by Pravin Mahajan, Director of Product and Corporate Marketing
These are exciting times at Infinera. In 2012 Infinera entered the 100G long-haul market with a PIC based platform, skipping over 40G. It was a big bet and one that paid off. We believe our DTN-X platforms played a significant role influencing the acceleration of the 100G long-haul market. Today, Infinera is an industry leader, having shipped over 30% of all 100G WDM ports outside of China since 2012.
At Insight Infinera 2014, held in September 2014, Infinera introduced the industry’s first metro Datacenter Interconnect (DCI) platform to help service providers with burgeoning datacenter to datacenter traffic. Similarly, the Cloud Xpress, a PIC-based platform purpose-built for the metro DCI application, also intersected the inflection point of this rapidly growing 100G market. Today we are shipping multiple variations of Cloud Xpress with 10GbE, 40GbE and 100GbE interfaces, offering truly optimized rack-and-stack style DCI platforms. At Insight Infinera 2014, we were clear about signaling Infinera’s entry to the broader metro market at what we and most analysts believed to be the 100G metro inflection point, sometime in late 2015 or early 2016.
Today at Insight Infinera 2015 we have done just that with two key metro announcements. The first is our announcement of new products for the surging 100G metro market with XTC-2 and XTC-2E platforms – an extension of the DTN-X family. Our customers benefit from the scalability, flexibility and simplicity in long-haul using the DTN-X XTC-10 and XTC-4 and they’ve often pressed us to extend that benefit for smaller sites with lower bandwidth needs. Today we address that requirement with the DTN-X XTC-2 and XTC-2E, which provide right-sized 100G packet-aware OTN switched WDM transport.
Second, we unified our portfolio by introducing new hardware and software capabilities for the DTN-X Family and the recently acquired TM-Series metro packet-optical platforms for efficient interoperability. Over a month ago we closed the Transmode acquisition. As a result of working with customers, we identified an immediate integration requirement between the TM-Series and the DTN-X Family, the need to aggregate traffic from the service-rich metro aggregation network to the OTN-switched metro core and regional networks that use coherent 100G technology for efficiency. To meet that need, today we introduced WDM line-side data plane interoperability between the TM-Series and the DTN-X XTC series managed by the Infinera DNA.
The TM-Series opens up new opportunities in metro access and aggregation as the metro network is increasingly becoming fiber-deep as a number of networking applications are being virtualized and centralized to run on commodity server hardware which in turn is increasing the need for WDM optical connectivity. According to Heavy Reading analyst Sterling Perrin, metro networks are being disrupted due to the evolution of cloud, video scale and the explosion of mobile and the Internet of things (IOT) – all catalysts in the emergence of cloud services (Layer C) and intelligent transport (Layer T). Cable and mobile infrastructure is being built with mini datacenters in the metro, containing cloud functions such as virtual CCAP, vRAN to name a few. Mini datacenters need to be connected by fiber to end points close to the user. Sterling observes that the metro aggregation networks are being built with packet-optical transport systems and router functions are being converted to software running on commodity x86 server hardware. Infinera’s TM-Series offers packet-optical transport for a rich set of metro applications including mobile fronthaul and backhaul, business Ethernet and triple play aggregation for cable and service provider networks. The recently announced PT-Fabric for the TM-series is designed to extend these capabilities to terabit packet switching with flexible, granular control. Complementing the TM-Series, the DTN-X XTC series platforms map the packet switched traffic to digital OTN containers, and efficiently switch and transport the traffic over core and regional networks. We believe in responding to our customer needs with speed. The unified packet-optical solution for the metro is the first amongst many upcoming integration milestones for Infinera.
At the Insight Infinera 2015 event earlier today, our executives shared Infinera’s plans for the expanded WDM transport market opportunity ahead, which is three times more with our expansion in metro. As we proceed with the Transmode integration, it is important for our customers to have a precise set of tools for their networks. We continue to invest in our long haul portfolio. There is tremendous growth in traffic in the metro as central offices and headends transform into datacenters. We are seeing a similar evolution of traffic patterns in the long-haul network. Based on feedback from customers, we realized that some need a long-haul point-to-point interconnect solution that can be used to connect datacenters and other large bandwidth sources over long haul optical fibers and line systems (used for signal amplification) which are already in place. To meet that need, today we introduced the DTN-X XT-500, a compact ultra-dense platform for long-haul point-to-point interconnect including DCI. This platform is an extension of the DTN-X Family, working over FlexILS line system and manageable as a single node along with the rest of the platforms in the family. The XT-500 uses PIC-500 and supports Instant Bandwidth™ enabling customers to deploy cash-flow efficient bandwidth matching end-customer needs.
While some in the industry announce capabilities far ahead of shipping, we at Infinera believe in introducing the right product at the right time. The solutions we introduced today are shipping this quarter. The entire unified portfolio is managed by the Infinera DNA management system and can be viewed live on the Infinera Express, our mobile demonstration units in USA and Europe.
Infinera now provides network operators with a unified end-to-end portfolio of products to address the entire packet-optical and WDM market, whether for metro access, metro aggregation, metro core, long-haul, subsea or datacenter interconnect segments. Infinera’s unified portfolio is powered by key foundational attributes of scalable photonics, flexible and granular control, and programmable, open interfaces.
Kudos to the entire Infinera team, from Sunnyvale, Stockholm and other locations around the world, that continue to deliver on Infinera’s founding vision of enabling an infinite pool of intelligent bandwidth that the next communications infrastructure is built upon. We look forward to continuing to enable carriers, cloud operators, governments and enterprises to scale network bandwidth, accelerate service innovation and simplify optical network operations with Intelligent Transport Networks.
by Pravin Mahajan, Director, Product & Corporate Marketing
Industry analysts have highlighted the unrelenting growth in bandwidth demand across subsea, long-haul and metro, due to video, mobile traffic and cloud architectures driving east-west datacenter to datacenter traffic running over transport networks. While long-haul networks are transitioning to 100 gigabits per second (100G), the explosive growth in traffic and a shift to cloud-based delivery of applications is now forcing a similar transformation of metro networks from 10G to 100G. Many studies show that metro traffic is now growing much faster than long-haul traffic.
Mark your calendars for Insight Infinera 2015 when Infinera executives and leading operators share their insight on network transformation and the evolution of metro networks from 10G to 100G.
Numerous applications are on the cusp of being virtualized, such as C-RAN in mobile networks and the move toward remote PHY in cable networks. These applications are leading to the creation of mini datacenters closer to the end-user and a shift towards a cloud based infrastructure across the metro network. As metro network architectures evolve to Layer C (cloud services) and Layer T (intelligent transport), I believe that the strategic importance of the transport layer has never been greater and as a consequence it is vital to build Layer T with the right attributes: scalable optics, flexibility with granular control, and finally, programmability.
Last year, at Insight Infinera 2014, Infinera announced a strategic pivot to move from a single-market focused business to a multi-market company, addressing the end-to-end opportunity in Layer T. Infinera was the first to develop a purpose-built DCI platform – the Infinera Cloud Xpress – for this fast growing market segment. And today we are still the only company shipping a server-like rack-and-stack style DCI platform.
Analysts believe the metro market will start its transition to 100G in earnest in 2016. Infinera is poised to enter this market with our recently completed acquisition of Transmode, a leader in metro packet optical.
We are excited about what the future portends with new technologies, new products and now an entire metro portfolio to address the metro network’s transition to 100G and cloud-based architectures. Please register for our live webcast to hear key Infinera executives talk in detail about this future, about the progress we are making integrating our new metro portfolio, and about new technologies that will help operators scale bandwidth, accelerate services and simplify network operations.
The future of metro unveiled. It all begins on October 6. Join us by registering for the live webcast.
by Vinay Rathore, Sr. Director, Field and Segment Marketing
Ahh, New York, New York. What a great place to celebrate another fantastic milestone in our partnership with Arista.
Yesterday we joined Arista at an exclusive event at the New York Stock Exchange (NYSE) to advance Arista’s CloudVision solution. Infinera CEO Tom Fallon followed Arista CEO Jayshree Ullal on stage to talk about our partnership and showcase our latest collaborative project: integration of Infinera’s Cloud Xpress optical datacenter interconnect platform with Arista’s CloudVision datacenter automation software.
We announced our technology alliance with Arista in May with an initial focus on end-to-end solutions that are high-performance and pre-tested for interoperability. In June we joined Arista’s initial CloudVision launch, expressing our support and alignment with their direction. And now we’ve delivered a concrete demonstration of how integrating Cloud Xpress into CloudVision can simplify provisioning of datacenter interconnection and bring it into a datacenter automation framework.
The demo integrates a new Infinera-developed application into Arista’s CloudVision and enables new datacenter to datacenter connections to be set up with a few simple clicks on a graphical user interface, eliminating cumbersome one-at-a-time provisioning of each individual Arista and Infinera platform. To see it in action, check out the demonstration video below, introduced by Stu Elby, Senior VP of Cloud Network Strategy and Technology.
Video: Integrating Cloud Xpress with Arista CloudVision
While this is only the latest step in our developing partnership, it is in some ways the most significant because it signals a commitment to tighter integration of our joint solutions. We are aligned and driven to help our mutual customers simplify their datacenter operations so they can continue to scale their datacenter networks efficiently.
At the end of the day, when Tom joined Jayshree for the NYSE closing bell, it was a fitting way to ring in a new era of partnership.
by Geoff Bennett, Director Solutions and Technology
While Indium Phosphide has ruled the roost in terms of “production grade” photonic integration for over a decade, silicon is rapidly becoming the dominant material in certain optical applications. I thought it might be useful to review some common misconceptions about silicon photonics in particular.
Is Silicon Photonics Cheaper Than InP?
The opportunityfor cost savings in silicon photonics is a much-misunderstood area. Silicon is a great material for many optical applications, and it’s relatively simple to work with. Silicon can’t be used for lasers or direct amplification, of course, and therefore much of the complexity (and R&D investment) at the complex component end involves the need to integrate an active III-V material (typically an InP heterostructure) onto a silicon “motherboard”.
When it comes to coherent super-channels the key to low cost is to integrate the maximum number of optical elements onto a single chip. For that, I believe InP is without doubt the best material.
Is silicon always cheaper in other, less demanding applications? This is hard to say for sure, because much of the cost of the final component is in the packaging, and any packaging technology that reduces cost could be applied generally to both silicon and InP. So in less demanding applications I’d say it’s a tie.
Does Silicon Photonics Consume Less Power Than InP?
The claim tends to be that silicon is “thermally insensitive”, and so it can be used in uncooled packages which will save a considerable amount of power (note that active coolers, such as Peltier devices, can be quite power-hungry).
This is mostly true for simple optics, but silicon still has a thermal sensitivity – it’s just much lower than InP. But as soon as we add III-V active elements to the silicon “motherboard” we have to provide temperature stabilization if we need to control the wavelength closely. Fortunately for the types of applications being promoted by Intel and IBM (i.e. evolving to optical computing over short distances) temperature control is less important because these devices tend to operate as Coarse WDM (CWDM), with “free running” wavelengths. The result of that decision is that individual fiber connections are limited to about 100Gb/s today. This may scale in the future by, I would say, as much as a factor of ten using known technologies, but that would still be a world away from the several terabits of capacity offered by InP based PICs over regional and long haul distances today.
From my perspective, this is not an “apples to apples” comparison. Silicon has advantages over short distances, but in my opinion InP is the only option over longer distances.
If we look at InP’s application for multi-carrier super-channels, which is admittedly a real sweet spot for this material, the combination of multiple active components onto one chip will drastically reduce the cooling required per active element.
If InP PICs are so good, why doesn’t everyone make them?
This is a question I’ve faced many times in the past ten years with Infinera.
First let’s be clear. The evidence of the value of InP-based PICs is now overwhelming – and this is not the part of the question that’s in play. In other words – PICs really are “that good”. It’s the second part of the questions that needs some explanation.
So why don’t optical component companies build PICs? They do! JDSU is a classic example of a component vendor that has developed integrated devices (tuneable laser plus Mach Zehnder modulator) based on InP for single wavelength operation. I can speculate on why JDSU and other component companies don’t go a stage further and develop multi-wavelength “large scale” PICs – and I think it boils down to two reasons.
First is that designing and manufacturing large scale PICs in commercial yields is incredibly difficult. Infinera is the only company to have demonstrated this capability publicly to date, and we’ve been doing it for more than a decade.
The second issue is about business strategy. A component vendor does not design the systems that these large scale PICs will go into. So if they do choose to develop PICs, they need to make them as widely applicable as possible so that they can address a wider market of vendors, and that means stepping back to the lowest common denominator of wavelengths…which is one!
Unlike any other company in the industry, Infinera has the benefit of being vertically integrated. We design and build our own large scale PICs (which, as far as I know, is unique), and we design and build our own coherent ASICs (less than a handful of optical transport vendors have this capability). Then we take these devices and design and build our own super-channel optical transport systems. This means that we can optimize the number of wavelengths we need per chip, and match them to the transport system they will go onto in a way that I believe no other company is able to do.
Finally we have the small matter of funding – Infinera needed hundreds of millions of dollars to complete the process from PIC design through manufacturing and then system design and its manufacturing. It would be extraordinarily difficult for any other company or VC fund to commit to this scale of investment, and the associated risk of failure, today.
Will there be “more Infineras” emerging in the optical world as the value of photonic integration spreads from transport to data processing? Time will tell.
Integration of Optical and Electronic functions
Here is my take …since silicon could be used for both the electronic and optical functions, this means that we could have computer chips that use light instead of electrons. The idea is great in theory, except that the feature size (i.e. the size of the silicon features used to implement optical functions on the chip) is in the range of 100 nm. Today’s Broadwell CPUs from Intel use a 14nm process that is currently many times too small for optical applications. The good news is that an older 100nm silicon CPU fab could be given a new lease on life to build silicon photonics, but the electronics it could support would be extremely dated, and simply not powerful enough to keep up with the optical functions integrated onto the same chip. To achieve modern electronic performance it would be better to connect the silicon optical chip to a separate silicon electronic chip, which is the architecture you will see being used by silicon photonics proponents.
Both InP and silicon play a vital role in today’s optical component toolkit, but as I see it here are the chief differences . InP rules for any active functions, and over distances longer than a few kilometers. InP is also the only material today that can be used to build large scale devices for super-channels. Silicon is ideal for simpler, single wavelength applications, and for co-packaging with active devices in the “pluggable” market for client optics and metro transponders.
Infinera welcomes the advances offered in silicon photonics because we use pluggable devices as client interfaces across our product range. And as silicon photonics on the server motherboard helps to continue the growth of data center capacity, so we will benefit by offering terabit scale interconnect for this growing market.
By Pravin Mahajan, Director Product and Corporate Marketing
A few months ago, I ordered pizza while watching the FIFA women’s world cup soccer final between USA and Japan. I enjoyed the pizza and the USA victory while my spouse was on the web searching for vacation rentals at Airbnb. Simultaneously, I was receiving emails about a critical situation with one of our subsea customers, which was resolved with a new solution that we were about to announce in few months. All these came together in an “aha” moment – the new on-demand economy is real and transforming businesses.
Let me explain, beginning with the blue waters of the Pacific Ocean.
Recently, one of our subsea customers – Australia Japan Cable (AJC) – experienced a network path failure following a subsea cable fault in 7000 meters to 8000 meters in the Pacific Ocean. The Infinera team and AJC rerouted over 400 gigabit per second (Gb/s) of traffic from the impacted AJC section to alternate paths on the AJC optical fiber ring configuration. The rerouting utilized the capabilities of the Infinera DTN-X platform featuring Infinera’s new on-demand Time-based Instant Bandwidth™ technology. The rerouting of the lost traffic was able to commence within minutes after receipt by AJC of the Time-based Instant Bandwidth license. This avoided a route outage while the cable repair ship was mobilized, transited to the repair site, and implemented in the challenging deep water repair.
Since late 2000, the cloud computing pioneer Amazon Web Services (AWS) was touted as a model for the adoption of on-demand rentable solutions. Infinera introduced the on-demand solution in the network realm in November 2012 with announcement of the Instant Bandwidth solution. Many customers have adopted this capability realizing tremendous value in their networks. Working with them, Infinera continues to extend the solution and Time-based Instant Bandwidth is one such outcome.
It enables the on-demand activation of bandwidth in various increments, including 100 Gb/s, depending on the application for specified durations of time, with the deactivation of the line-side bandwidth at the end of the time period. This is an expansion of the Infinera Instant Bandwidth solution, which allows permanent software activation of line-side bandwidth in various increments. This solution creates on-demand bandwidth using software based licensing without needing any new hardware. Customers can now effectively rent network capacity on their transport infrastructure.
The Instant Bandwidth solution takes advantage of pre-deployed optical network capacity enabled by Infinera’s photonic integrated circuit (PIC) based super-channel technology on the DTN-X platform, part of the Infinera Intelligent Transport Network architecture. Thus operators can deploy slices of capacity over time as they turn on the services. In contrast, competitive solutions use conventional technologies, like discrete optics instead of the PIC, resulting in an operational “penalty”. This is because they require multiple time consuming activities for each additional 100G line card that needs to be deployed such as ordering, shipping, installation and provisioning taking several weeks to fulfill. A potential alternate approach for a competitive solution would be to over-order WDM line-cards for a chassis incurring a significant money “penalty”.
Traditional on-demand delivery options have been limited to local pizza and Chinese takeout. The technology industry is disrupting businesses with the on-demand model resulting in the new rental economy. It is being exemplified by companies such as Airbnb and Zipcar in the consumer space; Amazon, Dropbox and Microsoft in the business segment. Each company is creating and capturing its own blue ocean.
Infinera pioneered the new approach to turn up network capacity with Instant Bandwidth. The recent enhancement with the Time-based offering now allows operators to deploy network capacity for a defined duration, with its trial by fire being deep underneath the blue waters of the Pacific Ocean.
For more information on Time-based Instant Bandwidth please visit www.infinera.com/go/instant.
by Dave Welch, co-founder and president, Infinera
Today, Infinera completed its offer for the shares of Transmode, a leader in metro packet-optical.
The acquisition of Transmode marks a significant milestone for both our companies and our customers. By bringing together two industry innovators, Infinera now offers an end-to-end product portfolio to address the $15.3 billion global optical networking market* including long-haul, Cloud and metro.
Transmode’s service-rich metro platforms, broad European customer base and shared culture of innovation is an ideal complement to Infinera – a true 1+1=3 scenario. By joining together, we can leverage the strengths of both companies, expand markets and will continue to serve our combined customer base with the passion you have come to expect from us.
As the optical networking market continues the once-in-a-decade transition from 10G to 100G, the metro market is at the beginning of this shift to 100G and I believe that the timing is right for Infinera and Transmode to join forces. I’m excited by the opportunity to combine Transmode’s metro portfolio with the advanced technologies developed by Infinera for the long-haul and Cloud. The expanded portfolio is compatible with our vision of using software defined networks to unleash the power of programmable optical networks.
Infinera now offers a comprehensive portfolio of solutions for the vast metro aggregation market, including metro core, edge and access, across a diverse set of fast growing applications like mobile fronthaul and backhaul, broadband aggregation and business Ethernet services with Metro Ethernet Forum (MEF) certification.
We are excited about the opportunity to deliver our Intelligent Transport Network to our expanded customer base which now includes more than 650 Transmode customers. Similarly, we look forward to offering Transmode’s extensive metro solutions to Infinera’s customers.
A key benefit of the acquisition is the immediate expansion of our skilled resource pool – a team focused on developing innovative Intelligent Transport Networks and providing the Infinera Experience to our customers. As part of this expansion, Karl Thedéen, CEO of Transmode, joins Infinera as the managing director of our Metro Business Group. We welcome Karl and all of the Transmode employees to the Infinera team as we continue to advance the optical networking industry together.
We asked Andrew Schmitt, research director, Carrier Transport Networking at IHS Research what he thought of the deal and here is what he told us:
“The combination of Infinera and Transmode makes sense from both a market and cultural perspective and I am excited to see how the teams will build off of each other’s strengths. This acquisition positions Infinera to capitalize on the growing metro market as it begins a transition to 100G.”
We believe that when our customers succeed, we win. Infinera and Transmode share a culture of customer first, strong technology and product excellence. The scale of the combined company allows for further investments that strengthen our differentiation in the optical networking market. We look forward to working with the team at Transmode to deliver Intelligent Transport Networks that accelerate service innovation and simplify optical network operations.
* Source: Infonetics Q4 2014 Market Forecast for 2019