In this post we present the TitanMIMO: A 100×100 Massive MIMO Testbed Based on xTCA Standards. The content of this blog is taken directly from our new whitepaper which you can download here.
Massive MIMO Can Help Achieve the 5G Goals of Tomorrow
Demand for wireless data connectivity is outpacing available capacity that can be supplied by today’s networks, and the gap will only increase as we seek to connect tens of billions of wireless devices over the coming years.
To help address this expected capacity deficit, the 3GPP standards body has set an aggressive 5G target for a 1000x capacity increase by 2020. Massive MIMO (Multiple Input Multiple Output) is a popular potential enabling technology to simultaneously increase capacity and peak data rates, while reducing energy consumption and latency.
Conventional MIMO solutions are already being deployed in networks today, typically using from 2-16 antennas with RF bandwidths up to 20 MHz. The 5G Massive MIMO approach proposes a tremendous increase in the number of antennas at a base station, potentially scaling to arrays of 100×100 or more, while supporting RF bandwidths of 100 MHz or greater.
Throughput Requirements of a Massive MIMO Testbed
Theoretical research into Massive MIMO has been rapidly advancing over the past several years; however there hasn’t been a hardware testbed that would allow researchers to accurately prototype a complete 100×100 system when dealing with over-the-air (OTA) bandwidths on the order of 20 MHz or greater.
This is in large part due to the fact that in order to fully realize the potential of Massive MIMO, baseband processing implementations of the algorithms will typically require the aggregation of all 100 RF channels into a single common processing element.
This ultimately dictates the required over-the-air (OTA) RF-to-baseband processing system throughput of a testbed, which will limit the combined maximum number of supported channels and real-time, OTA bandwidth that can be handled by a given system architecture.
To better understand the throughput demands which must be met by Massive MIMO testbeds, Table 1 illustrates the OTA RF-to-baseband processing throughput required for an 8×8 antenna 4G/WLAN MIMO system with 20 MHz bandwidth, as well as throughput calculations for a 100×100 Massive MIMO testbed with today’s 4G/WLAN radios and with the 5G/802.11AC radios of tomorrow (100 MHz bandwidth).
Table 1: OTA RF-to-Baseband Processing Throughput Requirements, 4G/WLAN & 5G/802.11AC Massive MIMO Testbed
Current Bus Technologies Cannot Meet Massive MIMO Throughput Requirements
Fixed, shared backplane bus technologies like PCIe or network infrastructures such as GigE have been very well suited to various 3G & 4G physical layer data transfer requirements. However, they require a master switch to manage all of the data coming from the N different radio front ends, and this N:1 aggregation can impose a significant physical constraint on throughput as the radio count scales to tens or hundreds of units.
For the 5G/802.11AC Massive MIMO data throughput requirements of 240 Gbps and above, it can be readily seen from Table 2 that another data transfer solution will be required to break the bottleneck which currently impedes real-world, wideband, very large antenna array testing.
Table 2: PCIe & GigE Theoretical Throughputs
Processing Requirements of a Massive MIMO Testbed
As the physical layers which would benefit from Massive MIMO technology are neither standardized nor implemented, we cannot know today the exact computational resources which will be needed for a 100×100 testbed configuration.
To ensure the future of a Massive MIMO testbed aiming to support 5G, while maximizing the return on today’s investment, it is important to avoid central baseband processing limits, such as would be imposed with an architecture that solely supported a single processor without a path for expansion.
It is important to note that when considering central baseband processing engine expansion, upgrading isn’t achieved through simple addition of more processor cards in a backplane infrastructure. As noted previously, each added processor within the central baseband processing unit must have access to information from all RF channels in order to realize the optimal Massive MIMO algorithmic implementations. This leads to an inter-baseband processor constraint related to the OTA RF-to-baseband throughput.
RF Requirements of a Massive MIMO Testbed
In order to allow Massive MIMO developers to begin real-world prototyping with today’s 4G/WLAN radios, while supporting future 5G radios at higher frequencies and bandwidths of 100 MHz or greater, the RF transceivers of a Massive MIMO testbed need to be easily replaceable. Additionally, upgrading of the radio should have minimal impact on the baseband processing related hardware, in order to maximize ROI and retain as much existing IP as possible.
The content of this blog is taken directly from our new whitepaper which you can download here.