ICONIC centre is in the process of building extensive experimentation infrastructure that aims to integrate end emulate complete end-to-end open source IoT/Edge/Cloud environment.
The infrastructure consists of radio-access layer in the form of three different testbeds:
Software-Defined Radio testbed:
This testbed consists of 16 USRP nodes (8 NI 2901 and 8 NI 2920) which are flexibly used in different radio access configurations. Using open source components such OpenAirInterface, we are able to emulate lab version of 4G LTE network and make experimentations across the 3GPP LTE protocol stack, where core network components reside in virtualized environment of our Cloud testbed.
Our testbed is currently evolving to major LP-WAN technologies: 3GPP NB-IoT and LoRaWAN. For NB-IoT, in joint project with a mobile operator, we designed and fabricated a testbed of 100 NB-IoT nodes
that are distributed across the FTN building to support Smart Building concept experimentation. Another 50 NB-IoT nodes are currently in fabrication for mobility applications such as smart logistics. For LoRaWAN, we are i
n the process of installing Kerlink base station to support future LoRa node design, development and experimentation. All the data collected from NB-IoT/LoRaWAN nodes are directed via LP-WAN networks to our Cloud testbed (see below).
UAV Communications Testbed: ICONIC centre is using DJI Matrice 600 Pro drone to experiment with UAV-based communication systems by on-board software-based 4G/5G/NB-IoT equipment.
IoT/WSN testbed:The lab possesses more than 100 Crossbow sensor boards of different generations.
The edge computing infrastructure is evolving towards two different testbed implementations:
Mobile Edge Computing (MEC) testbed:
In MEC testbed, we integrated general-purpose server capabilities in close proximity of radio access layer to experiment with low end-to-end delay IoT applications. In addition, we use hardware-accelerated (FPGA) nodes to implement and test different information processing and machine learning algorithms in the context of edge computing.
Distributed computing testbed:
To experiment with large-scale distributed information processing algorithms, we use emulation environment containing large number of Raspberry Pi nodes connected to a full-mesh network via high-speed networking equipment. This testbed serves for testing distributed optimization and inference algorithms under emulated communication constraints and impairments.
The Cloud computing and storage testbed consists of the single testbed:
Cloud testbed is the core of ICONIC testbed infrastructure. The testbed primarily serves to accumulate local computing capacity in the form of general-purpose servers (including many-core HP server containing 168 cores), graphical processing stations with graphical processing unit (GPU) cards, etc. The testbed is used for experimentation in the domain of software-defined networking (SDN), e.g., using Open Flow, network function virtualization (NFV), and cloud-based services enabled by open-source tools such as Open Stack.