Slice Admission Control is a critical decision point in the 5G core that validates whether a new user session can be admitted to a Network Slice Instance without violating the existing Slice SLA guarantees. It evaluates the slice's current utilization of Guaranteed Bit Rate (GBR) and non-GBR resources, the maximum number of allowed PDU sessions, and the specific NSSAI parameters requested by the user equipment against the slice's provisioned capacity.
Glossary
Slice Admission Control

What is Slice Admission Control?
Slice Admission Control (SAC) is the network function responsible for accepting or rejecting requests to establish new Protocol Data Unit (PDU) sessions within a specific network slice, based on real-time resource availability and configured policies.
The SAC function prevents resource overbooking and ensures strict Slice Isolation by rejecting requests that would degrade performance for active sessions. It works in concert with the Network Slice Selection Function (NSSF) and the Slice Orchestrator, enforcing admission policies that consider both static capacity limits and dynamic conditions like current load, priority levels, and energy-efficiency targets defined in the Slice-Level Energy Model.
Key Characteristics of Slice Admission Control
Slice Admission Control (SAC) is the gatekeeping function that determines whether a new PDU session can be established within a network slice. It enforces resource boundaries, SLA guarantees, and isolation policies to prevent slice congestion before it occurs.
Resource Quota Enforcement
SAC validates every incoming session request against the slice's maximum resource quota—a hard ceiling on compute, storage, and radio resources. If accepting the session would exceed the quota, the request is rejected immediately. This prevents the noisy neighbor problem, where one slice's traffic surge starves others sharing the same physical infrastructure. Quotas are typically defined in terms of Physical Resource Blocks (PRBs) in the RAN, virtual CPU cores in the core, and throughput guarantees in the transport network.
SLA-Aware Decisioning
Admission decisions are not binary accept/reject—they are SLA-aware. The controller evaluates whether the new session can be accommodated without violating existing Guaranteed Bit Rate (GBR) commitments or latency bounds for in-progress sessions. For example, a URLLC slice may reject a new session even if resources are nominally available, because the additional scheduling overhead would risk breaching the 1ms latency SLA for existing industrial automation traffic. This requires real-time telemetry from the Network Data Analytics Function (NWDAF).
Priority-Based Preemption
When resources are scarce, SAC implements priority-based preemption. Sessions are assigned an Allocation and Retention Priority (ARP) value. A high-priority emergency services session can trigger the preemption of a lower-priority eMBB streaming session to free resources. The preempted session is not dropped silently—it receives a cause code indicating resource unavailability, allowing the UE to reattempt on a different slice or after a backoff period. This ensures critical services maintain continuity during congestion events.
Energy-Aware Admission
In energy-efficient slicing architectures, SAC incorporates power consumption models into its decision logic. Rather than simply checking resource availability, the controller evaluates the marginal energy cost of accepting a session. It may steer the request to a slice instance hosted on hardware operating at a more efficient point on its Dynamic Voltage and Frequency Scaling (DVFS) curve, or defer admission to consolidate load and enable Cell Discontinuous Transmission (Cell DTX) on underutilized carriers. This aligns session admission with sustainability targets.
Cross-Domain Coordination
A slice spans RAN, transport, and core domains. SAC must therefore coordinate admission across all three simultaneously. A session accepted in the RAN but rejected in the core due to User Plane Function (UPF) overload creates a wasteful signaling storm. Modern SAC functions interface with the Slice Orchestrator to perform atomic admission checks—resources are reserved across all domains before a success response is sent to the UE. This transactional approach prevents stranded resources and ensures end-to-end slice integrity.
Predictive Admission Control
Reactive SAC—checking current load—is insufficient for slices with bursty traffic. Predictive admission control uses time-series forecasting models, often deployed within the NWDAF, to anticipate resource contention seconds or minutes in advance. If a surge in URLLC traffic is predicted, the SAC can proactively throttle eMBB admissions before congestion materializes. This closed-loop approach, combining prediction with preemptive action, is a cornerstone of Zero-Touch Network Provisioning and self-optimizing slice management.
Enabling Efficiency, Speed & Accuracy
Intelligent Analysis, Decision & Execution
We build AI systems for teams that need search across company data, workflow automation across tools, or AI features inside products and internal software.
Talk to Us
Search across company data
Give teams answers from docs, tickets, runbooks, and product data with sources and permissions.
Useful when people spend too long searching or get different answers from different systems.

Automate internal workflows
Use AI to route work, draft outputs, trigger actions, and keep approvals and logs in place.
Useful when repetitive work moves across multiple tools and teams.

Add AI to products and internal tools
Build assistants, guided actions, or decision support into the software your team or customers already use.
Useful when AI needs to be part of the product, not a separate tool.
Frequently Asked Questions
Explore the core mechanisms that govern whether a new session request is accepted or rejected within a 5G network slice, ensuring resource availability and strict adherence to Service Level Agreements.
Slice Admission Control (SAC) is a network function that accepts or rejects a request to establish a new Protocol Data Unit (PDU) session within a specific network slice instance based on current resource availability, configured slice policies, and active Service Level Agreement (SLA) guarantees. It operates as a real-time gating mechanism, preventing resource overbooking and ensuring that admitting a new session does not degrade the performance of existing sessions. The process typically involves the Slice Orchestrator or a dedicated admission control function querying the Network Data Analytics Function (NWDAF) for predictive load data, then comparing the requested slice profile (e.g., Guaranteed Bit Rate (GBR) or Ultra-Reliable Low-Latency Communication (URLLC)) against the slice's maximum quota. If the required resources exceed the available headroom, the request is rejected with a specific cause code, triggering the User Equipment (UE) to attempt a different slice or fallback mechanism.
Related Terms
Slice Admission Control is a critical gatekeeper that interacts with numerous other network functions and policies. Explore the key concepts that govern its decision-making process.
Network Slice Selection Assistance Information (NSSAI)
The NSSAI is the primary identifier used by User Equipment (UE) to request a specific slice. During the admission control process, the Slice Admission Control Function (SACF) validates the requested Single-NSSAI (S-NSSAI) against the UE's subscription and the slice's available resources. A mismatch or unauthorized request results in immediate rejection, enforcing strict tenant isolation from the very first signaling message.
Slice SLA
The Service Level Agreement defines the quantitative boundaries for admission control. It specifies the Guaranteed Bit Rate (GBR), maximum packet loss, and latency budgets. The admission control algorithm uses these thresholds to calculate the required resources for a new PDU session. If admitting the session would cause the slice to violate its SLA for existing users, the request is denied to maintain contractual performance guarantees.
Resource Overbooking
A capacity management strategy where the orchestrator allocates more virtual resources than physically available, relying on statistical multiplexing. Slice Admission Control must be tightly integrated with the overbooking policy to decide when to safely oversubscribe. A conservative admission policy rejects sessions early to prevent congestion, while an aggressive one leverages overbooking to maximize infrastructure utilization, risking SLA violations during peak loads.
Network Data Analytics Function (NWDAF)
The NWDAF provides predictive analytics that transform admission control from a reactive to a proactive function. By consuming NWDAF insights on predicted slice load and user mobility patterns, the admission controller can preemptively reserve resources or redirect requests. For example, if the NWDAF forecasts a surge in URLLC traffic, the admission control can temporarily restrict new eMBB sessions to safeguard mission-critical communications.
Slice Elasticity
The ability of a slice to dynamically scale its resources is a direct input to admission control logic. When a slice reaches its current capacity limit, the admission controller can trigger a scale-out request to the Slice Orchestrator instead of rejecting the session. The admission decision then becomes a function of the orchestrator's response time and the feasibility of acquiring additional virtualized resources without violating the Slice SLA of co-located instances.
Slice Isolation
A core design principle enforced by admission control. The mechanism must ensure that a new session in one slice does not degrade the performance of another slice sharing the same physical infrastructure. This requires the admission controller to have a global view of resource partitioning. It rejects requests not only when the target slice is full, but also when admitting the session would consume shared physical resource blocks needed to maintain the isolation guarantees of a neighboring high-priority slice.

About the author
Prasad Kumkar
CEO & MD, Inference Systems
Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.
His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.
Partnered with leading AI, data, and software stack.
How We Work
Custom AI workflows for your Business
One-fit-all AI don't work for modern businesses. At Inferensys, we aim to understand your business & custom requirements; which we use to define most efficient agentic workflows, the data, and the tools for your business.
01
Review the use case
We understand the task, the users, and where AI can actually help.
Read more02
Pick the right approach
We define what needs search, automation, or product integration.
Read more03
Build the first useful version
We implement the part that proves the value first.
Read more04
Improve from there
We add the checks and visibility needed to keep it useful.
Read moreThe first call is a practical review of your use case and the right next step.
Talk to Us