Q. What kind of technologies are used by engineers in the optical communications field?
A. A company into optical communications builds transmission systems for voice, data and video based on optical fibre, and designs electronics that forms the data transmission backbone. In terms of specific technologies, considering the voice networks, some of the prominent technologies worked with are synchronous optical networking (SONET), synchronous digital hierarchy (SDH) and optical transport network (OTN). For data transmission, we have Carrier Ethernet technologies like multi-protocol label switching transport layer and Ethernet ring protection switching.
Within the communication products portfolio, we develop and design high-speed telecom systems. This requires multidisciplinary skills. We have high-speed PCBs, FPGAs (the silicon component that goes into designing the product), huge amount of embedded software that runs on this product and network management software (the software that runs outside the product managing larger networks).
Q. What is the most exciting technology in optical communications at the moment?
A. Micro packet optical transport is one of the technologies that we have been working on for the last few years. At the moment, we are working on next-generation transmission technologies based on packet transport network, and efficient energy management solutions for telecom tower operators.
Q. What are the challenges faced in transitioning from 2G to 3G/4G, and how do your solutions tackle them?
A. Initially, for a 2G network, sending data over time-division multiplexing (TDM) network is the conventional way. When the traffic increases, it is better to send data over a separate packet network rather than over the TDM network.
The voice infrastructure is supported in the usual way, i.e., TDM, but sending data over a separate packet data network rather than over a voice network is far more cost-efficient as well as bandwidth-efficient. When 4G will come and when the dominant traffic in the network would be packets/data, voice transmission is going to be more efficient over the data network by packetising it.
Our product allows the network to transition seamlessly from one network architecture to another without having to change your hardware. It is basically the reprogramming of your hardware that allows you to do that.
We achieved this through a lot of programmable silicon like FPGAs and by reprogramming the data path to move from TDM to packet and a combination of both, and so on.
Q. What makes transition from one communication technology to the next so cumbersome?
A. The development of packet optical system architecture has itself been a really interesting challenge. This is a system that helps in transition from a 2G to 3G to 4G network, which requires an enormous amount of flexibility in the product in terms of hardware as well as programmability.
A few years ago, the conventional network was mainly voice and the efficient way of transmitting voice traffic was through time-division multiplexing (TDM) technologies like SONET and SDH. When people started accessing data with Internet connectivity and mobile data applications via cell phone, there was some data bandwidth in the network that could still cater to this demand.
We need to remind ourselves that the infrastructure here was optimised for voice, but data got mapped onto the voice circuits and transmitted due to the demand. For a long time, Ethernet over SONET/SDH technology was prevalent in terms of how data was backhauled or transmitted over a network that was actually built for efficient voice transmission. However, when the amount of data in the network increased with 3G coming in, and a lot more Internet connectivity and broadband penetration, that was not the most efficient way of transmitting data. That was an expensive alternative and we had to change it.
Q. So what’s the more efficient way to transmit data?
A. Native packets. There are a lot of technologies like Carrier Ethernet that enable you to do that while bringing the best of TDM technologies in terms of protection, fault and monitoring. However, there is still a very large voice network and huge number of 2G subscribers who need to be supported. At some point of time when the data grows, it becomes very expensive to transmit it over a TDM infrastructure. So a separate packet infrastructure has to be built.
Q. What are native packets?
A. What I mean by native data is that instead of sending it over a voice circuit, we send it over a fibre circuit as an Ethernet packet. Of course, it is necessary to build a lot of other things along with the packet so that issues like fibre cut, data re-routing or fault location in the network are handled. When a network of thousands of network elements is built and spread across the country, it is very important to detect those faults (no traffic, no service) and locate the problem.
Q. How did Tejas Networks contribute to the Indian government’s RailTel in setting up a digital TV infrastructure using SDH/DWDM-based transmission systems?
A. It was built around ELAN technologies that not only provide an efficient transmission of video but also technologies like video multicast that are required for a television transmission infrastructure. ELAN helps to leverage the statistical nature of data traffic and builds data pipes that leverage the ‘burstiness’ of data traffic (where at some point of time there is very high data traffic and at some other time there is no data traffic at all). Multiple people can share this fixed pipe. So, there is efficient use of bandwidth over a shared infrastructure while transmitting data.
The video is transmitted as packets. So a very high amount of video traffic is transmitted efficiently over the shared packet infrastructure enabled by ELAN. An important benefit of this is the efficient use of bandwidth.
When multicasting of video is done, there is a source from where the TV traffic is derived. The source need not send the same amount of traffic to every user. The data can be sent as common traffic until a certain point and then it can be split. Being able to service so many viewers from a single head-end source but not burning up too much bandwidth is possible by knowing where the traffic should be replicated and transmitted efficiently.
Q. What challenges crop up while implementing communication systems in the field?
A. Our products are used for multiple applications. One of them is mobile backhaul, for which telecom service providers use our equipment. One of the biggest applications is connecting a base station to the switching centre. The challenges here are setting up a multitude of base stations simultaneously connected to the same switching centre, and ensuring that this set-up performs as a resilient network that is protected against hardware and software failures. For 3G and 4G services, when the user talks to the base station via his handset, a part of the network connects, takes that data from the base station and brings it to all the switching centres. This is known as backhaul.
Q. Could you elaborate further on the backhaul network?
A. It relates to the transfer of data from any user access point to a central point where it is switched or routed. For instance, we shall consider the backhaul of Internet services. People have broadband connectivity from home, so taking the data from your home to the Internet router is also called backhaul. It is a generic term for taking data from a user access point to a central switching station and redistributing it (wherein it goes to other ends of the network).
Q. What is the most innovative product or technology from Tejas Networks?
A. An interesting innovation that comes to my mind is the double-bandwidth protected ring (DBPR), which we have patented. It allows doubling of the microwave network capacity. Portion of the network built on microwave is not very high-speed as it is not fibre and goes over the air. If people have a network that drives traffic to packet instead of doing an overhaul of the network by deploying all the IP radios and so on, we provide them a technology called DBPR that lets them double the capacity of data traffic, which could be used intelligently for implementing protection architectures and more. So that is a very interesting technology because without changing the infrastructure or ripping/rebuilding the microwave infrastructure, people can use our products as an add-on to their network and double the capacity of microwave while transmitting data.