NB-IoT offers three deployment modes that leverage existing cellular infrastructure: in-band (allocating one PRB within an LTE carrier, most common), guard-band (using unused spectrum at LTE carrier edges), and standalone (dedicated 200 kHz channel, often repurposed GSM spectrum), each with different trade-offs between deployment speed, spectrum efficiency, and impact on existing LTE services.
30.1 Learning Objectives
By the end of this chapter, you will be able to:
Compare Deployment Modes: Distinguish in-band, guard-band, and standalone NB-IoT configurations by spectrum usage and trade-offs
Calculate Resource Allocation: Compute PRB capacity impact for NB-IoT within LTE carriers and justify the 2% loss trade-off
Select Optimal Mode: Evaluate deployment options against infrastructure constraints and select the best configuration for a given scenario
Design Spectrum Strategy: Construct a phased deployment plan that exploits existing 4G/LTE infrastructure for NB-IoT rollout
In-band NB-IoT: NB-IoT using resource blocks within the LTE carrier; most common deployment mode, requires LTE carrier coordination.
Cat-NB1 vs Cat-NB2: 3GPP Release 13 (Cat-NB1) and Release 14 (Cat-NB2) NB-IoT device categories; Cat-NB2 adds multicast, 2 antenna ports, and OTDOA positioning.
Non-IP Data Delivery (NIDD): An NB-IoT data path bypassing the IP stack, suitable for ultra-simple devices sending small non-IP data via SMS-like delivery.
Network Coverage Classes: NB-IoT coverage classes (Normal Coverage, Extended Coverage Level 1, Extended Coverage Level 2) with progressively more signal repetitions for deeper indoor coverage.
30.3 NB-IoT Deployment Modes Overview
NB-IoT supports three flexible deployment options leveraging existing cellular infrastructure. Each mode offers different trade-offs between deployment speed, spectrum efficiency, and LTE impact.
Figure 30.1: NB-IoT Deployment Modes: In-Band, Guard-Band, and Standalone
30.4 Deployment Mode Selection
Alternative View: Deployment Mode Decision Flowchart
This flowchart guides deployment mode selection based on existing infrastructure and capacity constraints.
30.5 Detailed Deployment Modes
30.5.1 Mode Comparison Table
Mode
Spectrum Usage
Pros
Cons
1. In-Band
Uses PRB from LTE carrier (e.g., PRB 25 from 10 MHz/50 PRB carrier)
Rapid deployment, Software upgrade only
2% LTE capacity loss
2. Guard-Band
Uses unused guard band between LTE carriers (200 kHz)
No LTE impact, Efficient spectrum use
Limited scalability
3. Standalone
Repurposed GSM carrier (200 kHz @ 900 MHz)
No LTE impact, Excellent coverage
Spectrum refarming needed
30.5.2 In-Band Deployment Example
For a 10 MHz LTE Carrier:
PRB Range
Usage
Bandwidth
PRB 1-24
LTE Data
4.32 MHz
PRB 25
NB-IoT
180 kHz
PRB 26-50
LTE Data
4.5 MHz
For Beginners: Understanding PRBs
What is a PRB (Physical Resource Block)?
Think of an LTE carrier like a highway with multiple lanes. Each lane (PRB) is exactly 180 kHz wide:
A 10 MHz LTE carrier has 50 PRBs (50 lanes)
NB-IoT uses just 1 PRB (1 lane)
This leaves 49 PRBs for regular LTE traffic
The 2% Calculation:
1 PRB out of 50 PRBs = 1/50 = 2%
This small sacrifice enables thousands of IoT devices
Real-World Analogy: It’s like reserving one lane on a 50-lane highway exclusively for bicycles. You lose 2% of car capacity, but gain dedicated infrastructure for a different type of traffic.
Net benefit: \(\$50{,}000 - \$875 = \$49{,}125/\text{month}\) per cell
Key insight: NB-IoT generates 57x more revenue than the lost LTE capacity while using only 0.7% of that capacity. The massive overprovisioning (180 kHz supports 40 MB/day but carries only 1.2 GB/month) is by design - IoT devices transmit infrequently, allowing extreme oversubscription.
Lost LTE capacity: ~540 kbps continuous
Gained IoT capacity: 50,000+ devices per cell
Revenue from IoT subscriptions often exceeds lost LTE revenue
This makes in-band mode the most popular initial deployment strategy.
30.7 Standalone Mode Deep Dive
Standalone mode uses dedicated spectrum, typically repurposed from legacy 2G/GSM networks:
Guard-band deployment places NB-IoT in the unused spectrum at LTE carrier edges:
Requirements:
Guard band width >= 200 kHz
Careful interference management at band edges
May vary by LTE carrier configuration
Advantages:
Zero impact on LTE data capacity
Uses otherwise wasted spectrum
No spectrum refarming needed
Limitations:
Not all LTE configurations have suitable guard bands
Limited scalability (can’t add more NB-IoT capacity easily)
More complex RF planning
30.9 Worked Example: Deployment Mode Selection for Smart Metering
Scenario: A European water utility needs to connect 200,000 smart meters across a mixed urban-rural service area. Their carrier partner has both active LTE (Band 20, 800 MHz) and legacy GSM (Band 8, 900 MHz) infrastructure.
30.9.1 Step 1: Analyze Coverage Requirements
Putting Numbers to It
Why does the rural area require so many cells (2,400) compared to urban (40)? Let’s calculate the coverage area per cell.
Urban cell coverage:
Cell radius: 1.5 km (dense base station deployment)
Actual: 2,400 cells means average 50 meters/cell → \(\frac{4{,}800}{2{,}400} = 2 \text{ km}^2\) per cell
Key insight: Rural “cell” density of 50 meters/cell implies coverage holes (2 km² << 227 km² possible). This is why standalone mode on 900 MHz (12 km radius = 452 km² coverage) is attractive - reduces cells needed from 2,400 to \(\frac{4{,}800}{452} = 11\) cells.
Service area: 8,000 km² (60% rural, 30% suburban, 10% urban)
Meters per cell (average):
Urban: 500 meters/cell (high density, good coverage)
Suburban: 200 meters/cell (medium density)
Rural: 50 meters/cell (low density, coverage-limited)
Total cells needed:
Urban: 200,000 x 0.10 / 500 = 40 cells
Suburban: 200,000 x 0.30 / 200 = 300 cells
Rural: 200,000 x 0.60 / 50 = 2,400 cells
Total: 2,740 cells
30.9.2 Step 2: Evaluate Each Mode
Option A: In-Band (Band 20, 800 MHz LTE)
Existing LTE coverage: 2,200 cells (covers urban + suburban + some rural)
Rural gap: ~540 cells without LTE coverage
LTE carrier: 10 MHz = 50 PRBs
NB-IoT allocation: 1 PRB = 2% capacity loss
Lost revenue: 540 kbps x 2,200 cells x $0.01/Mbps/month = $14,256/year
Deployment: Software upgrade only (2-3 months)
Option B: Standalone (Band 8, 900 MHz GSM)
Legacy GSM coverage: 3,100 cells (includes deep rural)
Covers entire service area including 540 rural cells missing LTE
GSM carrier: 200 kHz (perfect fit for NB-IoT standalone)
Coverage enhancement: ~+1 dB free-space path loss advantage at 900 MHz vs 800 MHz
(20×log10(900/800) ≈ 1 dB; real-world gain includes additional building penetration
at lower frequency, contributing to larger effective cell radius for basement meters)
Deployment: Requires GSM spectrum refarming (6-12 months)
In-Band: 2,200 cells (urban + suburban) - 2 months to deploy
Standalone: 540 cells (rural) - 8 months to deploy
Total coverage: 2,740 cells = 100% service area
Phased rollout: Start revenue in urban areas while rural deploys
30.9.3 Step 3: Cost Comparison (5-Year TCO)
Cost Item
In-Band Only
Standalone Only
Hybrid
Software upgrade
$220K
$0
$220K
Spectrum refarming
$0
$930K
$162K
New tower equipment
$0
$0
$0
Coverage gap solution
$810K (new rural LTE)
$0
$0
Lost LTE revenue
$71K
$0
$52K
Total 5-year
$1,101K
$930K
$434K
30.9.4 Recommendation
Hybrid deployment wins – deploy in-band across existing LTE footprint for fast urban/suburban rollout, then standalone on GSM spectrum for rural coverage. The phased approach generates revenue 6 months earlier than standalone-only, and avoids the $810K cost of extending LTE to rural areas.
30.10 Real-World Operator Deployment Patterns
Major NB-IoT operators have chosen different modes based on their spectrum assets:
Fast deployment via software upgrade; existing LTE covers most meters
T-Mobile
USA
Guard-Band
700 MHz (LTE Band 12)
Available guard band; no LTE impact; good building penetration
Deutsche Telekom
Germany
In-Band
900 MHz (LTE Band 8)
Wide existing LTE coverage at low frequency
KT (Korea Telecom)
South Korea
In-Band
900 MHz
Dense urban deployment; software-only upgrade
Key Pattern: In-band is preferred for quick deployment in well-covered areas. Standalone is chosen when legacy 2G/3G spectrum is being retired, providing both IoT coverage and a migration path. Guard-band is niche but valuable when suitable guard spectrum exists.
30.11 Knowledge Check
30.12 Real-World Case Study: Vodafone’s NB-IoT Rollout Across Europe
Vodafone’s European NB-IoT deployment (2017-2021) provides a textbook example of how operators select deployment modes based on existing spectrum assets and market priorities across different countries.
Phase 1 – In-Band Launch (2017-2018):
Vodafone launched NB-IoT in Spain, Ireland, and the Netherlands using in-band deployment on existing LTE Band 20 (800 MHz). This mode required only a software upgrade to existing eNodeB base stations – no new hardware, no spectrum refarming.
Spain rollout timeline:
Week 0: Software update package distributed to 8,200 eNodeBs
Week 2: First 2,000 cells activated (major cities: Madrid, Barcelona)
Week 6: 5,500 cells (all urban areas)
Week 10: 8,200 cells (national coverage)
LTE capacity impact: 1 PRB from 50-PRB carrier = 2% loss
Revenue impact: -$180K/year LTE revenue
NB-IoT revenue (Year 1): +$2.1M from utility metering contracts
Net impact: +$1.92M in first year
Phase 2 – Standalone for Rural Coverage (2019-2020):
For rural areas in Germany and Italy where LTE coverage was sparse, Vodafone deployed standalone NB-IoT on GSM Band 8 (900 MHz) spectrum being freed by the 2G sunset.
Germany rural deployment:
GSM 900 MHz towers: 4,100 (covering agricultural regions)
Spectrum refarming time: 8 months per region
Coverage advantage: 900 MHz provides ~+1 dB free-space path loss gain vs 800 MHz
(20×log10(900/800) ≈ 1 dB; ~3–5% larger effective cell radius in typical terrain)
= ~9 km rural range vs 8.5 km on 800 MHz (modest gain; main benefit is existing
GSM infrastructure penetrating areas without LTE coverage)
Target market: Agricultural sensors (soil moisture, weather stations)
Devices connected (Year 1): 340,000 sensors across 12,000 farms
Phase 3 – Guard-Band in Dense Urban (2020-2021):
In central London and Amsterdam where LTE capacity was already strained, Vodafone used guard-band deployment to avoid any LTE throughput impact.
London deployment:
LTE Band 20 carrier: 10 MHz with 250 kHz guard bands on each side
Guard-band NB-IoT: 200 kHz placed in upper guard band
LTE capacity impact: 0% (using otherwise wasted spectrum)
Target: Smart parking sensors in 45,000 spaces
Limitation: Only 1 NB-IoT carrier per LTE carrier guard band
(can't scale beyond ~50,000 devices per cell without switching to in-band)
Key Deployment Lessons:
Lesson
Detail
Start with in-band
Software-only deployment generates revenue while planning standalone
Standalone requires 2G sunset
Cannot refarm GSM spectrum until 2G services are migrated – plan 12-18 months ahead
Guard-band is situational
Only viable where LTE guard bands are wide enough (>200 kHz) and device density is moderate
Hybrid wins at scale
By 2021, Vodafone ran all three modes simultaneously across 9 countries, selecting per-cell based on local spectrum conditions
Common Pitfalls
1. Selecting Deployment Mode Without Carrier Coordination
In-band and guard-band NB-IoT deployment modes must be coordinated with the LTE carrier — they cannot be independently configured by device deployments. Confirm supported deployment modes with your carrier before module procurement.
2. Not Testing All Coverage Classes
NB-IoT devices may work in normal coverage conditions but fail in extended coverage scenarios. Test devices in environments requiring Extended Coverage Level 2 (basements, underground) to validate behavior under maximum repetition.
Label the Diagram
💻 Code Challenge
30.13 Summary
Three deployment modes provide flexibility: in-band (fastest), guard-band (zero LTE impact), standalone (best coverage)
In-band mode is most popular due to software-only deployment, with acceptable 2% LTE capacity trade-off
PRB allocation of 180 kHz supports 50,000+ IoT devices per cell while minimizing LTE impact
Standalone mode offers excellent coverage at low frequencies (900 MHz) when GSM spectrum is available
Guard-band mode efficiently uses otherwise wasted spectrum but has limited scalability
Mode selection depends on existing infrastructure, spectrum assets, and coverage requirements
NB-IoT Fundamentals introduces the 180 kHz bandwidth requirement. Deployment modes chapter shows three ways to allocate this bandwidth within existing cellular spectrum.
Physical Resource Block (PRB) concept: LTE divides spectrum into 180 kHz blocks. Understanding “1 PRB from 50 PRBs” requires knowing LTE carrier structure (10 MHz = 50 × 180 kHz). This is NOT NB-IoT-specific - it’s LTE radio access network architecture.
Deployment Mode Trade-offs:
In-band (Mode 1): 1 PRB from LTE carrier (e.g., PRB 25/50) → 2% capacity loss. CRITICAL DEPENDENCY: Requires active LTE carrier. If operator has no LTE in 800 MHz, in-band deployment impossible in that band.
Guard-band (Mode 2): Uses 200 kHz guard band between LTE carriers. PREREQUISITE: LTE configuration must have >200 kHz guard band. Not all LTE deployments do (10 MHz carrier with 9.0 MHz resource block allocation leaves only 2 × 500 kHz guard bands, suitable for guard-band deployment).
Standalone (Mode 3): Dedicated 200 kHz channel, often repurposed GSM carrier. DEPENDENCY: Requires spectrum refarming (migrating 2G services to 3G/4G first). This is a multi-year regulatory and technical process.
Economic Relationships:
Scale effect: Worked example shows hybrid deployment (in-band + standalone) has LOWEST 5-year TCO (€434K) for 200K devices because it combines fast urban rollout (in-band software upgrade) with rural coverage (standalone on refarm spectrum). Pure standalone costs more (€930K) due to refarm delay; pure in-band costs more (€1,101K) due to new rural LTE build-out.
Sunk cost fallacy: Operators with existing LTE naturally prefer in-band (leverage sunk infrastructure). New entrants without LTE may choose standalone (avoid LTE dependency). This explains market patterns: Vodafone (LTE-dominant) uses in-band; China Telecom (CDMA sunset) uses standalone.
Coverage Architecture:
NB-IoT Architecture shows how deployment mode affects eNodeB configuration. In-band uses SOFTWARE configuration of existing eNodeB (add PRB scheduling rule). Standalone may require HARDWARE upgrade if old GSM BTS lacks NB-IoT support.
MCL (Maximum Coupling Loss): All three modes achieve same 164 dB MCL, BUT standalone often uses lower frequency (e.g., 900 MHz Band 8 vs 1800 MHz Band 3), providing approximately +6 dB free-space path loss advantage (20×log10(1800/900) = 6 dB), translating to roughly 20–25% larger cell radius independent of mode choice.
Practical Constraints:
Cellular IoT Fundamentals explains regulatory spectrum licenses. Deployment mode selection MUST respect license terms (e.g., 800 MHz Band 20 license may prohibit standalone use, requiring in-band only).
Real-world case (Vodafone): 3-phase rollout demonstrates migration strategy - start in-band (fast revenue), add standalone for coverage gaps (phased investment). Hybrid deployment isn’t just technical - it’s a staged business model.
Key Insight: Deployment mode is NOT a one-time decision. The worked example (European water utility) shows EVOLUTION: initial in-band (months 0-2, urban), then standalone (months 6-12, rural). Static analysis (“which mode is best?”) misses the dynamic optimization (“which sequence of modes minimizes time-to-revenue while controlling cost?”).
30.15 See Also
Technical Deep Dives:
NB-IoT Power and Channel - Deployment mode doesn’t affect device power (180 kHz operation identical across modes), but frequency choice does (900 MHz standalone vs 1800 MHz provides approximately +6 dB free-space path loss advantage, allowing lower TX power for equivalent range).
LTE Network Architecture - Understanding eNodeB, PRB scheduling, carrier aggregation essential for in-band deployment planning. NB-IoT in-band is a LTE resource allocation problem.
Comparative Analysis:
LoRaWAN Deployment Planning - LoRaWAN requires greenfield gateway deployment (12-18 months, high capex). NB-IoT in-band leverages existing LTE (2-3 months, software-only). Deployment time often matters more than protocol features.
Sigfox Network Coverage - Sigfox operator-deployed (no mode choices). NB-IoT’s three modes provide flexibility but add complexity. Trade-off: simplicity vs control.
Spectrum Management:
Cellular Spectrum Bands - Global Band 8 (900 MHz), Band 20 (800 MHz) allocation maps determine standalone viability. Europe has harmonized 800/900 MHz for LTE/NB-IoT; Asia-Pacific varies (Band 8 in Australia, Band 5 in US).
TV White Space - Alternative unlicensed spectrum approach. Weightless-W uses cognitive radio for TVWS; NB-IoT uses licensed cellular spectrum. Deployment mode chapter shows licensed spectrum still requires strategic allocation.
Case Studies:
Cellular IoT Applications - Smart metering (Vodafone Denmark) uses in-band for 95% coverage, accepts 5% gap rather than deploy standalone. Agricultural sensors (China Telecom) use standalone 900 MHz for deep rural. Application requirements drive mode selection.
NB-IoT vs LTE-M Comparison - LTE-M also has three deployment modes with SAME trade-offs. Deployment mode selection is orthogonal to NB-IoT vs LTE-M technology choice.
Hands-On Tools:
Try the NB-IoT Deployment Mode Simulator (Simulations Hub) to calculate coverage, capacity, and cost for different mode combinations across custom service areas.
Use Spectrum Refarming Timeline Tool to visualize GSM sunset schedule and standalone deployment windows in different countries.
Advanced Topics:
5G NB-IoT Migration - NB-IoT modes persist in 5G (now called “NB-IoT in 5G NR”). In-band becomes “NB-IoT within 5G carrier.” Standalone becomes “NB-IoT standalone using 5G spectrum.”
Network Slicing - 5G network slicing creates virtual in-band deployment (dedicated slice within shared spectrum). Conceptually similar to in-band PRB allocation but at higher abstraction.
