Ericsson: 5G migration won't be a terrifying slog. No. We have ‘plug-ins’
Perfect for finding MIMO... and also RAN
Analysis All the network equipment providers are engaged in major operator projects which they hope will guarantee them a place in those MNOs’ 5G rollouts in the coming years. This week, it was certainly the turn of Ericsson to score 5G marketing points, with a series of operator engagements around the world, and the announcement of 5G Plug-Ins.
These appear to tick the mobile operators’ key boxes – to boost capacity and efficiency in current networks, without waiting for new standards, but to future-proof those networks so the migration to 5G is a painless one. By calling them plug-ins, a term that usually applies to web applications, Ericsson made its offerings sound simple, modern and unthreatening. They are “network software - a way to deploy 5G technology concepts on today's networks, while preparing them for 5G use cases and deployments," said the company.
But in fact, it claims these plug-ins will, among other things, enable operators to support Massive multiple-input and multiple-output (MIMO) radio antenna arrays and Multiuser MIMO as early as 2017 – surely technologies which will require more than just a painless software upgrade. The Massive MIMO Plug-In is “a combination of Single-User MIMO (SU-MIMO) and beam-forming* supported by advanced antennas with a large number of steerable ports,” says the product description, which suggests a fair dose of hardware and engineering.
Joachim Bildt, head of product marketing at Ericsson’s radio business unit, told Telecom.com: “With 5G we’re moving from the network-focused concept of cells to the user-focused concept of beams. Today’s cellular network sends out energy in all directions while beams concentrate this energy, sending the signal more directly to the user.”
The other plug-ins, all of which will be in tests this year and commercially available next year, are:
- Multiuser MIMO Plug-In: “Transmits data to multiple user devices using the same time and frequency resources and coordinates beam-forming”, extending the Massive MIMO effect still further.
- RAN Virtualization Plug-In: "Improves network efficiency and performance by enabling Virtual Network Functions to be centralized on a common platform supporting both 4G and 5G."
- Intelligent Connectivity Plug-In: Where 5G and 4G coverage areas overlap, the network can "intelligently route data based on application requirements and network resource availability, increasing the combined data throughput of 4G and 5G resources”.
- Latency Reduction Plug-In: “Shortens access procedures and modifies the frame structure to enable instant network access and more frequent transmissions. This in turn reduces time-tocontent while enabling real time communications for key 5G applications such as smart vehicles."
All these are central attributes of 5G networks, as far as those are understood in a pre-standardized period, but they are also goals for many MNOs within their current LTE-based deployments. Virtualization, IoT support, flexible use of all available resources, greater capacity – these could enhance the business case now, if they can be achieved in 4G. Ericsson’s plug-ins do illustrate how the value and efficiency in the network is increasingly delivered by software, but radios and antennas, even when software-controlled, remain pieces of hardware which are difficult to deploy and optimize.
Ignoring the physical aspects of networks in an age of virtualization and software is not helpful because, like the ‘standards-ready’ claims of proprietary products through the ages, it promotes an illusion that deploying a new network architecture can ever be simple – and risks a backlash from operators.
Yet the delicate trick which the major network OEMs need to pull of this year is to promote themselves as trailblazers in 5G, even before core standards are defined, while not deterring operators from putting 4G expansion on hold while they wait for the next generation.
This entails the tried and tested ‘xxx-ready’ approach – if you buy equipment now, it will improve your networks and your economics immediately, and will then be smoothly upgradeable to the next generation when that is available.
Is a ‘5G-ready’ approach believable?
In the past, this has usually turned out to be a false promise. One of the WiMAX industry’s fatal mistakes was to pretend that all kinds of proprietary broadband wireless systems were "WiMAX-ready" when they were not, and then to introduce a second WiMAX standard, 802.16e, and pretend that it was an easy matter to upgrade, in software, from the earlier, incompatible one. Similarly, promises of simple transitions from 3G to LTE were illusory.
The "pre-standard" argument only really works when that standard is so close that only minor adjustments are likely to be made – WiFi equipment is often released ahead of the final ratification, with few, if any, modifications needed to be fully compliant.
The 5G situation is somewhat different. This time, vendors have to deliver, not just promise, a smooth transition from 4G to 5G, because operators will not tolerate another big bang upgrade. Instead, they need to introduce new capabilities one by one, according to their business requirement, so that a current HetNet gradually morphs into one with 5G characteristics.
Approaches like Ericsson’s plug-ins, once the marketing hype is stripped away, do help address this challenge. In the vendor’s words, its new offerings “are software-driven innovations supported by the Ericsson Radio System, and are specifically focused on capabilities that operators can leverage within current networks to facilitate their evolution to 5G.” Useful certainly, because if the RAN does not evolve at the same rate as other parts of the network – the core and the IT platforms, where virtualization is being applied rapidly – the operator will not be able to reap the value of these changes.
Most of the work of delivering that kind of evolutionary approach does rely on software. For instance, small cells will individually become extremely cheap now that economies of scale are kicking in, but densification will require huge numbers of them, so total cost will still be significant. It is the software which will make them affordable.
Virtualizing most of the network functions on shared controllers will allow the access points to be reduced to something very simple and costing a few hundred dollars; SON (selfoptimizing network) will slash the operating cost of planning and tuning large numbers of cells.
New hardware is important too
But while 5G will almost certainly center on increasingly dense networks of virtualized small cells, there will also be evolution in the macro layer in order to support wide area coverage and full mobility. The macrocell is where the OEM gets its richest revenues and the vendors will all be keen to push ‘SuperMacro’ technologies like Massive MIMO, as indicated in Ericsson’s choice of plug-ins.
Vendors still need to shift expensive hardware, and in the age of virtualized RANs, that means increasingly complex antenna/radio units at the cell site, and computing hardware from central servers to base station processors, to support virtualized network functions, to deliver cloud services, and to process data.
Massive MIMO will be an important element of 5G and even some pre-5G deployments, but it will not be simple. ZTE has been a leader here and this week it announced the recent completion of verification and performance tests for Massive MIMO technologies in Shenzhen, under the auspices of the Chinese IMT-2020 (5G) Promotion Group, as part of its three-year R&D plan. The tests, said ZTE, showed that with Massive MIMO, cell throughputs can be increased exponentially and downlink gains can be up to three times greater than on conventional macrocells, while uplink can be four times higher. ZTE says it is the only vendor to perform Massive MIMO uplink tests so far.
No escape from premium hardware then, just new areas of focus – as Nokia’s AirScale platform, integrating RAN and IT platforms, highlights, as do Ericsson’s and Huawei’s increasingly aggressive moves into hyperscale cloud platforms for carriers, and the territory of Hewlett-Packard.
Copyright © 2016, Wireless Watch
Wireless Watch is published by Rethink Research, a London-based IT publishing and consulting firm. This weekly newsletter delivers in-depth analysis and market research of mobile and wireless for business. Subscription details are here.
* using the way that signals cancel each other or reinforce each other to shape different beams towards different clients of the same base station. Read more about the technologies that support 5G standards here