I was recently thinking a bit about the 5G NR Control Channel inside the CORESET and wondered what its datarate would be. This seems to be a bit of a theoretical exercise at first. On the other hand, the Control Channel is a significant portion of the channel overhead which is lost for user data. In LTE, the PDCCH spans the complete channel bandwidth and at least one symbol. Perhaps reasonable for a maximum channel bandwidth of 20 MHz. However, I expected that on an 80, 90 or 100 MHz 5G NR carrier, the control channel would not span the complete channel bandwidth. I was quite surprised however, when I had a closer look.
Jaeku over at Sharetechnote has a nice overview of where to find the information about how many RBs are used for the downlink shared channel in an RRC-Reconfiguration message when the 5G channel is added to an LTE connection. The relevant parameter is frequencyDomainResources in the ControlResourceSet section. And to my big surprise the parameter indicated that pretty all RBs on the frequency axis are used for the channel.
ControlResourceSet frequencyDomainResources: ffffffffff00 [1 bit represents 6 RBs] duration: 1 [symbol]
In other words, since there are 14 symbols in a slot on band n78 with a sub-carrier spacing of 30 kHz, 1/14 of the channel capacity is used for the control channel, which is around 7%. Or in terms of bandwidth:
As each each Resource Block contains 12 symbols on the frequency axis, the CORESET contains 240*12*2 = 5760 bits in the example above. Based on the downlink slot duration of 0.5 ms, the raw datarate of the channel that assigns downlink and uplink resources for all devices in the cell is thus (4/5 (DDDSU) * 5760 bits) / 0.0005 s = 9.2 Mbit/s*.
O.k. so the overall channel with 256QAM and 4×4 MIMO can yield more than a Gbit/s for user data, but still, 9 Mbit/s for just assigning uplink and downlink resources!? Wow!
* Yes, I know, CCEs are grouped to further increase redundancy which lowers the actual data rate of the channel, but I’ll stick with the raw datarate in this post.