The move to LTE will require devices and network equipment that perform perfectly from the first deployment, despite large increases in complexity and functionality. Spirent’s Nigel Wright tells Keith Dyer how the industry can assure the quality of…
The move to LTE will require devices and network equipment that perform perfectly from the 
first deployment, despite large increases in complexity and functionality. Spirent’s Nigel Wright tells Keith Dyer how the industry can assure the quality of the next wireless generation
Keith Dyer:
Nigel, you are marketing director for Spirent’s wireless business. Can you remind us how Spirent fits into the wireless world, and your role in helping the industry develop and bring products and solutions to market?
Nigel Wright:
Of course. In Spirent’s Performance Analysis Division, our Wireless business has products across two main areas. The first is on the device test side, carrying out performance testing of CDMA, UMTS and LTE devices. We also have a Fader product for receiver testing, which is popular with infrastructure manufacturers for next-generation base station testing, among other applications. Our Broadband business carries out performance analysis and testing of broadband networks and services, including core networks for wireless. One of our Broadband product families tests wireless packet core network elements – for standards compliance, performance, scalability and interoperability.
Keith Dyer:
And what do you think are your growth drivers, whether in Europe or in other markets?
Nigel Wright:
One of the biggest current drivers has been the surge in smartphone market share. The big name players continue to increase the level of competition in this sector, with new launches and functionality. But there were some high-profile performance mishaps with early smartphone launches that manufacturers and some operators are anxious not to repeat, and that helps a test vendor like us.
Our test solutions help to identify and analyse performance issues that can plague device users, such as frequent dropped calls. A recent report from Broadband Testing here in Europe used our equipment to do call reliability testing on some leading European-market smartphones, and found significant differences between devices in something as fundamental as their voice call reliability.
As more complex applications reside on the phone, so the interaction between the device and the network goes up significantly. An application like push email is continually communicating with the network for updates, meaning that the device has to be able to handle that data interaction in the background during voice calls to avoid call drops. Service interaction issues like this are only going to increase as more applications are embedded and downloaded to these complex devices.
Keith Dyer:
We’ve seen in the US operators take a very hands on role with device performance testing, while in Europe operators are more aligned to a certification-based model. With the increasing complexity you have mentioned, do you think one model is more likely to ensure device performance, and therefore customer satisfaction and retention.
Nigel Wright:
Yes, the US has a stronger model of operators working in an aggressive role to make sure that the devices they sell perform well. Verizon plays a strong role, and AT&T has made similar moves, as have some others. We are also seeing this model extend to some extent to Japan and China. In Europe we still see the device manufacturers remaining responsible for producing certified devices. Operators then focus on testing with their own infrastructure and networks, as well as ensuring that their own applications and portals run well. This model in Europe has not changed that much in recent years.
The problem is that, as LTE comes along, the current scope of device testing under that model is unlikely to be enough – even though devices will have passed certification testing there will be so many variables affecting their performance on real networks with real applications. There’s already a recognition of this today, with 3G devices – it’s part of the reason that US operators drive their own device performance requirements. However, many more operators realise there are poorer-performing devices out there that have higher return rates. And performance differences are only likely to increase with HSPA+ and LTE. Operators will need to prioritise what to do about this, detailing their requirements in terms of their own application portfolios, before they can make a business case.
Keith Dyer:
So how can the industry make that business case?
Nigel Wright:
One approach is to wait for broader standardised performance testing. Some organisations are looking at producing industry recommendations in this area, including the GCF for example, but this process will take a considerable time and there are many areas that require detailed performance testing.
In the meantime, although some operators are already starting to use headline LTE maximum data rates in their publicity, it’s clear that this is a risky strategy, since there will be large differences between headline rates and what subscribers will typically experience. These differences will depend on the network topology, channel bandwidth and frequency bands that operators use. In Nordic countries they plan to use 2.6 GHz spectrum for initial LTE rollouts, while in the USA the 700 MHz band will be used. The new spectrum used for LTE may mean devices will eventually need to support 10-12 bands or more for global roaming. Trying to optimise the breadth of device performance that will be required going forward is going to be a big challenge – something that Spirent’s experience can play a huge part in.
You also have the critical issue of new air interface techniques that come with LTE. MIMO is intended to be a major contributor to LTE data performance but by its nature the technology requires multiple antennas in a device. This means that complex physical designs and algorithms are needed for MIMO to work well across multiple bands, especially in a handset form factor.
Then there are other network issues such as backhaul limitations, and the fact that LTE, like HSPA before it, shares its data pipe across multiple subscribers. Going forward, many of these subscribers will be passive users, as their resident smartphone applications constantly poll the network, quietly chewing up bandwidth.
So the industry will need to find a way to get past this confluence of issues if LTE is to deliver on its performance and business objectives.
Keith Dyer:
So where would you place the industry in terms of how it is getting to terms with these issue? After all, the first LTE networks are slated to go live next year.
Nigel Wright:
So right now, we have more than two years’ work with the LTE R&D teams at network infrastructure manufacturers under our belt, during which we have helped to work through many of the network issues. For MIMO, our fader solutions have been testing the performance of multiple antenna arrangements. We can simulate accurately and repeatably the multipath fading that the OEMs need to use to optimise their designs and algorithms for MIMO.
Our Landslide product has also been very involved in LTE infrastructure testing, to help ensure standards compliance in all the little grey areas that become evident when systems try to interoperate, especially as the specifications continue to evolve. So there is an interoperability path for the ENodeBs, serving gateways, MMEs and other network elements from different vendors, that takes into account slightly different interpretations of standards that always arise when implementing a complex new technology. Beyond these issues, we also help test the performance and scalability of the evolved packet core networks that will power LTE.
On the device side, I’d say things are ramping up now, though we’ve been involved with chipset development for a while. For the earliest releases of LTE in the US, there are specific interference challenges at 700 MHz that must be tested for, since public services and digital TV use adjacent bands, and the second harmonic of an LTE device’s transmitter falls in the GPS band. Because we’re a major player in CDMA device testing we are also able to provide early support for testing the critical device handover performance between LTE and CDMA networks.
Another important consideration on the device side is over the air, or OTA, test capability. Regular lab-based conformance and performance testing is carried out over a cabled connection from the test solution directly to the device antenna port, bypassing the antenna itself. But with MIMO, antenna performance quickly become the limiting factor that will dictate how well the device performs overall, so performance testing needs to include the antenna. You can do this in two ways. The first is in a drive test – but this can make it very difficult to repeat issues you uncover so you can pinpoint the root cause, because with drive testing the environment is constantly changing.
The alternative is lab-based OTA testing, which is carried out in an anechoic chamber fitted with reference antennas, allowing you to quantify the performance of the device’s antenna when it’s transmitting or receiving. However, when compared to current technologies supporting LTE and MIMO with a chamber-based approach requires multiple additional antennas and faders, and the resulting cost and complexity will limit the total market.
So while we’re involved in this type of OTA testing, we also a leader in research into alternative methods, such as virtual OTA testing. With this approach, the device’s antenna characteristics need to be established only once in a chamber, then you use that data in a lab-based device performance test solution to do ‘virtual OTA testing’, which can give a very good idea of how well the whole device will perform in the real world. In the future this approach could help operators reduce their need for drive testing – which of course could reduce the total time and cost for them to bring a quality device to market.
So despite the challenges, there is a huge effort across the industry to make sure that there is a holistic approach to make LTE successful. Verizon, for example, is really committed to putting its money where its mouth is and to driving a successful initial rollout.
Keith Dyer:
Do you think that the approach European operators are taking is to wait a while, to extract what they can from HSPA and then to move to LTE when some of these issues have been resolved?
Nigel Wright:
That seems to be the approach that the bigger European operators are taking. As you say, many are still looking to maximise their return on investment in UMTS technology, with some going to HSPA+ as an intermediate step. There are still major issues with spectrum availability, which are not settled in many major markets. I think the other open issue for LTE in the longer term is support for voice. Whether the answer is IMS or one of the alternative proposals, I think this is another issue that has caused some to hesitate.
Keith Dyer:
So while you work on LTE, and perhaps wait for some larger European operators to enter the market, what else is driving business for you at the moment?
Nigel Wright:
One area to watch is GPS. Sales of GPS-equipped handsets have quietly trebled year on year, and at last LBS are becoming widespread. The standardization of Assisted-GPS, or A-GPS, has made it easier for users to get the technology’s benefits, but the performance of some devices has been disappointing because of poor GPS antenna implementations. The CTIA in the US released a new OTA test plan for A-GPS this past April to take into account GPS antenna performance. The 3GPP/GCF are looking at a similar approach – which should help to drive improved performance of LBS. We’re an industry leader in A-GPS testing, so these new test methodologies will drive this market for us and our customers.
The other driver in this market results from recent Russian investment in its competitive system to GPS, called GLONASS. Receivers which can use GLONASS as well as GPS effectively double the number of satellites in the sky that they can use to obtain a position fix, which can greatly improve performance in urban locations. So I expect to see the first GPS/GLONASS receivers appear in cellular devices by the end of 2010. The ramp-up in the use of multiple global navigation satellite systems is another driver for us and for LBS market success in general.