffic in core networks, driven by video and the move to cloud computing among others,” says Alberto Valsecchi, vice president of marketing, optics activities at Alcatel-Lucent.
The 1870 TTS supports next-generation Optical Transport Network (OTN), carrier Ethernet and SONET/SDH protocols, as well as generalized multiprotocol label switching/ automatically switched optical network(GMPLS/ ASON) control plane technology to enable network management and traffic off-load between the IP core and optical layers.
"
It [the 1870 TTS] is designed to address the explosion of traffic in core networks"
Alberto Valsecchi, Alcatel-Lucent
Central to the 1870 TTS is an in-house-designed 1Tbps switch integrated circuit (IC). The switch chip is non-blocking and by switching at the OTN level supports all traffic types. The device is designed to limit power consumption and is claimed to consume 0.04 Watts per Gbps. Four such ICs are required to achieve the 4Tbps switch capacity.
Each platform line card has a 120Gbps capacity and supports 1, 2.5, 10 and 40Gbps interfaces with a 100Gbps interface planned. The line card’s optical transceiver interfaces include 12 XFPs or two CFP modules. Three 40Gbps interfaces will be supported in future and a 240Gbps line card is already being mentioned (Alcatel-Lucent describes the platform as ‘8Tbps hardware ready’).
The cards also use tunable XFP modules. The 1870 TTS can thus be used alongside existing optical platforms for long-haul dense wavelength division multiplexing (DWDM) transport or support its own links. “There is urgency for this [platform] to manage bandwidth in the central office,” says Michael Sedlick, head of cross-connect product line management at Alcatel-Lucent. “It can solve both requirements: working with installed based WDM platforms and enabling a more integrated implementation [using tunable XFPs].”
“It is already in trials and is selected by tier one service providers,” says Valsecchi.
Why is it important:
Gazettabyte asked three experts to give their views on the announcement. In particular, to position the 1870 TTS platform, discuss its significance, and the importance of GMPLS/ ASON support.
Ron Kline, principal analyst, network infrastructure at Ovum
The platform announcement is essentially about solving three key [service provider] dilemmas: scaling to meet the huge growth in traffic, working through the transition from SONET/SDH to Ethernet and driving efficiencies in the network through automation that helps reduce capital expenditure and operational expenditure.
The 1870 TTS is not a new class of platform but rather the next generation of bandwidth management systems that is using electrical OTN switching rather than STS-1 [SONET frame] switching.
Alcatel-Lucent’s existing 1850 packet optical transport system is really the next generation of aggregation (optical edge device/ multi-service provisioning platform) equipment. The difference is where the device goes in the network and the granularity of the switching.
For the 1870, you are switching bandwidth at wavelength rates (2.5G, 10G, 40G, etc.) There is also some sub-wavelength granularity as well. The device is protocol independent because client signals (SONET/SDH, Ethernet, video, etc) are all encapsulated in the OTN wrapper.
The most similar platforms are the Ciena 5400 introduced in September ’09 and Huawei’s OSN 8800. Tellabs also introduced a high-speed shelf for the 7100 that has a 1.2Tbps OTN matrix and ZTE has the ZXONE 8600 that it introduced in March ’09.
Momentum has been building for several years now. The current generation of optical core switches (Ciena's CoreDirector, Alcatel-Lucent's 1678 MCC, the Sycamore 16000) cannot scale large enough and are SONET/SDH based. The need is to be able to groom at the wavelength level. Current switch sizes (640Gbps, 1.2 Tbps for Sycamore) can’t scale so you have to place another switch and also use capacity to tie the switches together. The larger you grow the bigger the problem—you use 10% of capacity to tie two machines together, 20% to tie 3 together, etc.
In addition, older generation optical core switches are SONET/SDH-based and have trouble with Ethernet so they have to use the generic framing procedure/ virtual concatenation (GFP/VCAT) to manipulate the signal. When you move to OTN switching you don’t have to convert between protocols to switch through the matrix.
And yes, so far 4Tbps is the highest switch capacity per chassis.
As for the control plane, ASON automates configuration so it is more applicable for turning up and down bandwidth. In Alcatel-Lucent’s case, it is integrating the control planes across its product portfolio which gives visibility across the entire network. Although router offload is a key application for the device, you don’t necessarily need a control plane to do it.
IP routing is much more expensive (per bit) then wavelength switching. The idea is to switch at the lowest network layer possible. People have been using an inverted triangle with 4 layers to illustrate. IP routing is at the top followed by layer-2 switching, TDM/OTN switching and then wavelength switching.
Eve Griliches, managing partner, ACG Research
The 1870 TTS isn’t a new platform class but the platforms in general are new. Huawei, Fujitsu, Tellabs, Cisco and various others have platforms all geared towards this but most are missing some element today. In this case, Alacatel-Lucent is still missing the optical portion of the product [DWDM and ROADM].
The platform is starting out as a large OTN and packet switch which will eventually turn into a full packet optical transport product - that is my estimate. Each vendor is approaching this area differently, suffice to say, I think this is an OK and decent approach.
The GMPLS/ ASON control plane technology means the 1870 TTS can manage the optical and IP layers together. Some providers want that, some don't.
Andrew Schmitt, directing analyst, optical, Infonetics Research
The 1870 TTS is a lot like the 1850, just bigger. I suspect much of the new functionality in the 1870 will migrate down to the 1850.
It is significant as there are not that many boxes - none, really - that can do converged SDH/OTN/layer-2 switching all on one backplane. Several other vendors such as Ciena with its 5400, Cyan and Tellabs are going in this direction but Alcatel-Lucent has some legitimacy since it already was out with the 1850. Only the Fujitsu 9500 is in this class.
GMPLS/ ASON allow routers to communicate with the layer one infrastructure and set up and tear down paths as needed. It gives the router visibility into the lower layers of the OSI stack.
I think the key point is really the 1870 TTS’s 4Tbps switch capacity. This box represents the cutting edge of converged layer-1 plus layer-2 packet optical transport system technology.
Now we will see whether carriers adopt this architecture or whether they use IP over WDM or OTN switching only.
The Alcatel-Lucent 1870 TSS: the two central cards, larger than a shelf, each contain four 1Tbps universal switch ICs. There are two cards per platform as one is used for redundancy.
EPON becomes long reach
“Rural [PON deployment] is a tough proposition”
Barry Gray
Moreover, the TK3401 supports up to four such EPONs. The chip does not require changes to EPON’s optical transceivers although wavelength division multiplexing (WDM) transceivers are needed for the greater reach.
The TK3401 sits within what Barry Gray, director of marketing for Teknovus, calls the Intelligent PON Node (IPN). The IPN resides 20km from the subscriber’s optical network unit (ONU), where the PON’s optical line terminal (OLT) normally resides.
On one side of the IPN platform are sockets for up to four EPON OLT transceivers that support the PONs. On the other side are four SFP WDM transceivers that communicate with the central office up to 80km away and where the OLT platform is located. The OLT line card instead of using OLT optics uses WDM transceivers also in the SFP form factor. As such the line card does not require any redesign (see diagram).
Up to four point-to-point fibres can be used to connect the PONs’ traffic to the OLT, or a single fibre and up to 8 lambdas with coarse WDM (CWDM) technology to multiplex four PONs onto a single trunk fibre.
The 256 subscribers are achieved using a PX20+ specified optical transceiver. “It has a 28dB link budget such that going through 8 splitter stages is still sufficient for 2km distances [from the ONUs],” says Gray. “This is ideal for multi-dwelling unit deployments.”
Besides the pluggable optics, the IPN design includes the TK3401, a field programmable gate array (FPGA), and a flash memory.
The TK3401 comprises an EPON ONU media access controller (MAC), microprocessor and on-chip memory. The MAC registers the IPN with the central office OLT to set up remote IPN management and configuration communication links. The on-chip memory holds the firmware that configures the FPGA on start-up. The FPGA implements a crossbar switch to connect traffic from any of the EPONs to any of the WDM ports.
The IPN approach offers other advantages besides the 100km reach and increased subscriber count. It has a power consumption of 20W which means it can be powered from such locations as a telegraph pole. As the PONs are first populated, all four PONs’ traffic can also be aggregated into a single WDM link OLT port, with OLT ports added only when needed. In turn a fibre link can be used for protection with a sub-100ms restoration time.
However, unlike long reach PON or WDM-PON which also offer a 100km reach, the Teknovus scheme still requires the intermediate network node. The node is also active as it must be powered.
Teknovus claims it has strong interest from its IPN-based EPON architecture from operators in Japan and South Korea, while interest in China is for rural PON deployments. “Rural [PON deployment] is a tough proposition for service providers,” says Gray. “There is not the subscriber density and it is more expensive; the same is also true for mobile backhaul.”
The company is demonstrating the IPN to customers.
Click here for Teknovus' IPN presentation and White Paper
Retweet
"It's a bit like working with your wife; it has its ups and downs."
Siraj ElAhmadi, CEO of Menara Networks, on what it is like working with his brother, Salam, who is the company's CTO.
Titbits and tweets you may have missed
- ENISA has published a comprehensive report entitled Cloud Computing: Benefits, risks and recommendations for information security. The 125-page report can be downloaded from the ENISA site, click here.
- There is also a SecureCloud 2010 conference coming up in March involving ENISA and the Cloud Security Alliance, click here for details.
- A European court overruled German regulators that had allowed Deutsche Telekom to ban competitors from having access to its high-speed broadband network.
- Teknovus announced availability of the TK3401 EPON node controller that supports central office (OLT) to subscriber (ONU) distances of up to 100 km and up to 1,000 subscriber ONUs. And if you ever wondered what is the difference between the optical network unit (ONU) and optical network terminal (ONT), here is the answer.
- Broadcom announced its plan to acquire Dune Networks for $178m. Meanwhile, the 10GBASE-T copper interface got a shot in the arm with start-up Aquantia raising $44m in financing.
Retweet
Tidbits and tweets you may have missed
1. ENISA has published a comprehensive report entitled Cloud Computing: Benefits, risks and recommendations for information security. The 125 page report can be downloaded from the ENISA site, click here. http://tr.im/GC8x
2. A European court overruled German regulators that had allowed Duetche Telekom to ban competitors from having access to its high-speed broadband network. http://tr.im/GC7O
3. Teknovus announced availability of the TK3401 EPON node controller that supports central office (OLT) to subscriber (ONU) distances up to 100 km and connections to over 1,000 subscriber ONUs http://tr.im/GCe2 And if you evered wondered if there is a difference between the ONU and ONT here is the answer http://tr.im/GCgu
4. Broadcom announced its plan to acquire Dune Networks for $178m http://tr.im/GCf3
5. Meanwhile 10GBASE-T copper standard got a shot in the arm with start-up Aquantia gets $44m in financing http://bit.ly/6MCwdk
Quotes
"It's a bit like working with your wife; it has its ups and downs."
Siraj ElAhmadi, CEO of Menara Networks on what it is like working with his brother, Salam, who is the company's CTO.
Video compression: Tackling at source traffic growth
The issue is the same whether it is IPTV services and over-the-top video sent over fixed networks, or video transmitted over 3G wireless networks or even video distributed within the home.
According to Cisco Systems, video traffic will become the dominant data traffic by 2013. Any technology that trims the capacity needed for video streams is thus to be welcomed.
"The algorithm uses a combination of maths and how images are perceived to filter out what is not needed while keeping the important information.
Angel DeCegama, ADC2 Technologies
Massachusetts-based start-up ADC2 Technologies (ADC2 stands for advanced digital compression squared) has developed a video processing technology that works alongside existing video coder/decoders (codecs) such as MPEG-4 and H-264 to deliver a 5x compression improvement.
ADC2 uses wavelet technology; a signal processing technique that for this application extracts key video signal information. “We pre-process video before it is fed to a standard codec,” says Angel DeCegama, CEO and CTO of ADC2 Technologies. “The algorithm uses a combination of maths and how images are perceived to filter out what is not needed while keeping the important information.”
The result is a much higher compression ratio than if a standard video codec is used alone. At the receiving end the video is decoded using the codec and then restored using ADC2’s post-processing wavelet algorithm.
DeCegama says the algorithm scales by factors of four but that any compression ratio between 2x to 8x can be used. As for the processing power required to implement the compression scheme, DeCegama says that a quad-core Intel processor can process a 2Gbit/s video stream.
ADC2 Technologies envisages several applications for the wavelet technology. Added to a cable or digital subscriber line (DSL) modem, operators could deliver IPTV more efficiently. And if the algorithm is included in devices such as set-top boxes and display screens, it would enable efficient video transmission within the home.
The technology can also be added to smart phones with the required wavelet processing executed on the phone's existing digital signal processing hardware. More video transmissions could be accommodated within the wireless cell and more video could be sent from phones via the upstream link.
ADC2 Technologies demonstrated the technology at the Supercomm trade show held in Chicago in October. Established in 2008, the 5-staff start-up is looking to develop hardware prototypes to showcase the technology.
The first market focus for the company is content delivery. Using its technology, operators could gain a fivefold improvement in network capacity when sending video. Meanwhile end users could receive more high-definition video streams as well as greater content choice. “Everyone benefits yet besides some extra hardware and/or software in the home, the network infrastructure remains the same,” says DeCegama.