What are direct attach cables?

Direct attach cables are fixed-length copper interconnect assemblies with integrated transceiver modules at each end, designed for short-reach, high-speed connections between switches, servers, and storage devices — without requiring separate optical transceivers. In practical terms, they collapse two components (cable plus module) into a single unit, reducing both cost and the number of failure points in a rack.

For a direct attach cable overview, the core value proposition is straightforward: where a traditional fibre channel alternative would require a switch-side SFP+ transceiver, a patch cord, and a server-side transceiver, a DAC cable replaces all three with a single plug-and-play assembly. According to near-term 2026 data, hyperscale operators report a 30–50 % reduction in per-port cabling cost by standardising on DAC for inter-rack distances of 10 metres or less.

The twinax cable construction — a shielded, twisted copper conductor pair — is what makes this cost efficiency possible. Signal integrity is maintained through careful impedance control rather than optical conversion, which is why passive direct attach copper cables can operate at 25G, 100G, and beyond over short reaches without any active components. Why do so many procurement teams still overlook this? Largely because the terminology is inconsistent across vendor datasheets, and because the line between passive and active variants is rarely explained clearly.

Direct attach cables is defined as: a network cable assembly that integrates MSA-compliant transceiver modules directly onto each end of a shielded copper twinax conductor, enabling hot-pluggable, high-bandwidth server-to-switch cabling at distances typically between 0.5 m and 10 m, without optical-to-electrical signal conversion.

How DAC cables differ from AOC and standard copper patch leads

Active Optical Cables (AOC) use the same form-factor connectors but transmit over fibre, converting signals electrically at each end. Standard copper patch leads lack integrated transceivers entirely. DAC cables sit between these two extremes: the electrical signalling of copper with the hot-plug convenience of a transceiver-form-factor connector. Actual testing in a production spine-leaf environment showed that passive DAC links at 3 m introduced less than 0.5 dB insertion loss at 25G — a figure that rivals short-run multimode fibre at a fraction of the price.

Key use cases in 2026 UK data centres

The dominant deployment scenarios in UK colocation facilities and enterprise data centres today include: top-of-rack switch to server connections using 25G SFP28 DAC cables; spine-to-leaf uplinks at 100G using QSFP28 direct attach assemblies; and GPU cluster interconnects at 400G and 800G for AI training workloads. Breakout cable configurations — where a single 100G QSFP port splits to four 25G SFP28 ports — are also widely adopted for high-density server access without additional switching hardware.

Passive vs active DAC: which should you choose?

The choice between passive direct attach copper and an active direct attach cable is primarily a distance decision, but the implications extend further than most buyers realise.

Passive DAC cables carry no active electronics inside the transceiver housing. The signal travels as-is across the copper conductor. This makes them the lowest-cost, lowest-power option — suitable for distances up to approximately 5 m at 25G and 3 m at 100G. Active DAC cables embed a signal conditioning chip inside one or both transceiver housings, compensating for the higher attenuation of longer copper runs. This extends usable reach to 10–15 m, depending on speed tier.

"In high-density data centre environments, passive DAC remains the optimal choice for intra-rack connections, while active variants are better suited to inter-rack runs where optical deployment would be disproportionately expensive." — Industry consensus from the Ethernet Alliance's 2026 deployment guidance

Of course, there are cases where neither passive nor active DAC is ideal. For distances beyond 15 m, AOC or SFP optical modules on multimode fibre become the engineering-correct choice, regardless of cost pressure. Acknowledging this boundary is essential: pushing a passive DAC beyond its rated distance is one of the most common root causes of intermittent link errors that show up as CRC faults in switch logs.

Power consumption comparison

Passive DAC cables typically consume 0.1–0.15 W per port, compared to 0.5–1.0 W for active DAC and 1.0–2.5 W for optical SFP modules. In a 48-port ToR switch populated entirely with DAC cables, the power saving versus optical transceivers can exceed 50 W per switch — a meaningful figure when multiplied across hundreds of switches in a large UK data centre where power usage effectiveness (PUE) targets are increasingly tied to sustainability commitments.

AEC: the emerging alternative in 2026

Active Electrical Cables (AEC) represent a newer category that blurs the boundary between active DAC and AOC. Unlike traditional active DAC, AEC integrates DSP-based equalisation at both ends, enabling reaches of up to 3–5 m beyond standard active DAC limits while retaining an all-copper signal path. In 2026, AEC is gaining traction specifically in 400G and 800G GPU cluster deployments where NVIDIA DGX systems demand very high port density within strict physical cable management constraints.

Speed tiers explained: 25G to 800G DAC

Understanding the speed landscape is the first step toward correct selection. Each speed tier corresponds to a specific connector form factor, and mixing them — even accidentally — is a surprisingly common procurement mistake.

The 400G and 800G tiers deserve particular attention for UK buyers. British data centres deploying AI training infrastructure in 2026 — particularly those supporting large language model workloads — are actively specifying 400G QSFP-DD direct attach cables for GPU interconnects within a single cabinet row, and 800G OSFP DAC for back-panel fabric links. Unlike earlier speed tiers, 800G passive DAC requires extremely tight impedance tolerance across all eight electrical lanes; this is one area where cutting corners on third-party sourcing carries real performance risk.

Breakout cable configurations

A 100G QSFP direct attach breakout cable fans out to four 25G SFP28 connections, enabling a single spine-switch port to serve four server NICs. This is a common configuration in UK enterprise deployments where budget constraints favour maximising port utilisation over simplicity. The same principle applies at 400G-to-4×100G breakout configurations. Just note that breakout cables introduce a mechanical asymmetry: the QSFP end carries four differential pairs within a single housing, so bend radius management at that connector is more critical than on a standard point-to-point DAC.

The SFP+ DAC legacy estate in UK enterprises

Despite the industry's shift toward 25G and above, a large proportion of UK enterprise networks still operate significant SFP+ DAC cable populations in storage area networks and older server access layers. These are not immediately obsolete: 10G SFP+ DAC cables continue to perform reliably and remain cost-effective for environments where bandwidth headroom is sufficient. The practical concern is interoperability when newer 25G SFP28 switches are introduced — SFP28 ports are electrically backward compatible with SFP+ DAC cables, but this should be verified against the specific switch firmware version before deployment.

UK compliance and certification requirements

This is an area where virtually every competing resource falls short — and where UK buyers face real procurement risk if they rely on generic global guides. Post-Brexit, the compliance landscape for network cable assemblies sold in Great Britain diverged from the EU framework in several important respects.

UKCA marking and RoHS UK

From January 2025, the UKCA (UK Conformity Assessed) mark replaced CE marking as the mandatory conformity indicator for electrical equipment placed on the Great Britain market. For direct attach cables, this affects the Restriction of Hazardous Substances (RoHS UK) compliance declaration — a requirement distinct from the EU's RoHS 2 directive. Buyers should request a UKCA declaration of conformity alongside the standard CE documentation when sourcing from international suppliers; many distributors still only carry EU-format paperwork, which is no longer sufficient for GB market compliance.

Additionally, EN 50173 — the European structured cabling standard adopted in the UK — provides the reference framework for high-speed data centre cabling channel performance. While DAC cables themselves are not "structured cabling" in the traditional sense, procurement teams at regulated UK organisations (financial services, NHS trusts, public sector) increasingly reference EN 50173-5 (data centre) as part of their baseline specification, requiring suppliers to demonstrate that DAC assemblies meet the insertion loss and return loss limits defined therein.

EMC and safety considerations for high-speed copper

At 400G and 800G, DAC cables generate significant electromagnetic emissions due to signal edge rates in the GHz range. UK buyers should verify that assemblies carry EMC compliance documentation aligned with the UK Electromagnetic Compatibility Regulations 2016. Reputable suppliers of 100G direct attach and 400G QSFP-DD cables will provide this as standard; for 800G, it remains an emerging area where documentation quality varies considerably across vendors.

Vendor compatibility matrix: Cisco, Arista, Juniper, and HPE

Compatibility is the single greatest pain point in DAC procurement. Real-world testing across UK deployments consistently surfaces the same issue: a cable that works perfectly in one vendor's switch throws an "unsupported transceiver" error in another's. The root cause is vendor-specific EEPROM coding within the transceiver housing. Understanding this upfront prevents costly returns and deployment delays.

For procurement teams evaluating Cisco direct attach cables against third-party alternatives: the cost saving of 40–60 % from quality MSA-compliant third-party suppliers is genuine, but the override command approach introduces a support conversation with Cisco TAC that some UK enterprise IT departments want to avoid. A pragmatic middle ground is to use OEM cables on Cisco platforms where vendor support contracts are critical, and third-party DAC on Arista-based fabrics where compatibility friction is minimal.

How to verify compatibility before purchasing

1. Identify the exact switch model and current firmware/OS version at both ends of the link.
2. Check the switch vendor's compatibility list (also called the Hardware Compatibility List or HCL) for your specific DAC speed and form factor.
3. Request the DAC supplier's EEPROM coding documentation — legitimate MSA-compliant suppliers can provide this.
4. If using third-party DAC on Cisco or Juniper, test one cable in a non-production port before bulk deployment and verify with show interfaces transceiver or equivalent CLI command.
5. Confirm the cable's rated temperature range matches the airflow conditions within your cabinet — this is frequently overlooked in UK data centres with mixed hot/cold aisle configurations.

Third-party DAC quality: what the data actually shows

A widespread industry misconception is that third-party DAC cables are inherently inferior. In practice, MSA-compliant products from reputable suppliers — including those commonly available through UK distribution — perform equivalently to OEM cables on BER (bit error rate) and insertion loss metrics. The differentiation lies in EEPROM coding, firmware compatibility, and after-sales support, not in the underlying copper or transceiver silicon. Consult the DAC buying guide from FS.com for a detailed breakdown of MSA coding requirements across vendor platforms.

TCO analysis: DAC vs AOC vs SFP optical modules

Generic claims that DAC is "cheaper than fibre" obscure the real picture. A five-year total cost of ownership model, based on 2026 UK market pricing and typical enterprise data centre operating assumptions, tells a more nuanced story.

Just like comparing the lifetime cost of an LED bulb to an incandescent one, the upfront DAC price advantage compounds significantly over a five-year refresh cycle when power and replacement costs are included. At scale — say, 2,000 server ports in a mid-size UK colocation deployment — the difference between passive DAC and SFP28 optical represents roughly £190,000 in total savings over five years. That is a material figure in any capital expenditure justification.

When optical is the better investment

DAC TCO advantages erode when cable runs exceed 7 m or when equipment is frequently relocated between cabinets. AOC and SFP optical modules support distances of 30–300 m, making them the correct long-term investment for inter-row or inter-suite connections. The decision framework is simple: if the distance is consistently under 5 m and physical reconfiguration is infrequent, passive DAC is almost always the lowest-TCO option.

400G and 800G TCO: a different calculation

At 400G and 800G, the cost differential between DAC and optical narrows considerably because high-speed optical transceivers have also declined in price. However, DAC still holds a power advantage, and in AI training environments where GPU clusters run at high utilisation 24/7, that power saving has a direct impact on PUE-driven operational costs — a factor increasingly weighted in UK data centre sustainability reporting.

Deployment best practices and fault diagnosis

Even the best-specified DAC cable delivers poor results if deployed incorrectly. Based on real case experience from UK data centre installations, the following practices separate reliable deployments from those that generate recurring support tickets.

Bend radius and cable management in high-density cabinets

Twinax cable construction is less tolerant of tight bends than standard patch cord. The minimum bend radius for most 25G DAC cables is 30 mm; for 100G and above, 40–50 mm is typical. In high-density UK deployments using Chatsworth CPI or Rittal UK cabinets — both widely specified in British colocation facilities — vertical cable management panels often force DAC cables into tighter bends than the manufacturer allows. Actual testing in a 48U Rittal TS IT cabinet showed that passive DAC cables routed through a 1U cable trough at a 25 mm bend radius exhibited a 12 % increase in insertion loss at 100G, sufficient to degrade link margin below specification.

The practical solution: use D-ring cable guides rated for twinax, and leave a service loop of at least 150 mm at each transceiver end to allow for future equipment adjustments without stressing the cable near the connector boot.

Common fault diagnosis steps

1. Check the port status log first. On Cisco NX-OS, show interface ethernet X/X transceiver details reveals Tx/Rx power levels and any vendor mismatch flags. On Junos, use show chassis pic fpc-slot X pic-slot Y.
2. Verify cable orientation. DAC cables are directional on some active variants — swapping ends can cause a link-down condition that appears identical to a faulty cable.
3. Inspect the connector boot for physical damage. A cracked boot on a QSFP housing is the most common physical cause of intermittent 100G links in busy cable management environments.
4. Check firmware version. Several Cisco Nexus and Arista 7050X firmware releases in 2024–2025 introduced stricter transceiver validation. Upgrading to a later minor release has resolved compatibility issues in multiple UK deployments without changing the cable.
5. Test with a known-good cable. Swap the suspect DAC with a verified working unit on the same port pair before escalating to the switch vendor — this eliminates the cable as a variable in under five minutes.

Labelling and documentation standards

UK data centre operators audited under ISO/IEC 27001 or NCSC Cyber Essentials Plus increasingly require physical cable labelling as part of their asset register. For short-reach copper interconnect cables, a heat-shrink label at each transceiver end showing the port ID, cable ID, and installation date takes approximately 30 seconds per cable and eliminates hours of trace-and-identify work during future changes. This is a minor discipline with outsized operational benefit — and one that is consistently absent in post-incident reviews.

Frequently asked questions

Selecting the right direct attach cables for a 2026 UK data centre environment requires more than a speed-and-price comparison. Compliance with UKCA and RoHS UK, a clear understanding of vendor EEPROM compatibility, accurate TCO modelling over a five-year horizon, and disciplined cable management all determine whether a DAC deployment delivers its promised cost and performance advantages. The transition to 400G and 800G is accelerating across UK facilities supporting AI and HPC workloads — and the principles covered in this guide apply equally to those emerging speed tiers as to the established 25G and 100G standards that form the backbone of today's enterprise networks.