Hello!
Yes, you can convert the 400Gbps ports to 100Gbps ports simply by inserting a 100Gbps QSFP28 transceiver/optic; no breakout cable is needed. To remove any doubt, we have a N9K-C93600CD-GX in our lab environment running NX-OS 10.5(1):
N9K-1# show module
Mod Ports Module-Type Model Status
--- ----- ------------------------------------------------ --------------------- --------
1 36 28x100/40G + 8x400G QSFP-DD Ethernet Module N9K-C93600CD-GX ok
27 0 Virtual Supervisor Module N9K-C93600CD-GX active *
Mod Sw Hw Slot
--- ----------------------- ------ ----
1 10.5(1) 3.0 NA
27 10.5(1) 3.0 VSUP
Mod MAC-Address(es) Serial-Num
--- -------------------------------------- ----------
1 54-51-de-f9-fb-10 to 54-51-de-f9-fb-df FDOREDACTED
27 54-51-de-f9-fb-10 to 54-51-de-f9-fb-df FDOREDACTED
Mod Online Diag Status
--- ------------------
1 Pass
27 Pass
* this terminal session
A 100Gbps transceiver (specifically, a QSFP-100G-SM-SR, which leverages single-mode duplex fiber with LC connectors; put another way, no breakout cable is used here) is inserted in Ethernet1/29, which is one of the 400Gbps uplink ports on the switch.
N9K-1# show interface Ethernet1/29
Ethernet1/29 is up
admin state is up, Dedicated Interface
Hardware: 10000/25000/40000/50000/100000/200000/400000 Ethernet, address: 5451.def9.fb17 (bia 5451.def9.fb98)
MTU 1500 bytes, BW 100000000 Kbit , DLY 10 usec
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA, medium is broadcast
full-duplex, 100 Gb/s, media type is 100G
Beacon is turned off
Auto-Negotiation is turned on FEC mode is Auto
Input flow-control is off, output flow-control is off
Auto-mdix is turned off
Rate mode is dedicated
Switchport monitor is off
EtherType is 0x8100
EEE (efficient-ethernet) : n/a
admin fec state is auto, oper fec state is Rs-fec
Last link flapped 00:31:44
Last clearing of "show interface" counters never
1 interface resets
Load-Interval #1: 30 seconds
30 seconds input rate 0 bits/sec, 0 packets/sec
30 seconds output rate 0 bits/sec, 0 packets/sec
input rate 0 bps, 0 pps; output rate 0 bps, 0 pps
Load-Interval #2: 5 minute (300 seconds)
300 seconds input rate 40 bits/sec, 0 packets/sec
300 seconds output rate 40 bits/sec, 0 packets/sec
input rate 40 bps, 0 pps; output rate 40 bps, 0 pps
RX
0 unicast packets 34 multicast packets 0 broadcast packets
34 input packets 10812 bytes
0 jumbo packets 0 storm suppression bytes
0 runts 0 giants 0 CRC 0 no buffer
0 input error 0 short frame 0 overrun 0 underrun 0 ignored
0 watchdog 0 bad etype drop 0 bad proto drop 0 if down drop
0 input with dribble 0 input discard
0 Rx pause
0 Stomped CRC
TX
0 unicast packets 34 multicast packets 0 broadcast packets
34 output packets 10812 bytes
0 jumbo packets
0 output error 0 collision 0 deferred 0 late collision
0 lost carrier 0 no carrier 0 babble 0 output discard
0 Tx pause
Here's further evidence that a QSFP-100G-SM-SR is inserted in this port.
N9K-1# show interface Ethernet1/29 transceiver
Ethernet1/29
transceiver is present
type is QSFP-100G-SM-SR
name is CISCO-INNOLIGHT
part number is TR-FC13H-HCI
revision is 1C
serial number is INLREDACTED
nominal bitrate is 25500 MBit/sec per channel
Link length supported for 9/125um fiber is 2 km
cisco id is 17
cisco extended id number is 220
cisco part number is 10-3220-02
cisco product id is QSFP-100G-SM-SR
cisco version id is V02
CDP shows us that on the remote end of the switch, we have another N9K-C93600CD-GX with a remote interface of Ethernet1/29 as well.
N9K-1# show cdp neighbors interface Ethernet1/29
Capability Codes: R - Router, T - Trans-Bridge, B - Source-Route-Bridge
S - Switch, H - Host, I - IGMP, r - Repeater,
V - VoIP-Phone, D - Remotely-Managed-Device,
s - Supports-STP-Dispute
Device-ID Local Intrfce Hldtme Capability Platform Port ID
N9K-2(FDOREDACTED)
Eth1/29 156 R S s N9K-C93600CD-GX Eth1/29
Total entries displayed: 1
With respect to your second question - Generally speaking for modern Nexus switches running NX-OS, you can safely use one of the "uplink ports" as an access or downlink port without any loss in features or functionality (and vice-versa - use a downlink as an uplink). Depending on the specific model of switch and the specific use case you're working with, there may be scenarios where a feature does not work on an uplink, but does work on a downlink (or vice-versa).
To use an arbitrary example, let's say you're working with an N9K-C93180YC-FX3 running NX-OS 9.3(14). You may have a use case where you need to connect an uplink port (e.g. Ethernet1/49, a 40/100Gbps QSFP28 port) to a WAN (like a dark fiber or DWDM circuit) at 10Gbps by using a QSA (CVR-QSFP-SFP10G). In this scenario, you cannot use a downlink port because all of them are already in use elsewhere to connect downstream servers or network devices. Additionally, you must satisfy a business requirement dictating that any traffic over a WAN must be encrypted, so the use of MACSec is required in this instance.
In this hyper-specific instance, you would encounter the following documented limitation:
"Cisco Nexus 9000 Series switches do not support MACsec on any of the MACsec capable ports when QSA is being used."
Luckily, this limitation is lifted for the N9K-C93180YC-FX3 starting with NX-OS 10.1(2), so a software upgrade is all that is needed to resolve this blocker:
"Beginning with Cisco NX-OS Release 10.1(2), MACsec is supported by Cisco Nexus 9300-FX3 platform switches when QSA is being used."
However, not all limitations like this can be worked around through software; some are rooted in hardware limitations (the design of the underlying ASIC, the length of circuit traces between the ASIC and MACs/PHYs of specific ports, etc.).
In any case, this is a niche example that demonstrates why I would say that generally speaking, you will see feature and functionality parity between downlinks and uplinks. You will typically see any limitations surrounding the lack of parity documented under the Guidelines and Limitations section of a given feature's configuration guide. In your case, since you're looking for a spine switch for an EVPN fabric, the Cisco Nexus 9000 Series NX-OS VXLAN Configuration Guide is a great place to start. Note that the aforementioned guide is for the NX-OS 10.5(x) train; make sure to reference the configuration guide for the NX-OS train you plan on deploying into production!
Apologies for the long-winded explanation - networking is a complex field from bottom to top, which makes it difficult to answer simple questions with straightforward, absolute answers. In any case, I hope you find this useful!
-Christopher
