In modern SMB and edge data centres, once you move beyond simple 1G/10G copper patch cords, three options appear very quickly for short, high-speed links: DAC cables, AOC cables, and the classic combination of optical transceivers plus fiber jumpers. They all plug into SFP+/SFP28/QSFP ports and they all say “10G/25G/40G” on the label, but they behave very differently in cost, reach, power and day-to-day handling.

This article focuses on DAC cables from a practical point of view. We will look at what DAC assemblies actually are, how they differ from AOC and pluggable optics, and which patterns in real projects tend to favour DAC over the alternatives. The goal is not to memorise every part number, but to make future decisions about “which cable to use” in your racks much faster.

1. What Exactly Is a DAC Cable?

A DAC cable is a factory-terminated twinax copper cable with integrated transceiver-style ends on both sides. Instead of buying two separate optical modules and a patch cord, you receive a single, fixed-length assembly where the “modules” and the cable are permanently attached.

Electrically, the DAC plugs into SFP+/SFP28/QSFP cages and presents itself to the switch or NIC in almost the same way as an optical transceiver. Internally, however, there is no optical conversion: high-speed electrical lanes travel directly over the twinax conductors. This direct path is exactly what gives DAC its characteristic strengths and limitations.

In practice you will come across two broad DAC categories:

Passive DAC. The cable assembly contains no active signal conditioning. It relies on the host ports’ own equalisation to open the eye diagram. Passive DACs are typically used for the very shortest links in a rack – often 1–3 metres – and are valued for their low cost, zero power consumption and low latency.

Active DAC. The ends of the cable include active components (such as equalisation or pre-emphasis) to extend the useful reach of copper. Active DACs can support longer lengths at higher speeds than passive ones, but they draw a small amount of power from each host port and are usually slightly more expensive.

2. DAC vs AOC vs Optics + Fiber: How They Really Compare

From the outside, DAC, AOC and transceiver-plus-fiber assemblies all look like “something you plug into SFP or QSFP ports”. Inside the network, however, they serve different technical and economic niches. The table below summarises the main contrasts.

For SMB and edge environments the pattern is straightforward: DAC for very short, high-speed links inside and between racks; AOC for medium-reach links where you do not want heavy copper bundles; and discrete optics plus structured fiber for anything beyond that.

If you are at the stage of choosing physical cables for an existing switch layout, our dedicated collection of high-speed assemblies can be a useful reference. You can see all DAC and AOC cables here to get a sense of common lengths, gauges and form factors.

3. Common Types of DAC Cables in Real Projects

In theory there is a wide variety of DAC options. In real projects you see the same combinations again and again:

SFP+ to SFP+ 10G DAC. The workhorse for 10G short-reach links, often 1–3 metres, connecting servers to a Top-of-Rack (ToR) switch or linking switches inside a rack. Most of these assemblies are passive twinax.

SFP28 to SFP28 25G DAC. The same pattern at 25G. The physics are less forgiving, so cable length and gauge matter more. It is common to see a mix of passive DACs at 1–3 metres and active DACs when you push towards 5 metres.

QSFP+ or QSFP28 DAC. These handle 40G/100G links, either as 1:1 connections between aggregation switches or as breakouts to multiple 10G/25G ports (for example, QSFP28 to 4×SFP28). They are very useful for spine and leaf designs inside a single rack or row.

Gauge and flexibility. Thicker twinax (for example 24 AWG) supports longer reaches but can be stiff in dense cable managers. Thinner gauges (for example 30 AWG) are more flexible but are normally used only for shorter lengths. In tight racks, cable manageability can become just as important as the extra metre of reach.

4. When DAC Cables Are the Best Choice

Once you understand the basic trade-offs, situations where DAC is the best answer start to look very similar. A few examples illustrate the pattern.

4.1 In-Rack Server to ToR Links

The classic DAC use case is a server two units below a ToR switch. The cable path is short, both ends sit in the same rack, and the likelihood of repeated re-routing is low. In this scenario 1–3 metre passive DACs provide:

predictable latency, minimal power draw, low material cost and very straightforward deployment. There is little to be gained from using AOC or full optical modules for these links unless you have a very unusual rack layout.

4.2 Short Inter-Rack Links in the Same Row

In many SMB and edge rooms, two or three racks make up the entire environment. Switches in neighbouring racks are often linked using short 10G/25G or 40G DAC cables that run through an overhead or under-floor tray. Total length may be 3–5 metres, squarely in active DAC territory at higher speeds.

These links benefit from DAC’s low latency and simplified bill of materials. Instead of managing separate optical modules and patch cords, you deal with a single coded assembly per link. The trade-off is cable bulk, which becomes noticeable if you run many parallel DACs between the same racks.

4.3 Stacking and Short Uplinks Between Switches

Some switch families support dedicated stacking ports or front-panel ports that are commonly connected over DAC assemblies. When all stack members sit close together, DAC once again offers a good balance of simplicity and performance.

For short uplinks from access switches to an aggregation switch in the same row, DAC is also a natural fit, provided the length and routing keep you within safe limits for the chosen speed and gauge.

5. When AOC or Optics Make More Sense

There are also scenarios where DAC is technically possible but no longer the most practical option. As your links become longer, more numerous or more physically exposed, AOC or optics plus structured fiber tend to win.

The first point is reach. Even with active equalisation, DAC links are generally constrained to a handful of metres at current speeds. If your link budget or physical layout calls for tens of metres, it is usually more efficient to switch to AOC or to standard optics with well-planned fiber runs.

The second point is cable bulk. High-density DAC bundles are heavy and stiff, which can make it difficult to maintain bend radius, preserve airflow and keep racks tidy. AOC and fiber patch cords, in contrast, are lighter and more flexible in large bundles.

Finally, there is the question of routing flexibility. Fiber-based links can be re-routed across patch panels and consolidation points with more freedom than fixed-length DAC assemblies. In environments where layouts change frequently or where you want a clear separation between backbone and equipment cabling, the optical approach fits more naturally into structured cabling practices.

6. Design Checklist for SMB and Edge Data Centres

Rather than debating DAC vs AOC in the abstract, it helps to run each candidate link through a short checklist. The answers tend to make the right choice obvious:

1) What is the port type and speed? Are you dealing with SFP+, SFP28, QSFP+ or QSFP28 ports, and what line rate do they run? Not all DAC part numbers are available for every combination.

2) What is the realistic cable path length? Measure the actual route, not just the straight-line distance. For 10G/25G passive DAC, keeping to three metres or less gives comfortable margin in most hardware designs. Longer paths may call for active DAC or optical alternatives.

3) How many parallel links run in the same pathway? A single DAC is easy to route; twenty in the same vertical manager is a different story. If you expect many parallel links, factor cable bulk and bend radius into the decision.

4) How stable is the physical layout? DAC assemblies work best in stable, well-known layouts where the endpoints will not move frequently. In more dynamic environments, optics plus patch panels are easier to reconfigure.

5) Are there strict power or latency constraints? Where both power budgets and latency budgets are tight, DAC has an advantage over solutions that require optical conversion – especially on dense ToR switches.

7. Choosing DAC Cables Step by Step

Once you have decided that DAC is appropriate for a given link, the actual selection process is relatively mechanical. A repeatable approach helps keep ordering consistent across different projects and sites.

Step 1 – Confirm the form factor and speed at each end. Note whether the ports are SFP+, SFP28, QSFP+ or QSFP28, and confirm that they are configured for the same line rate. For breakout applications (for example QSFP28 to 4×SFP28) check that the platform explicitly supports this mode.

Step 2 – Measure the route and select length. Follow the intended cable path between devices, including vertical and horizontal managers, and pick the next standard length up from the measured value. Leaving a small amount of slack is better than forcing a short DAC to stretch across the rack.

Step 3 – Choose passive or active and gauge. For very short, in-rack runs at 10G, passive 30 AWG DACs are typically fine. As length and speed increase, moving to a heavier gauge or to active DAC extends the reach. Reviewing the switch vendor’s guidance for maximum DAC length at each speed is a good practice.

Step 4 – Confirm compatibility expectations. Many switches and NICs are tolerant of standards-based third-party DACs, but some enforce vendor coding. Checking the platform’s compatibility list and documenting any constraints upfront avoids unpleasant surprises during installation.

Step 5 – Standardise part numbers where possible. Once you find combinations that work well in your environment, standardising on a small set of DAC assemblies for common link types simplifies both operations and spare-stock planning.

To see what these combinations look like as real products – among them SFP+, SFP28 and QSFP options in common lengths – you can browse our DAC and AOC cable collection as a practical reference.

FAQ: DAC Cables in SMB and Edge Networks

Do DAC cables always have lower latency than optics?

DAC cables avoid optical conversion, so they remove a small amount of processing delay compared to solutions that use optical modules. In practice, the latency difference is modest in absolute terms, but in very latency-sensitive designs every microsecond can matter.

How far can I reliably run a DAC cable?

For 10G passive DAC, many deployments stay at or below three metres. At higher speeds or longer distances, active DACs can extend the reach to around five to seven metres, depending on platform and gauge. Beyond that range, AOC or fiber-based solutions are usually more robust.

Are DAC cables part of structured cabling?

DAC assemblies are usually treated as equipment cables rather than part of the permanent structured cabling system. The permanent plant is still built from copper or fiber cabling terminated on patch panels, with DAC used between active devices at the rack level.