Dell PowerSwitch Z9332F-ON 400GbE Switch (New Sealed)

The Dell PowerSwitch Z9332F-ON is a 1RU data centre leaf aggregation switch with 32 QSFP56-DD ports carrying 400GbE - and, through breakout, 10, 25, 40, 50, 100 or 200GbE - plus two 10GbE SFP+ ports, two hot-pluggable redundant power supplies and seven hot-pluggable fan modules.

It was Dell's first 400GbE fixed-format switch, and it is the model that established the QSFP56-DD cage in the PowerSwitch line. That makes it the cheapest route into a 400GbE-capable chassis on the secondary market, and it makes its breakout behaviour worth understanding properly, because it is not the simple four-lanes-per-port arithmetic people assume.

The breakout matrix, and the constraint nobody mentions

Dell publishes a per-speed breakout table for this switch rather than a single ratio, and the differences between the rows are the whole design story:

• 10GbE: 1, 4 or 8 way per QSFP56-DD port
• 25GbE: 1, 4 or 8 way per QSFP56-DD port
• 40GbE: 1 or 2 way per QSFP56-DD port
• 100GbE: 1, 2 or 4 way per QSFP56-DD port
• SFP+ (1 and 10GbE): 2 way

From that, Dell lists the reachable configurations directly: 32 x 400GbE, 32 x 100GbE, 32 x 40GbE, 64 x 40GbE, 64 x 100GbE, 128 x 100GbE, 128 x 25GbE and 128 x 10GbE.

Now the constraint. The 32 ports are arranged in two rows. If you use any top-row port as an 8x breakout, the port directly beneath it is restricted to a single interface only - one 400GbE link, or one 100GbE link, or one 40GbE link, and nothing more. It is a physical pairing limit, not a configuration setting, and it is documented in the installation guide rather than in any marketing material.

The practical effect is that a design assuming eight-way breakout across all 32 cages does not exist. Sixteen ports at 8x plus sixteen ports at 1x is the shape you actually get, and that is where the 128-port ceiling comes from. Plan the port map before you buy the optics.

Where a first-generation 400GbE switch earns its place

Dell positions the Z9332F-ON for data centre leaf aggregation - the tier that takes 25 and 100GbE from the racks and hands 400GbE upward. In a two-tier fabric that is exactly the job: 128 ports of 25GbE facing the servers, 400GbE facing the spine, and one rack unit doing it.

The QSFP56-DD ports also carry QSFP28 support directly, which is the detail that makes this switch practical in a mixed estate. You do not have to arrive at 400GbE to justify the chassis; you can run it as a 32-port 100GbE aggregation switch today and grow into the 400GbE cages as the spine catches up. Buying capability ahead of the migration is the entire argument for a multi-rate switch, and it is a stronger argument on the used market than on the new one, because the capability premium has already been depreciated by someone else.

Power: AC only, highline only

Two constraints on this model deserve to be read before anything is ordered, because both are absolute and both are easy to miss.

DC power is not supported. Unlike most of the Dell Z-series, this chassis has no DC option at all. If your facility is DC-fed, this is not the switch.

AC must be highline. Dell states that the switch requires 200-240 VAC and that 120 VAC does not provide adequate power for the system. A 120V rack circuit will not run this switch - not degraded, not non-redundant, simply not adequate. Maximum current draw is 8.5A at 200 VAC, maximum power consumption is 1244.86W, and typical consumption is 1004W measured at 45 C with sixteen ports on active optical cables and sixteen on SR8 optics running snake traffic. That is an unusually honest typical figure, because Dell states the load conditions it was measured under.

Thermally the chassis puts out 635W, or 2167 BTU/hr, and runs continuously to 113 F (45 C) with a 1 C derate per 125 m above 950 m of altitude. Seven hot-pluggable fan modules handle cooling, and the unit includes temperature monitoring, a software-readable thermal monitor and power management monitoring.

Open networking and what you are actually buying

The "-ON" suffix is Dell's open networking designation. The switch boots the Open Network Install Environment, which fetches and installs a network operating system over the network - ONIE is a bootloader, not an operating system. Dell SmartFabric OS10 and Enterprise SONiC Distribution by Dell Technologies are both documented for this platform, and the ONIE bootloader will equally install other ONIE-compatible network operating systems.

The hardware underneath is a 2-core Broadwell-DE control plane with 16GB of DDR3 and a 64GB SSD, a single RJ45 serial console port, a dedicated 10/100/1000BASE-T management port and a USB Type A storage port, all on the I/O side of the chassis. The switch is 1.70 in high, 17.26 in wide and 25.88 in deep - noticeably deeper than a typical 1RU switch, so check your cabinet depth and leave the five inches of front and rear clearance Dell specifies. It weighs 31 lbs with power supplies and fans installed.

Where it sits in Dell's 400GbE line

The Z9332F-ON is the first generation. Z9432F-ON is its direct successor at the same 32 x 400GbE port count, in the same 1RU, with a substantially larger fabric and buffer. Z9664F-ON doubles the port count to 64 x 400GbE in 2RU, and Z9864F-ON moves to 800GbE on OSFP112 cages for AI training fabrics. Below them the Z9264F-ON and Z9100-ON hold the 100GbE tier.

This unit is new condition, fully tested and shipped with warranty. It is quote-only - tell us the port map, optics and quantity you need and a specialist will come back to you with availability.

Specifications