Dell Networking S6000-ON 40GbE Switch

The Dell Networking S6000-ON is a 1RU switch with 32 line-rate 40GbE QSFP+ ports on a 2.56 Tbps cut-through fabric, forwarding 1,462 Mpps at sub-600 nanosecond latency. Every port breaks out to four 10GbE, so the same chassis can present 32x 40GbE, 96x 10GbE plus 8x 40GbE, or 104x 10GbE with no 40GbE remaining.

It was Dell's first "-ON" open-networking switch and, on the secondary market, it is one of the cheapest ways to add 40GbE aggregation or a large block of 10GbE ports to a rack.

Be honest about 40GbE: this is a legacy-fabric play

40GbE is not where new data centre builds go. A 40GbE QSFP+ port is four 10G lanes bonded together; the industry moved to 25G-lane serdes because a single 25G lane carries 2.5x what a 10G lane does on the same cable and the same cage. That is why 25GbE access with 100GbE uplinks displaced 10GbE access with 40GbE uplinks, and why a greenfield build today starts at 25/100G rather than 10/40G.

So we will not pretend the S6000-ON is a forward-looking purchase. Buy it if you already have a 40GbE estate. Specifically, it earns its place when:

• You are expanding an existing 10/40GbE fabric and need another spine or aggregation switch that speaks the same optics, DACs and breakout cables you already own.
• You need a large block of 10GbE ports cheaply - 96 or 104 of them from one rack unit, using breakout cables rather than a chassis switch.
• You are replacing a failed unit in a live fabric and need matching hardware rather than a redesign.
• You need a lab, staging or DR switch that mirrors production without production budget.

If none of those describe you and you are starting fresh, look at 25/100G instead - the S5048F-ON for 25GbE server access, or the current-generation S5248F-ON. The optics are cheaper per gigabit and the platform has a longer runway.

What has not changed is the physics. 2.56 Tbps and sub-600ns latency are still fast, and a fabric built on this generation does not get slower because a newer generation exists.

Breakout: one chassis, three port layouts

Each QSFP+ port carries four 10G lanes. A breakout cable splits one port into four SFP+ ports, which is what gives the S6000-ON its three published configurations:

The 104-port mode is the one people overlook. 104 ports of 10GbE in a single rack unit is a density that otherwise costs a chassis switch, and it is the reason this model still sells years after its 40GbE role was superseded. A QSA adapter also lets a QSFP+ port take an ordinary SFP+ or SFP module, so 1GbE connections are possible where you need them, and there is a QSFP+ to 4x RJ45 breakout cable for copper gigabit.

One planning note: breakout consumes the port. A port broken out to 4x 10GbE is no longer available at 40GbE, which is why the three configurations above are alternatives rather than additive.

Fabric, latency and scale

The switch runs 2.56 Tbps of switching I/O bandwidth full-duplex with a 1,462 Mpps forwarding rate and sub-600 nanosecond latency in cut-through mode. Dell rates the non-blocking claim over aggregate operation with average packet transfers greater than 200 bytes.

Behind it: a 160K MAC table, a 128K ARP table, 128K IPv4 routes, 64K IPv6 routes and 64K IPv6 hosts, with 4 GB of CPU memory and a 12 MB packet buffer. VRF-lite scales to 511 instances, which is a genuinely large number for a switch of this vintage and the reason Dell positioned it for multi-tenant network-centric virtualization. Layer 2 VLANs top out at 4K, MST at 64 instances, and link aggregation runs 16 links per group across 128 groups.

Stacking, VLT and Active Fabric

This model supports user port stacking for up to six units - a shared control plane across six chassis, managed as one. That is worth flagging, because it is the clearest functional difference between this switch and its successor: the S6010-ON specification makes no stacking claim at all.

For multi-chassis redundancy without a shared control plane, the switch does Virtual Link Trunking, Routed VLT and VLT Proxy Gateway, which give you an active-active pair with no spanning-tree blocked links and let you stretch a Layer 2 VLAN within or across two data centres. Dell also positioned the S6000-ON inside its Active Fabric designs - as a leaf or spine alongside the Z-series for large flat two-tier 10/40GbE networks, or in a small-scale leaf-and-spine with 1/10GbE S-series switches beneath it.

Virtualization, storage and SDN

Dell's own framing for this switch was that it supported both approaches to network virtualization rather than forcing a choice: VRF-lite for network-centric isolation and a line-rate VXLAN gateway for hypervisor-centric overlays, bridging virtualized and non-virtualized infrastructure.

For storage, the full Data Center Bridging set is present - priority flow control (802.1Qbb), enhanced transmission selection (802.1Qaz), DCBx and the DCBx application TLV for iSCSI and FCoE - and the switch supports Routable RoCE so RDMA traffic can cross a routed boundary. OpenFlow 1.3 is supported for SDN deployments and interoperates with standard OpenFlow controllers. The Embedded Open Automation Framework includes an in-box Puppet agent, and the Fastboot feature allows a minimum-loss software upgrade on a standalone unit that is not in a VLT pair or a stack.

Airflow: standard and reverse builds

This listing covers both airflow builds of the S6000-ON. Select the direction under the Airflow option: air direction is set by the power supplies and fan modules physically fitted, so it is chosen when the unit is built rather than switched afterwards. Power supplies and fans must both face the same way - a mismatched pair raises the switch's incompatible-airflow alarm.

Standard airflow (I/O to PSU)

Unusually, Dell's specification for this model documents a conversion. In-field change of airflow direction is supported only when the unit is powered down and all fan and power supply units are replaced with airflow moving in a uniform direction. In practice that means buying a full set of direction-matched fans and power supplies and swapping them on a dead chassis - which is a real option if you already own the parts, and rarely worth it if you do not. Buying the chassis in the direction you need is simpler, which is why we list the two builds separately.

The one thing you cannot do is mix. Every fan and every power supply in the chassis must move air the same way.

Reverse airflow (PSU to I/O)

Dell names airflow directions by where the air goes. I/O Panel to PSU airflow is the standard build: cool air in at the port face, exhaust at the power-supply and fan face. PSU to I/O Panel airflow is the reverse. This -RA SKU is the PSU-to-I/O-Panel unit.

Which one you need is decided by which way the switch faces in the rack, not by preference. In a hot-aisle/cold-aisle row, servers pull cool air from the cold aisle at the front and exhaust into the hot aisle at the back. Rack a top-of-rack switch with its port face toward the cold aisle and standard airflow is correct. Rack it with the port face toward the hot aisle - very common, because it puts switch ports on the same side as server NICs and halves cable runs - and a standard-airflow switch would be inhaling your own hot exhaust from day one, running hot and spinning its fans harder for the rest of its life.

A quick field test if the unit is in front of you: stand at the port face with the switch powered on. Air blowing out at you means reverse airflow. Air being pulled in means standard.

Open networking, ONIE and what a licence means second-hand

The "-ON" in the model name is the point of the product. This was Dell's first disaggregated data centre switch: the hardware ships with the Open Network Install Environment, which supports zero-touch installation of an alternate network operating system, and the factory software is Dell's own Networking Operating System with its full routing stack - OSPFv2/v3, BGP-4 with IPv6 multiprotocol extensions, IS-IS, RIP, VRRP, PIM-SM, IGMP and MSDP, plus policy-based routing.

What you are buying from us is the hardware. Dell software entitlements are bound to a unit's service tag and to a Dell account, and they do not automatically travel with a switch on the secondary market. A refurbished unit may arrive with its factory software installed and fully functional and still carry no active entitlement or support contract. We do not promise a transferred Dell licence, a software subscription or a support contract with a refurbished switch, and you should treat any reseller who does with suspicion. Tell us what you intend to run and we will confirm the state of a specific unit before it ships.

Common questions

What is the difference between the S6000-ON and the S6010-ON?

Both are 32x 40GbE QSFP+ in 1RU at 2.56 Tbps, 1,462 Mpps and sub-600ns. The S6000-ON has a smaller 12 MB buffer and 4 GB of CPU memory but larger IPv6 tables (64K routes and 64K hosts against 8K and 24K), a much larger ingress ACL (2.5K against 716), lower power draw, and support for user port stacking up to six units. The S6010-ON has a 16 MB buffer, 8 GB of CPU memory, and its specification names four supported third-party network operating systems. They are frequently confused; the two are not interchangeable on paper.

How many 10GbE ports can it provide?

104 with every QSFP+ port broken out and no 40GbE remaining, or 96 while keeping eight ports at 40GbE.

Does it stack?

Yes - user port stacking for up to six units. It also supports VLT, Routed VLT and VLT Proxy Gateway for multi-chassis active-active pairs.

Can I use it as a spine?

Yes. 32 ports of 40GbE was a standard spine radix in this generation, and Dell positioned it exactly that way in its Active Fabric designs alongside the Z-series.

Is 40GbE worth buying in 2020s infrastructure?

Only if you already run it. See the section above - the honest case for this switch is expanding, replacing or lab-mirroring an existing 40GbE estate, or getting 96 to 104 cheap 10GbE ports out of one rack unit.

Where it sits in the Dell S-series

Condition, testing and what ships in the box

This is a refurbished, professionally reconditioned unit. Every switch is powered up, inspected and functionally tested before it leaves us, and it ships with warranty coverage. Power supplies, fan modules, rail kits, console cables, optics and breakout cables are configurable above - choose the airflow direction that matches the chassis, because power supplies and fans are direction-specific parts.

If you need a specific firmware state, a particular breakout loadout, reverse airflow, or matched units for a stack or a VLT domain, tell us before you order and we will build to it.

Specifications