Skip to main content

Managing Bandwidth in Student Accommodation Networks

This guide provides IT managers, network architects, and property operations directors with a vendor-neutral technical reference for managing WiFi bandwidth in high-density student accommodation environments. It covers VLAN segmentation, Quality of Service (QoS) policy design, identity-based traffic shaping, and application-layer visibility โ€” the four pillars of a scalable, fair-access network. With real-world deployment scenarios, measurable outcomes, and decision frameworks, this is the operational playbook for any team responsible for residential network infrastructure at scale.

๐Ÿ“– 8 min read๐Ÿ“ 1,982 words๐Ÿ”ง 2 worked examplesโ“ 3 practice questions๐Ÿ“š 9 key definitions

Listen to this guide

View podcast transcript
Welcome back to the Purple Technical Briefing. I'm your host, and today we're tackling one of the most persistent headaches for property managers and IT directors in the high-density residential sector: Managing Bandwidth in Student Accommodation Networks. If you're managing connectivity for hundreds or thousands of digital-native residents, you already know the pain points. The sheer volume of concurrent connections, the proliferation of IoT devices, and the insatiable demand for streaming and gaming can bring even a robust network to its knees. Today, we're cutting through the noise. No academic theory โ€” just practical, vendor-neutral strategies for bandwidth shaping, Quality of Service, and fair access policies that you can implement this quarter. Let's dive straight into the technical deep-dive. The core challenge in student housing isn't just raw throughput; it's contention and fairness. A flat network architecture with basic throttling is a recipe for disaster. When you simply apply a global 20 megabits-per-second cap on every device, you're not solving the problem โ€” you're just equally distributing the misery during peak hours. What you need is a layered approach. First, VLAN segmentation is non-negotiable. You must isolate student traffic from administrative, IoT, and building management systems. This isn't just about performance; it's a fundamental security requirement. Under IEEE 802.1Q, each VLAN operates as a logically separate broadcast domain, meaning a compromised student device cannot traverse into your building management network or administrative infrastructure. Once segmented, you implement intelligent traffic shaping. This means moving beyond static caps. We recommend dynamic bandwidth allocation. During low-usage periods โ€” say, between 2 and 9 in the morning โ€” let users burst up to higher speeds, perhaps double or triple their baseline allocation. But when contention hits 80 percent of your uplink capacity, your traffic shaping rules must aggressively prioritise latency-sensitive applications like VoIP and video conferencing over bulk downloads and peer-to-peer traffic. This brings us to Quality of Service, or QoS. You should be marking packets at the edge โ€” right at the access point โ€” using standard Differentiated Services Code Point, or DSCP, values. Voice traffic gets Expedited Forwarding, which is DSCP 46. Video conferencing gets Assured Forwarding. Background updates and bulk downloads get Best Effort or lower. This classification must happen at ingress, before the packet hits your core switching fabric, otherwise you've already lost the battle. Now, let's talk about the identity layer, because this is where most deployments fall short. The average student brings seven connected devices to their accommodation. Laptops, smartphones, tablets, smart TVs, gaming consoles, smart speakers, and wearables. If your bandwidth policy is built around per-device limits rather than per-user limits, you will exhaust your DHCP address pools and your bandwidth allocations will be gamed trivially. The solution is an identity-driven approach. Authenticate the user via IEEE 802.1X โ€” ideally using WPA3-Enterprise for the security benefits โ€” tie all their devices to a single user identity, and apply the bandwidth policy to the aggregate user session. When that user's combined device footprint exceeds their allocation, the policy applies across all sessions simultaneously. This is fundamentally different from per-MAC throttling, and it's the approach that scales. For devices that don't support 802.1X natively โ€” gaming consoles, smart TVs, IoT sensors โ€” implement MAC Authentication Bypass, or MAB, combined with a self-service registration portal. Students register their headless devices through a captive portal, those devices are placed into a specific device group, and tailored QoS profiles are applied. This gives you visibility and control without creating a support burden. Let's talk about application-layer visibility, because you cannot manage what you cannot measure. Deep Packet Inspection, or DPI, at the gateway gives you the application-layer telemetry you need to make intelligent policy decisions. If you can see that 60 percent of your uplink capacity is consumed by a single streaming service, you have options: you can cache that content locally using a transparent proxy, adjust your peering arrangements, or apply application-specific rate limits during peak hours. Platforms like Purple's WiFi Analytics provide exactly this kind of granular visibility โ€” not just raw throughput metrics, but application-layer intelligence that informs your bandwidth policy decisions in real time. Now, let me walk you through two real-world implementation scenarios. The first is a 400-bed purpose-built student accommodation block in Manchester. Prior to the engagement, the network was running a flat architecture with a single SSID and a global 10 megabits-per-second per-device cap. During peak hours โ€” typically 7 to 11 in the evening โ€” the network was effectively unusable for video conferencing. Support tickets were running at 40 per week. The remediation involved deploying VLAN segmentation across three logical networks: students, staff, and IoT. A per-user bandwidth policy of 25 megabits-per-second was implemented with dynamic burst capability up to 50 megabits-per-second during off-peak hours. QoS policies prioritised video conferencing traffic using DSCP marking at the access point layer. Within 30 days of deployment, support tickets dropped by 78 percent and average peak-hour throughput per user increased by 140 percent โ€” despite no change to the uplink capacity. The second scenario is a 1,200-bed university halls of residence in Edinburgh. The challenge here was more complex: the existing infrastructure was a mix of legacy 802.11ac access points and newer Wi-Fi 6 hardware, and the network had no application-layer visibility whatsoever. The approach was a phased migration. Phase one: deploy a unified network management platform with DPI capabilities and establish baseline telemetry over 30 days. The data revealed that 55 percent of peak-hour traffic was attributable to four streaming platforms. Phase two: implement application-aware QoS policies, throttling streaming traffic to 8 megabits-per-second per user during peak hours while maintaining full speed for video conferencing and academic platforms. Phase three: migrate authentication to 802.1X with per-user policy enforcement. The outcome was a 35 percent reduction in peak-hour congestion and a measurable improvement in resident satisfaction scores. Now let me address the common pitfalls and risk mitigation strategies. Pitfall one: blanket peer-to-peer blocks. Don't do it. Blanket bans on peer-to-peer traffic drive users to commercial VPN services, which completely blinds your deep packet inspection and analytics. Instead, throttle peer-to-peer to a trickle โ€” 1 to 2 megabits-per-second โ€” and deprioritise it to best-effort. You retain visibility, you reduce the bandwidth impact, and you avoid the arms race with VPN adoption. Pitfall two: ignoring the compliance dimension. If you're operating in the UK, you have obligations under the Investigatory Powers Act 2016 to retain connection records. Your network architecture must support this. Ensure your logging infrastructure captures the data required for compliance, and that your audit trail is tamper-evident. Pitfall three: failing to account for IoT growth. Building management systems, smart meters, CCTV, and access control are increasingly IP-connected. These devices must be on isolated VLANs with strict firewall policies. A compromised smart thermostat should never be able to reach your student authentication infrastructure. Time for a rapid-fire Q&A. Question one: Should we publish our bandwidth policies to residents? Yes, absolutely. Transparency reduces complaints and sets expectations. Include bandwidth allocations in your tenancy agreement or welcome pack. Question two: How do we handle VPN traffic that bypasses our QoS marking? Implement traffic shaping at the IP flow level, not just at the application layer. VPN-encapsulated traffic can still be rate-limited based on flow characteristics, even if you can't inspect the payload. Question three: What's the right uplink sizing for student accommodation? A reasonable baseline is 1 megabit-per-second per bed, with the ability to burst to 3 megabits-per-second. For a 400-bed property, that means a minimum 400 megabits-per-second uplink with a 1.2 gigabit-per-second burst capacity. To summarise the key takeaways from today's briefing. Flat networks fail at scale โ€” segment your traffic with VLANs from day one. Move from per-device to per-user identity-based policies to prevent gaming of your bandwidth allocations. Implement dynamic traffic shaping with time-of-day rules rather than static caps. Use DSCP marking at the access point edge to enforce QoS before traffic hits your core. Deploy application-layer visibility to make data-driven policy decisions. And don't block peer-to-peer โ€” throttle and deprioritise it instead. For the full technical reference guide, including architecture diagrams, configuration templates, and worked implementation examples, visit the Purple website. Until next time, keep your networks fast, your policies fair, and your residents connected.

header_image.png

Executive Summary

Managing WiFi bandwidth in student accommodation is one of the most technically demanding challenges in the residential property sector. A single 400-bed block can generate over 2,800 concurrent device connections during peak hours, with traffic profiles spanning latency-sensitive video conferencing, high-throughput streaming, online gaming, and background IoT telemetry โ€” all competing for the same uplink capacity.

The failure mode is predictable: flat network architectures with per-device throttling degrade during peak hours, generate disproportionate support overhead, and expose operators to compliance risk. The solution is equally well-defined: VLAN segmentation, identity-based QoS policy enforcement, dynamic traffic shaping, and application-layer analytics.

This guide provides the technical architecture, implementation sequence, and operational decision frameworks required to deploy a bandwidth management strategy that scales. Whether you are remediating a legacy flat network or designing a greenfield deployment, the principles here apply across vendor stacks and property sizes. For operators already using Guest WiFi infrastructure, these policies integrate directly with existing captive portal and authentication workflows.


Technical Deep-Dive

The Contention Problem

The fundamental challenge in student accommodation is not raw bandwidth โ€” most operators have access to gigabit uplinks at competitive pricing. The challenge is contention management: ensuring that the available capacity is distributed fairly and intelligently across hundreds of concurrent users with wildly different traffic profiles.

A flat network architecture โ€” a single SSID, a single IP subnet, a global per-device cap โ€” fails for three compounding reasons. First, per-device limits are trivially gamed: a student with seven devices effectively receives seven times the allocation. Second, without traffic classification, a single user running a large torrent download can saturate the uplink queue and introduce latency for every other user on the segment. Third, without application-layer visibility, the operator has no data to inform policy decisions or identify chronic offenders.

VLAN Segmentation Architecture

The first architectural requirement is logical network separation using IEEE 802.1Q VLANs. At minimum, a student accommodation deployment should operate three distinct VLANs:

VLAN Purpose Bandwidth Policy Security Posture
VLAN 10 โ€” Students Resident internet access Per-user cap, dynamic burst Isolated, internet-only
VLAN 20 โ€” Staff/Admin Property management systems Dedicated allocation Restricted access
VLAN 30 โ€” IoT/BMS Building management, CCTV, access control Strict rate limit Air-gapped from student VLAN

This segmentation is non-negotiable from both a performance and a security standpoint. Under IEEE 802.1Q, each VLAN operates as a separate broadcast domain, eliminating cross-segment broadcast storms and preventing lateral movement between user classes. A compromised student device cannot reach building management infrastructure if the VLANs are correctly configured with inter-VLAN routing policies at the firewall layer.

qos_architecture_diagram.png

Quality of Service Policy Design

Once traffic is segmented, QoS policies must be applied to prioritise latency-sensitive applications over bulk transfers. The industry standard mechanism is Differentiated Services Code Point (DSCP) marking, defined in RFC 2474. Packets are classified and marked at the access point โ€” the ingress point โ€” before they reach the core switching fabric.

The recommended DSCP marking scheme for student accommodation is as follows:

Traffic Class Application Examples DSCP Value Per-Hop Behaviour
Voice VoIP, video calls EF (46) Expedited Forwarding
Interactive Video Video conferencing, remote desktop AF41 (34) Assured Forwarding
Streaming Video Netflix, YouTube, iPlayer AF21 (18) Assured Forwarding
Web / Email HTTP/S, SMTP, DNS CS0 (0) Best Effort
Bulk / P2P Torrents, large file transfers CS1 (8) Background / Scavenger

Critically, DSCP marking must occur at the access point layer, not at the core router. If classification is deferred to the core, packets have already traversed the wireless medium and the distribution switching fabric without priority treatment, negating the benefit.

Identity-Based Policy Enforcement

The most impactful architectural decision in a student accommodation deployment is moving from per-device to per-user bandwidth policy enforcement. The average student brings seven connected devices to their accommodation. Per-device caps are therefore both ineffective and unfair: a student with a single laptop receives one-seventh of the effective allocation of a student with a full device suite.

The correct approach is IEEE 802.1X authentication, ideally with WPA3-Enterprise for the cryptographic security benefits.

Under this model:

  1. The student authenticates once using their institutional or property credentials via a RADIUS server.
  2. All subsequent device registrations are tied to that user identity via MAC Authentication Bypass (MAB) for headless devices.
  3. The bandwidth policy โ€” say, 25 Mbps aggregate โ€” applies to the sum of all sessions associated with that user identity.
  4. When the aggregate exceeds the allocation, the shaping policy applies proportionally across all active sessions.

This model is fundamentally more scalable and equitable than per-MAC throttling, and it provides the identity layer required for compliance logging under the Investigatory Powers Act 2016.

Application-Layer Visibility

Deep Packet Inspection (DPI) at the gateway provides the application-layer telemetry required to make intelligent, data-driven policy decisions. Without DPI, bandwidth management is essentially blind: you can see that your uplink is saturated, but you cannot determine which applications or users are responsible.

With DPI-enabled analytics โ€” such as those provided by WiFi Analytics โ€” operators gain visibility into application distribution, peak usage patterns, top consumers, and traffic trends over time. This data directly informs policy decisions: if 55% of peak-hour traffic is attributable to four streaming platforms, you can apply application-specific rate limits during defined windows without impacting video conferencing or academic platforms.


Implementation Guide

Phase 1: Baseline Assessment (Weeks 1โ€“2)

Before deploying any new policies, establish a 14-day baseline of current network behaviour. Deploy a network management platform with DPI capabilities and capture: peak concurrent device counts, application distribution by traffic volume, per-floor and per-AP utilisation, and uplink saturation frequency. This data is the foundation for all subsequent policy decisions and provides the before/after comparison required to demonstrate ROI.

Phase 2: VLAN Segmentation Deployment (Weeks 3โ€“4)

Deploy the three-VLAN architecture described above. This requires configuration changes at the core router/firewall (inter-VLAN routing and ACL policies), distribution switches (trunk port configuration and VLAN tagging), and access points (SSID-to-VLAN mapping). For existing deployments, this can typically be completed in a maintenance window without requiring new hardware, provided the existing switching infrastructure supports 802.1Q trunking.

Phase 3: QoS Policy Activation (Week 5)

Activate DSCP marking at the access point layer and configure per-hop behaviour at the core router. Validate that DSCP markings are being honoured end-to-end using a packet capture tool. Common failure modes at this stage include upstream ISP routers remarking or stripping DSCP values โ€” verify with your ISP whether DSCP is honoured on your transit link.

Phase 4: Identity-Based Bandwidth Policies (Weeks 6โ€“7)

Migrate authentication from PSK or MAC-based access to 802.1X. Deploy a RADIUS server (FreeRADIUS or a cloud-hosted equivalent) and configure per-user bandwidth attributes using the standard RADIUS attributes: WISPr-Bandwidth-Max-Up and WISPr-Bandwidth-Max-Down. Implement a MAB self-registration portal for headless devices. Test with a pilot floor before full rollout.

Phase 5: Dynamic Shaping Rules (Week 8)

Configure time-of-day shaping rules on the core router or bandwidth management appliance. A recommended policy structure:

  • Off-peak (00:00โ€“08:00): Burst to 2ร— baseline allocation, P2P unrestricted.
  • Standard (08:00โ€“18:00): Baseline allocation, P2P throttled to 5 Mbps.
  • Peak (18:00โ€“23:00): Baseline allocation, P2P throttled to 1 Mbps, streaming capped at 8 Mbps, video conferencing prioritised.

bandwidth_policy_comparison.png


Best Practices

Publish your bandwidth policy. Transparency reduces resident complaints and sets expectations. Include bandwidth allocations and fair-use policies in tenancy agreements and welcome packs. This is also a risk mitigation measure: documented policies reduce exposure in the event of a resident dispute.

Size your uplink correctly. A practical baseline is 1 Mbps per bed, with burst capacity to 3 Mbps per bed. For a 400-bed property, this means a minimum 400 Mbps uplink with a 1.2 Gbps burst circuit. Underprovisioning the uplink makes all downstream QoS policies less effective.

Do not block P2P traffic entirely. Blanket bans drive users to commercial VPN services, which blinds your DPI analytics and makes traffic management significantly harder. Throttle P2P to a scavenger-class allocation (1โ€“2 Mbps) and deprioritise it. You retain visibility, reduce the bandwidth impact, and avoid the arms race with VPN adoption.

Plan for IoT growth. Building management systems, smart meters, CCTV, and access control are increasingly IP-connected. Ensure these devices are on isolated VLANs with strict firewall egress policies. Review your IoT VLAN policy annually as the device population grows.

Maintain an audit trail. Under the Investigatory Powers Act 2016, UK operators are required to retain connection records. Ensure your logging infrastructure captures the data required for compliance, and that your audit trail is tamper-evident. For a detailed breakdown of audit trail requirements, see Explain what is audit trail for IT Security in 2026 .


Troubleshooting & Risk Mitigation

Common Failure Mode 1: DSCP Remarking by ISP

Many ISPs remark or strip DSCP values at the transit boundary, rendering your QoS policies ineffective for traffic traversing the internet. Mitigation: verify DSCP behaviour with your ISP before relying on it for end-to-end QoS. For internal traffic (e.g., local caching servers), DSCP will always be honoured. For internet-bound traffic, rely on queue management and shaping at your own gateway rather than expecting DSCP to be honoured upstream.

Common Failure Mode 2: DHCP Pool Exhaustion

With seven devices per student and hundreds of residents, DHCP pool exhaustion is a real operational risk. Ensure your student VLAN subnet is sized with sufficient headroom: a /21 (2,046 usable addresses) is a reasonable minimum for a 200-bed property. Implement short DHCP lease times (4โ€“8 hours) to reclaim addresses from inactive devices promptly.

Common Failure Mode 3: VPN Bypass

Students using commercial VPN services will encrypt their traffic, bypassing application-layer classification. Mitigation: implement flow-based shaping at the IP level โ€” VPN traffic can still be rate-limited based on flow volume and duration, even without payload inspection. Additionally, ensure your P2P throttling policy applies to encrypted flows, not just identifiable P2P protocols.

Common Failure Mode 4: Connectivity Issues Post-Segmentation

After VLAN segmentation, residents may encounter connectivity issues if their devices are incorrectly placed in the wrong VLAN or if inter-VLAN routing is misconfigured. For a structured troubleshooting approach to connectivity issues, refer to Solving the Connected but No Internet Error on Guest WiFi .


ROI & Business Impact

The business case for a properly architected bandwidth management strategy is straightforward. The primary cost drivers are support overhead and resident satisfaction, both of which are directly impacted by network performance.

In a 400-bed deployment running a flat network, support ticket volumes of 30โ€“50 per week during term time are common. Post-remediation deployments consistently report ticket reductions of 60โ€“80%, representing a significant reduction in IT staff time and third-party support costs.

Resident satisfaction scores โ€” increasingly a competitive differentiator in the purpose-built student accommodation (PBSA) market โ€” are directly correlated with network performance. Properties with well-managed networks report higher renewal rates and stronger occupancy.

From a compliance standpoint, the cost of non-compliance with the Investigatory Powers Act 2016 or GDPR data handling requirements significantly exceeds the cost of implementing compliant logging infrastructure. The identity-based architecture described in this guide provides the audit trail required for compliance as a by-product of the bandwidth management implementation.

For operators in the hospitality sector managing mixed-use properties โ€” student accommodation with ground-floor retail or food and beverage โ€” the same VLAN segmentation principles apply, with the addition of PCI DSS compliance requirements for any payment-processing network segments.

The WiFi Analytics layer adds a further dimension of ROI: application-layer traffic data can inform infrastructure investment decisions, identify capacity upgrade triggers, and provide the evidence base for renegotiating ISP contracts based on actual usage patterns rather than estimates.

Key Definitions

VLAN (Virtual Local Area Network)

A logical network segment created within a physical switching infrastructure using IEEE 802.1Q tagging. Each VLAN operates as a separate broadcast domain, providing traffic isolation between user classes without requiring separate physical hardware.

IT teams use VLANs to separate student, staff, and IoT traffic on the same physical infrastructure. Without VLAN segmentation, a flat network exposes all traffic classes to each other and makes per-class bandwidth policies impossible to enforce cleanly.

QoS (Quality of Service)

A set of network mechanisms that prioritise certain traffic types over others to ensure latency-sensitive applications (VoIP, video conferencing) receive preferential treatment during periods of congestion.

In student accommodation, QoS is the difference between video conferencing being usable during peak hours and being unusable. Without QoS, a single user running a large download can introduce latency for every other user on the segment.

DSCP (Differentiated Services Code Point)

A 6-bit field in the IP packet header, defined in RFC 2474, used to classify packets into traffic classes. Each class receives a defined per-hop behaviour (PHB) at each network device โ€” Expedited Forwarding for voice, Assured Forwarding for video, Best Effort for standard web traffic.

DSCP is the standard mechanism for implementing QoS in enterprise networks. IT teams configure access points to mark packets with the appropriate DSCP value at ingress, ensuring priority treatment is applied consistently across the network.

IEEE 802.1X

An IEEE standard for port-based network access control that provides an authentication framework for devices connecting to a LAN or WLAN. It uses the Extensible Authentication Protocol (EAP) and requires a RADIUS server for credential validation.

802.1X is the foundation of identity-based bandwidth policy enforcement. When a student authenticates via 802.1X, their identity is known to the network, enabling per-user bandwidth policies rather than per-device policies.

Traffic Shaping

A bandwidth management technique that controls the rate and timing of traffic flows to conform to a defined policy. Unlike policing (which drops excess traffic), shaping queues excess traffic and transmits it when capacity is available.

Traffic shaping is preferable to policing for TCP-based traffic (web, streaming) because it avoids triggering TCP retransmission, which wastes bandwidth. Policing is appropriate for UDP-based traffic (P2P, some gaming) where retransmission is not a factor.

DPI (Deep Packet Inspection)

A network analysis technique that examines the full content of packets (beyond the header) to identify the application or protocol generating the traffic. DPI enables application-aware QoS policies and provides granular traffic analytics.

DPI is the technology that enables an operator to distinguish between Netflix traffic and a video call, even when both use HTTPS on port 443. Without DPI, application-aware bandwidth policies are not possible.

MAB (MAC Authentication Bypass)

A fallback authentication mechanism for devices that do not support IEEE 802.1X. The device's MAC address is used as the authentication credential, validated against a RADIUS server or local database.

MAB is used for headless devices in student accommodation โ€” gaming consoles, smart TVs, IoT sensors โ€” that cannot perform 802.1X authentication. Combined with a self-registration portal, MAB enables these devices to be tied to a user identity and subject to the same per-user bandwidth policies.

Bandwidth Contention

The condition that occurs when multiple users or devices compete for the same finite bandwidth resource, resulting in reduced throughput and increased latency for all parties. Contention is the root cause of most perceived network performance problems in high-density environments.

Understanding contention is essential for diagnosing bandwidth problems. A network with a 1 Gbps uplink and 400 concurrent users each consuming 3 Mbps is in contention (1.2 Gbps demand vs 1 Gbps supply). QoS and traffic shaping manage the contention; they do not eliminate it.

WPA3-Enterprise

The latest generation of the Wi-Fi Protected Access security protocol for enterprise networks, defined by the Wi-Fi Alliance. WPA3-Enterprise mandates 192-bit minimum-strength cryptography and provides stronger protection against offline dictionary attacks compared to WPA2.

WPA3-Enterprise is the recommended authentication mode for student accommodation deployments using 802.1X. It provides the cryptographic security required for GDPR compliance and protects against credential interception on the wireless medium.

Worked Examples

A 400-bed purpose-built student accommodation (PBSA) block in Manchester is running a flat network with a single SSID and a global 10 Mbps per-device cap. During peak hours (19:00โ€“23:00), the network is effectively unusable for video conferencing. Support tickets are running at 40 per week. The operator has a 1 Gbps uplink and a budget for software configuration changes only โ€” no new hardware. How do you remediate this?

Step 1 โ€” Baseline audit (Days 1โ€“7): Deploy DPI-enabled monitoring on the existing gateway to capture application distribution, peak concurrent device counts, and per-AP utilisation. This establishes the evidence base and identifies the primary bandwidth consumers.

Step 2 โ€” VLAN segmentation (Days 8โ€“14): Configure three VLANs on the existing switching infrastructure (assuming 802.1Q-capable switches, which is standard in any post-2015 deployment). Map the student SSID to VLAN 10, create a staff SSID mapped to VLAN 20, and migrate IoT devices to VLAN 30. Configure inter-VLAN routing at the firewall with appropriate ACLs.

Step 3 โ€” QoS activation (Day 15): Enable DSCP marking at the access point layer. Classify video conferencing traffic (Zoom, Teams, Google Meet) as AF41. Classify streaming as AF21. Classify P2P as CS1. Validate with a packet capture.

Step 4 โ€” Per-user bandwidth policy (Days 16โ€“21): Migrate authentication to 802.1X using the existing RADIUS infrastructure (or deploy FreeRADIUS on a VM). Set per-user bandwidth attributes: 25 Mbps aggregate during peak, 50 Mbps off-peak. Implement MAB portal for headless devices.

Step 5 โ€” Time-of-day shaping (Day 22): Configure peak-hour rules: P2P throttled to 1 Mbps, streaming capped at 8 Mbps per user, video conferencing prioritised with guaranteed minimum 5 Mbps per active session.

Outcome: Within 30 days, support tickets dropped by 78% (from 40 to 9 per week). Average peak-hour throughput per user increased by 140% despite no change to the physical uplink. Video conferencing became reliably usable during peak hours.

Examiner's Commentary: This scenario illustrates the critical insight that bandwidth problems in dense residential networks are almost never caused by insufficient uplink capacity โ€” they are caused by poor traffic management. The 1 Gbps uplink was more than adequate; the problem was contention and the absence of traffic classification. The remediation sequence is deliberately ordered: establish baseline data first, then segment, then classify, then enforce identity-based policies. Attempting to implement QoS before segmentation is a common mistake that results in policies being applied inconsistently across mixed traffic types. The 78% ticket reduction is a realistic outcome based on comparable deployments; the key driver is the shift from per-device to per-user policy enforcement, which eliminates the most common gaming vector.

A 1,200-bed university halls of residence in Edinburgh has a mixed infrastructure: legacy 802.11ac access points on floors 1โ€“4 and newer Wi-Fi 6 hardware on floors 5โ€“8. There is no application-layer visibility, and the network management team has no baseline data. The university IT director wants to reduce peak-hour congestion by 30% within 90 days without a full hardware refresh. How do you approach this?

Phase 1 โ€” Telemetry deployment (Days 1โ€“30): Deploy a unified network management platform with DPI capabilities across all access points, including the legacy 802.11ac hardware. Most enterprise NMS platforms support mixed-generation hardware via SNMP and syslog. Capture 30 days of baseline data: application distribution, per-floor utilisation, peak concurrent device counts, and top bandwidth consumers by user identity.

Phase 2 โ€” Data analysis and policy design (Days 31โ€“35): Analyse the baseline data. In this scenario, the data revealed that 55% of peak-hour traffic was attributable to four streaming platforms. Design application-aware QoS policies: streaming platforms throttled to 8 Mbps per user during 18:00โ€“23:00, video conferencing and academic platforms (VLEs, library databases) excluded from throttling and given AF41 priority.

Phase 3 โ€” Policy deployment (Days 36โ€“50): Deploy QoS policies starting with the Wi-Fi 6 floors (5โ€“8) as a controlled pilot. Monitor for 14 days. Validate that peak-hour congestion metrics improve before rolling out to legacy floors.

Phase 4 โ€” Identity migration (Days 51โ€“75): Migrate authentication to 802.1X with per-user bandwidth enforcement. This is the most operationally complex phase: coordinate with the university IT team for RADIUS integration with the student identity provider. Implement MAB self-registration for gaming consoles and smart TVs.

Phase 5 โ€” Validation and reporting (Days 76โ€“90): Compare post-implementation metrics against the 30-day baseline. Report on peak-hour congestion reduction, support ticket volume, and application distribution changes.

Outcome: 35% reduction in peak-hour congestion (exceeding the 30% target), measurable improvement in resident satisfaction survey scores, and a documented evidence base for the hardware refresh business case.

Examiner's Commentary: The phased approach is essential here for two reasons: the mixed hardware environment requires careful validation at each stage, and the 90-day timeline is tight. Starting the pilot on the Wi-Fi 6 floors is the correct decision because these APs have more sophisticated QoS capabilities and will produce cleaner results. The 30-day baseline phase is non-negotiable โ€” without it, you cannot demonstrate ROI or make defensible policy decisions. The identity migration phase is correctly placed last because it has the highest operational risk (authentication failures affect all residents) and requires the most coordination with third-party systems. The 35% congestion reduction is achievable through application-aware throttling alone, before the identity migration is complete.

Practice Questions

Q1. You are the IT director for a 600-bed PBSA operator. Your current network uses WPA2-PSK with a shared password changed monthly. Students are complaining about poor performance during evening hours. Your uplink is 500 Mbps. Before spending any budget, what is the first thing you should deploy, and what specific data are you trying to capture?

Hint: You cannot make defensible policy decisions without baseline data. What tool gives you application-layer visibility without requiring new hardware?

View model answer

Deploy a DPI-enabled network monitoring tool on the existing gateway โ€” most enterprise gateway appliances support this via software activation or a management platform integration. Run it for 14โ€“30 days to capture: (1) application distribution by traffic volume during peak hours, (2) peak concurrent device counts, (3) per-AP utilisation to identify hotspots, and (4) top bandwidth consumers by MAC address. This data will tell you whether the problem is uplink saturation (requiring a capacity upgrade or traffic shaping), contention on specific APs (requiring AP placement changes or load balancing), or a small number of heavy users consuming disproportionate bandwidth (requiring per-user policy enforcement). Without this data, any remediation is guesswork. The baseline also provides the before/after comparison required to demonstrate ROI to the property owner.

Q2. A student in a 300-bed hall reports that their gaming console cannot connect to the network after you migrated authentication to 802.1X. They are using a PlayStation 5, which does not support 802.1X natively. How do you resolve this without creating a security exception that bypasses your identity-based bandwidth policies?

Hint: The solution must maintain the link between the device and the student's identity for bandwidth policy enforcement purposes.

View model answer

Implement MAC Authentication Bypass (MAB) with a self-service device registration portal. The workflow: (1) The student visits a captive portal URL (e.g., register.accommodation.ac.uk) from an authenticated device (their laptop or phone). (2) They enter the MAC address of their gaming console and confirm ownership. (3) The portal adds the MAC address to the RADIUS database, associated with the student's user identity. (4) When the PlayStation connects, the network performs MAB โ€” it sends the device's MAC address to the RADIUS server, which returns the associated user identity and bandwidth policy attributes. (5) The console is placed in the same VLAN as the student's other devices and subject to the same aggregate per-user bandwidth policy. This approach maintains identity linkage for bandwidth enforcement, provides an audit trail for compliance, and does not require the student to contact IT support. Ensure the registration portal validates that the MAC address is not already registered to another user to prevent address spoofing.

Q3. Your DPI analytics reveal that 62% of peak-hour bandwidth on your student accommodation network is consumed by video streaming (Netflix, Disney+, YouTube). Your uplink is at 85% utilisation during peak hours. You have two options: (A) upgrade the uplink to 2ร— capacity, or (B) implement application-aware traffic shaping to cap streaming at 8 Mbps per user during peak hours. Which do you recommend, and why?

Hint: Consider both the short-term cost and the long-term scalability of each approach. What happens to demand if you simply increase capacity?

View model answer

Recommend Option B (application-aware traffic shaping) as the primary intervention, with Option A as a medium-term follow-on if required. The reasoning: (1) Increasing uplink capacity without traffic shaping does not solve the underlying problem โ€” it defers it. Streaming consumption will expand to fill available capacity (Jevons paradox applied to bandwidth), and you will be back at 85% utilisation within 12โ€“18 months. (2) Capping streaming at 8 Mbps per user during peak hours has a negligible impact on user experience โ€” Netflix recommends 5 Mbps for HD streaming and 25 Mbps for 4K. An 8 Mbps cap delivers a good HD experience. (3) The 62% streaming share means that an 8 Mbps per-user cap on streaming, applied to a typical peak concurrency of 200 active users, reduces streaming demand from approximately 425 Mbps to approximately 160 Mbps โ€” a 62% reduction in streaming traffic, bringing total utilisation to approximately 55%. (4) The cost of traffic shaping configuration is near-zero if the gateway hardware supports it; the cost of a 2ร— uplink upgrade is a recurring OpEx increase. Implement traffic shaping first, measure the impact over 30 days, and then make an evidence-based decision on whether an uplink upgrade is still required.