“We spend 50 percent of development costs on a product after the launch”
eeNews Europe: 5G wireless communications technologies represent an important emerging topic in today’s market environment. How does Rohde & Schwarz see this development, and what new worlds of application does it open up? How is Rohde & Schwarz preparing for 5G?
Roland Steffen: Naturally everything is expected to be much better with each new generation of wireless communications. 5G mobile communications is currently more in the marketing phase than the technology phase, and it is still unclear what direction it will take. The technology is about data rates and bandwidths. Another important topic is latency. This means reaction times should be as short as possible. But all this will be a gradual development process. Probably, mobile network operators will first improve their established networks. They will optimize frequency use and then gradually work their way toward higher frequencies. There is currently a lot of talk about 60 GHz. But that is for indoor applications, as such frequencies have a very limited range. Outdoors, we will likely be dealing with 20 GHz or 40 GHz, depending on which frequencies the regulator makes available. 60 GHz is more likely to be the last phase of development. There is still a lot of research and development to be done in this area.
EETE: 5G will also enable new network structures, and connections won’t necessarily be organized hierarchically. More importantly, it will be possible to form ad hoc networks. What is that about?
Steffen: At Rohde & Schwarz, we are concentrating on the air interface, but the challenges of 5G are behind the base stations. These networks will transmit massive quantities of data that have to be processed. That is why there will be a very well networked system behind the base stations. As far as the technology is concerned, the antennas will certainly be an issue at higher frequencies. There is currently a discussion about MIMO (multiple input, multiple output). Mobile phone manufacturers already have several antennas in their devices. The higher the frequencies, the smaller you can make the antennas. That takes us into the field of massive MIMO and beamforming. A mobile phone might have 64 antennas that generate a controlled antenna beam. The antenna would orient itself toward the access point based on where the user takes the device. This creates new challenges in terms of testing. Up to now there has only been one antenna and one antenna cable and that is where the tester is connected. It is unclear what this would look like with 64 antennas.
EETE: How could it look from a technical perspective?
Steffen: Such devices are currently measured with a robotic arm that moves along the antenna beam. That is very time consuming. You can of course pack a large number of sensors into a test chamber, but that is no less expensive. If you have a spherical test setup, you of course need the same number of receivers as sensors. Plus there are many challenges in production, development and research. In the research phase there is no time pressure when taking measurements so you might be able to afford a robotic arm, but it’s a different story in production.
EETE: In addition to 5G, one of the most prominent current trends in radiocommunications is the Internet of Things (IoT), which frequently runs via wireless connections and will require new wireless communications standards. How does this impact your business? Does it open up new business opportunities for you?
Steffen: Considering that there are more machines than people on this planet, IoT should actually boost our business. However, the modules that are to be used for this are often based on technologies that are already available. GSM, for example, is still very widespread. IoT applications typically communicate only limited data volumes at low data rates. Smart meters are a good example. They transmit the status of electricity meters every once in a while, and a few bytes suffice. This also applies to vending machines that report when they need to be restocked. That also doesn’t require large data volumes. Today, the Internet of Things consists mostly of applications that operate occasionally and tend to transmit small data volumes.
EETE: There are special networks for this, such as Sigfox.
Steffen: Many services that currently feature proprietary wireless technology will be implemented with 5G in the future. Mass production makes powerful platforms very affordable; proprietary technology is always costly. That is why applications with direct connections between user devices could be better implemented with LTE in the future. The standard already provides for this; but it is not well supported by network operators. The automotive industry is also pushing into this field with car to X (C2X) applications. You need very low latency when security related C2X services come to the point of being able to transmit to their environments, for example when a car reports, "Caution, I’m braking". In this case, the car should not have to first establish a connection with its network operator. The technologies currently under development with 5G would be able to do that. But we will have to wait to see what the business models look like. We expect that more applications will be implemented with 5G technologies than is possible in mobile communications networks today. It remains to be seen, however, how many of these functions will ultimately be available in mobile phones.
From a test and measurement standpoint, it is more of a technology issue. We have to handle bandwidths and latencies and implement what the standardization bodies specify. These are matters of modulation modes and physical levels. The physical layer is being redefined, and we naturally have to implement this in our products as well. But we already have T&M instruments that work at 60 GHz and can generate and receive the code. Seeing that the market is very small at the moment, everything is still relatively costly. But that will change. A return to expensive lab instruments is out of the question for production test and measurement equipment. But we have five or six years to develop suitable solutions before mass production begins.
EETE: Rohde & Schwarz has a branch that deals with security technology. Could this produce synergies associated with IoT? Security, encryption and authentication are also being intensely discussed in the IoT world.
Steffen: There are definitely synergies among our various activities, but it is not so much a matter of being able to offer a "security measurement instrument". For example, network operators have asked us to analyze exactly what smartphones apps do. Many apps communicate in the background. The question is what transmission bandwidths they use and how much network capacity they consume. The individual apps are not really coordinated and naturally require a certain transmission capacity. Even when only a few bits are transmitted, the need to connect and disconnect many times consumes a lot of resources. Network operators need to know what is happening. App behavior is also important to IT security personnel. They want to know which servers and services these apps connect to and what data they transmit.
The extensive expertise that is compiled in the Rohde & Schwarz group of companies has gone into the software in our products. It is possible to determine what kind of data a mobile phone is currently transmitting and via which servers without involving the phone user. There is demand for this on the market, for network traffic classification or how to detect anomalies, for example. The apps that tax the battery the most can be identified. That of course goes beyond the boundaries of conventional radiocommunications test equipment.
EETE: Government security agencies are now combating cybercrime. Does this open up new business opportunities for you? For instance, detecting apps that install malware without users knowing it?
Steffen: I don’t want to claim that our instruments will tell you that, but they do offer the possibility to view such processes in the first place. There are also unmeasurable aspects that play a role in analyzing measurement results. For example, the question of whether you have a "good" or "bad" server and where it is located. We simulate base stations and allow instruments to connect to the Internet in order to analyze data traffic. A user can assess how little or how much data an app is using and see how actively it is transmitting.
EETE: When designing T&M instruments, there is a visible trend toward moving control and evaluation logic to intelligent devices such as computers. This shrinks physical space requirements to fit in small, very handy and transportable devices. What is your view on this?
Steffen: It is beyond dispute that the computer world has a higher speed of innovation than conventional T&M equipment. Lab test equipment has a lifespan of seven to ten years. A computer lasts two to four. By choosing a solution such as you describe, you can certainly improve the performance and lifespan of a test instrument. That makes sense where speed and automated processes are concerned, such as in production. You can also increase performance by outsourcing certain algorithms to a high-performance computer. We extensively use solutions like these in the production environment. It’s a different story in the mobile field. For example, it is of course possible to separate the functions of a battery fed handheld device into a small test module and a tablet. But then technicians would have to carry two devices and handling is less convenient than with a conventional device. Our customers tell us that they prefer an instrument that includes all required features. That is why concepts like these do not play a role in handheld devices.
EETE: So there will never be a Rohde & Schwarz oscilloscope app?
Steffen: You should never say never, but at the moment I do not believe this is a business case. Of course, it is cool when you can operate your tester from a smartphone or tablet, but is it practical? We’ve already had the discussion about whether it’s practical to voice control a test instrument and have it voice report the measurement results. Now imagine a lab with fifteen employees, and each of them has a talking tester. And they’re all voice reporting at the same time. That may be fun, but I don’t see any value for the user.
EETE: But software is easier to upgrade than hardware. So if I have a new test sequence that might require another type of analysis, it might be actually easier to implement it with an app than by changing my hardware.
Steffen: No, that isn’t necessarily the case. At Rohde & Schwarz, we spend around 50 percent of our development costs on a product after it is launched. It’s no longer true that hardware is inflexible. Of course, analog hardware in the frontend can only do so much, but there are DSPs and FPGAs for analysis, and the instruments have built in computers with hard disks. We update the software every three months. We offer a software option if, for example, a customer wants to work with a new wireless communications standard and the instrument has the technical prerequisites. You could, of course, call this an app, but we stick to the traditional term of software option. These options make it possible to custom configure instruments without overloading them with expensive features. A product can have up to 100 different options over its lifespan. That is a key component of our business model. Only those customers who need these features have to pay for their development.
Of course, there are T&M equipment vendors that disclose the internal structure of their products to allow customers to develop software. We steer clear of that for a number of reasons. On the one hand, support is very costly. On the other hand, we think that we can make at least the standard applications better and cheaper because we know the instrument much better than the customer does. We can also distribute the cost of development over many customers.
EETE: What is your opinion on uploading test analysis to the cloud? There are possibilities for evaluating data via cloud models: for example, production testing in the semiconductor industry. They say that the advantage is better application scalability.
Steffen: Our observations tell us that many companies are quite cautious about these sorts of offers. Small companies may not be able to afford to analyze their measurement results in-house, but I would be very surprised if a large semiconductor manufacturer would let a service provider evaluate its production measurement results. All companies carefully protect their quality data. That is why I don’t believe that many companies would allow their confidential business data to be stored and processed by a service provider.
Then again, what we already have, and what I also believe has a future, is networking test instruments and integrating them into IT infrastructures. We have our own development databases for storing results and conducting centralized analysis. Remote access to test equipment is also useful for customer service. Technologies such as Remote Desktop allow you to display the results from a test instrument in Taiwan right in the developer’s office. Functions like this are standard today, at least when it comes to higher performance T&M equipment. However, this usually requires a bilateral agreement with customers since you are active within their IT network and have to transmit data out of it.