Infinera Product Evangelist
One thing I hear again and again from Infinera customers is that our gear is particularly easy to use. And not just a bit easier than other gear – they tell me that it’s a night and day comparison with conventional DWDM equipment. We’ve even been able to establish a Guinness World Record as a result of the scalability and ease of use of the Intelligent Transport Network.
Ease of use has many facets, but one obvious requirement is that it must be possible to automate day to day operations within the transport network. Automation is a major factor in reducing network operations costs, and delivering services more rapidly.
We achieve automation using Control Plane protocols that make the network self-aware, such as Generalized MPLS (GMPLS), so that every network element understands the status and capabilities of every other network element. But the reason that we believe Infinera’s GMPLS is dramatically more effective than other DWDM vendors’ is that we’ve always built OTN switching into every node we make. In fact our DTN-X platform is an OTN powerhouse, with 5Tb/s of non-blocking switching capacity in a single chassis, scalable up to 240Tb/s in multi-chassis configurations while remaining non-blocking. In other words we’ve been able to automate Layer 1 – or the digitally switched layer of the network – from the very beginning.
The next step is to allow Layer 0, the optical layer, to be automated as well. To do this we have introduced the industry’s first Colorless, Directionless and Contentionless (CDC) super-channel FlexROADM. So what does all that techno-speak actually mean?
Vice President, Corporate Marketing
Today Infinera and Telefónica announced a successful collaboration in which Telefónica I&D developed an MPLS Network as a Service (NaaS) application which provided point-and-click provisioning of MPLS services and automatically configured the underlying routers and Infinera Intelligent Transport Network™ to deliver those services. The NaaS application ran on top of a Telefónica developed multi-layer IETF-based Application Based Network Operations (ABNO) controller. You can see a video of the demo here.
Infinera’s Open Transport Switch (OTS) is a lightweight software construct that can either logically abstract and virtualize the resources of network element, or logically abstract the resources of multiple connected network elements leveraging a GMPLS control plane. The OTS has both OpenFlow and RESTful APIs for configuration, provisioning and monitoring. With this approach Infinera enables direct access to the programmable transport network so service providers can leverage any controller versus being forced to purchase a controller from a particular vendor.
The vision of SDN includes standardized open APIs to foster an open environment that enables third party application development, which will help SPs unleash rapid innovation to drive new revenues and reduce costs through automation. Vendor solutions featuring tightly coupled hardware and software have slowed down industry innovation for the last three decades. Because of this, many carriers desire to develop their own controllers or use third party controllers to avoid vendor lock in. Infinera has now completed two successful trials enabling multi-layer integration with the Infinera lightweight OTS and router network elements using a controller and applications developed by the service providers. The first trial was conducted using ESnet’s controller, and today we have announced a trial using Telefónica’s controller.
It is clearly time to move networking into the 21st Century – I believe SDN and Infinera’s Intelligent Transport Network can help with that transformation.
Senior Technical Recruiter
People are often amazed by the “can-do” spirit, tenacity and accomplishments of Infinera employees. Maybe it’s because we offer our employees broad responsibility and scope of work instead of confining them into narrowly defined areas of responsibility. We believe that by providing more exposure to varied projects and responsibilities and encouraging employees to dream big, we build not only the greatest products, but also the greatest careers. As we mount the campaign towards changing our industry, we are continually seeking the brightest minds and the best talent to get us there.
What can someone expect once they are hired at Infinera? For starters, we have an environment unlike most others. Our employees understand that we are leading the charge toward changing our industry with our breakthrough technology and that each of them is taking part in building the future of communications technology. Is this the easiest path? No. But why spend your career at a company that’s doing anything less? Our teams are filled with some of the most talented individuals on the planet and our goal, unlike that of many companies, is to find ways to allow them the freedom to create.
Like the Silicon Valley pioneers before us who dared to dream and created what was once thought unattainable, Infinera is reshaping the communications landscape by advancing technology that many believed was impossible. Infinera utilizes breakthrough semiconductor technology, which integrates hundreds of optical functions onto one small chip, the photonic integrated circuit (PIC). Infinera’s PIC technology is not sold directly but instead is used as a key differentiator inside our own Intelligent Transport Network™. Our solutions are market leading and widely recognized in the industry as the easiest to use and most reliable solutions available.
Would you like to join us in this journey and help advance the industry? Visit us at www.infinera.com/careers.
Sr. Director, Corporate Communications
On Thursday, Infinera Senior Vice President Fred Kish was elected as a member of the National Academy of Engineering (NAE), one of the highest professional accolades that can be awarded to an engineer in the United States.
Fred was one of 67 new members and 11 foreign associates who entered into this prestigious group, consisting of 2,250 members and 214 foreign associates in total from the fields of technology and engineering. The NAE honors “…those who have made outstanding contributions to ‘engineering research, practice, or education, including, where appropriate, significant contributions to the engineering literature,’ and to the ‘pioneering of new and developing fields of technology, making major advancements in traditional fields of engineering, or developing/implementing innovative approaches to engineering education.” You can catch the full press release here.
Fred was recognized for his contributions to high-brightness light-emitting diodes (LEDs). As part of this work, he co-invented and led the commercialization of the highest performance (efficiency) red-orange-yellow visible LEDs as well as the first high-power high-brightness LEDs. These were the first LEDs with efficiencies exceeding halogen lamps. These LEDs have been widely used in applications ranging from traffic signals to automotive brake lights to large-area displays, and as elements in solid-state lighting systems.
Currently, Fred is Senior Vice President of the Optical Integrated Components Group at Infinera. Here he leads the photonic integrated circuit (PIC) group, where he and his team co-invented the first practical (commercial) large-scale PICs, including 100Gb/s and, more recently, 500 Gb/s transmitter and receivers.
Congratulations again to Fred on receiving this prestigious honor! For more information about the National Academy of Engineering, please visit www.nae.edu.
Director, Corporate Marketing
A few weeks ago I wrote about pitfalls to watch for if you’re thinking about building your own fiber network. Some large US content providers have already built their own networks, and it may be that others are at least thinking about the prospect after January 14th when the D.C. Circuit ruled that the Federal Communications Commission (FCC) cannot require network neutrality in the US. The implication of the ruling is that broadband and backbone internet providers can develop new business models, such as charging a premium to internet content providers like Netflix or Google in exchange for high speed access to broadband customers. This issue was recently overviewed by Marguerite Reardon in her FAQ on the subject in CNET.
While the full implications of the ruling will take time to play out, it does change the risk profile for many bandwidth-intensive content providers. Internationally, the picture is much more muddled, with different degrees of network neutrality and networking shaping permitted (see http://en.wikipedia.org/wiki/Net_neutrality. )
If your business model assumes a level playing field in terms of getting your content to your customers, it’s probably a good idea to consider new scenarios such as:
- How vulnerable are you to price increases by carriers, particularly those running into capital expansion problems, resulting in changed tiered pricing for volume of traffic and level of service?
- How well positioned are you to take advantage of scale in your market and gain preferential access to customers and, in some cases, suppliers?
- If you’re a large content provider, what competitive effect would building your own transport network have?
Some (mostly large) content providers will probably figure out that owning their own optical fiber will reduce their operating costs and provide negotiating leverage with carriers. While owning the last mile of connection to homes and businesses is extremely capital-intensive, these companies may determine that paying carriers for only the last mile will be less costly over time than paying for the last mile plus metro and long haul packet transport. Who knows, in some cases, investing in wireless last mile connections may eventually be cost effective.
But it’s important to remember there are risk mitigation choices. In addition to acquiring their own fiber through Indefeasible Rights of Use, purchase or construction, companies have three other options: (1) negotiate longer term contracts that lock in today’s rates, (2) partner with other companies to acquire and operate capacity, and (3) do nothing and lease capacity as needed from wholesalers and new transport owners with excess capacity. It’s also possible that bandwidth derivatives may emerge for the first time as a mechanism for risk management.
Whichever path you take, the one thing you do not want to do is get stuck with hard to implement and user-unfriendly equipment. Sorting out your fiber strategy is going to be tough enough.
Sr. Director, Corporate Communications
For many years now, the Pacific Telecom Council holds their annual conference in Honolulu, Hawaii. This year PTC’14 brought together the leading submarine network operators from all reaches of the Pacific Ocean. Hawaii is a natural location for such an event as many submarine optical networks terminate in Hawaii.
The week running up to the show is typically filled with new announcements related to the subsea optical network market. We were pleased to announce Australia Japan Cable the week before the show, and were delighted to announce that Telstra Global joined the list of Tier-1 network operators that have selected and deployed an Intelligent Transport Network™ from Infinera.
Vice President, Network Strategy
2013 is now history, but there is an architectural generalization that some in the telecom industry will undoubtedly carry over into the new year that goes something like this: “Traffic is increasingly becoming IP-based, therefore in order to simplify the network, everything should be IP-over-WDM.”
It’s difficult to argue with the fact that IP is ubiquitous. But I feel that it is an overly simplistic approach to draw a conclusion that the ideal network architecture should be based on just the dominant traffic type. Here’s why:
The conventional 7-layer OSI stack has for a long time served as the de facto standard for describing a hierarchical set of functions – each layer of the stack is self-contained with clean abstraction delineating one layer from the other. Each layer serves a particular purpose in networking – this is an important factor often overlooked during planning and design. In the early days, each functional layer of a network often was implemented by a separate device. We had Layer 3 IP (and then Layer 2.5 MPLS) being served by routers, Layer 1 SONET/SDH being served by ADMs, and Layer 0 transmission being served by optical transmission systems. Today the Layer 1 SONET/SDH technology is being supplanted by OTN as advances in optics and signal processing extend wavelengths speeds beyond 100G to the world of super-channels.
So the key question is whether the dominance of Layer 3 packet-based services is best served by the elimination of the networking functions described by Layer 1 – e.g., adaptation, grooming, multiplexing, and switching of bandwidth “pipes” onto optical wavelength carriers.
As the world becomes increasingly connected, one can envision everything having their own IP address and being somehow connected to the Internet of things. But the fact that the world’s services are largely becoming packet-based does not necessarily mean a service provider should architect the network plumbing without any of the underlying network layer functions from the OSI stack. In my opinion, elimination of an equipment layer clearly makes good economic sense. But that is different from removing a network layer function. IP-over-optics, sometimes referred to as IPoDWDM, eliminates Layer 1 functionality, and consolidates all bandwidth management functions within routers.
Switching is an essential networking function for steering and managing traffic across the network, and can be achieved at different network layers. The elimination of Layer 1 switching, however, means that fine-grained Layer 3 / Layer 2.5 packet-level switching would need to be substantially augmented to support various transport, and bandwidth management functions operators need throughout their networks. This is a non-optimal use of network resources, especially at sites where traffic is largely transit in nature. Research, such as that used in Nick McKeown’s work , demonstrates that using Layer 1 switching in the core increases overall network efficiency and improves overall economics – here Layer 3 / Layer 2.5 is not eliminated, rather just used where needed. My colleague, Pravin Mahajan, has blogged on this key aspect within a historical context – Switch Where You Can, Route Where You Must.
Modern transport systems today provide state-of-the-art Layer 0 transmission technologies that enable multiple Terabits of capacity on long-haul fibers, and can also incorporate flexible Layer 1 switching capabilities while supporting variable sized bandwidth circuits (i.e., virtual wavelengths) ranging from 1Gb to 100Gb in increments of 1Gb. Carriers deploying such transport systems today are eliminating the traditional Layer 1 network layer as a separate device, but they aren’t forgoing Layer 1’s transport networking functionality. In fact, commercial transport systems today can deliver 500G per slot of transmission and switching, while packet systems scale to 100G-200G per slot of routing. This is a matter of simple engineering, where each network layer function is used to its fullest potential. Recently, Alcatel-Lucent concurred in their Technology Symposium that using IP-over-optics by combining the packet XRS7750 core router with state-of-the-art elements of the 1830 Photonic Service Switch does not make sense for 100G.
So how does OTN help service providers in their networks?