CATS Service Metric Agent (C-SMA) Functional Implementation

1.1. Scope and Relationship to Other CATS Documents

This document specifies the functional behavior of the C-SMA as a CATS component. As with the SCI document [I-D.zhangb-cats-sci-implementation], the scope is deliberately limited in two ways. First, the normative requirements concern the observable behavior of the C-SMA - the validation rules applied to incoming SCI reports, the soft-state management of the metric cache, and the synchronization semantics towards the C-PS. Second, how the C-SMA is internally structured is a local matter; Section 3 provides an illustrative component decomposition that is informative in nature.¶

The relationships to other CATS documents are as follows:¶

• [RFC10053] defines the CATS components and their interactions, including the C-SMA role of collecting service metrics and advertising them to C-PSes, and the centralized/distributed/hybrid deployment models (Appendix A of that document). This document provides the detailed C-SMA-side behavior that the framework leaves open.¶

• [I-D.ietf-cats-metric-definition-13] defines the CATS metric taxonomy, the CATS metric field template, and the "CATS Metrics" registry. The validation rules in Section 4.2 of this document reference the metric semantics and field definitions of that document; this document does not define new metric categories.¶

• [I-D.zhangb-cats-service-metrics-op] defines the service-oriented metrics (Global Available Slots, Computing Time, Price, Reputation, Security Level) and the C-PS side joint selection workflow. The metrics that the C-SMA collects, validates, and advertises are those defined in that document.¶

• [I-D.zhangb-cats-sci-implementation] specifies the SCI-side behavior: how an SCI derives and reports service-oriented metrics to the C-SMA. This document is the control-plane counterpart: it specifies how the C-SMA receives, validates, caches, and forwards those reports. The two documents share the same metric semantics, the same health state model, and the same update-control philosophy; the report fields defined in that document are the input to the validation and caching functions defined here.¶

• [I-D.ietf-cats-oam-fw] defines the OAM framework and requirements for CATS. The failure-detection, synchronization, and partitioned-operation behavior specified in Sections 9 and 10 of this document is a component-level behavior that complements, and is consistent with, the OAM requirements; the OAM mechanisms and management interfaces are out of scope here.¶

• [I-D.ietf-cats-data-model] defines a YANG data model for the management of CATS. The thresholds, policies, and configuration objects mentioned in Section 7 of this document should be mapped to that data model where applicable; this document does not define new YANG objects.¶

• [I-D.yxl-cats-protocols-applicability] discusses the applicability of existing protocols for CATS. This document does not define new protocol extensions or wire formats for metric distribution; it specifies the C-SMA-side behavior that is independent of the chosen transport, and leaves protocol selection to that document (see Section 6).¶

The framework states that how a service provider structures its services remains out of the scope of CATS. This document does not standardize the internal structure of a C-SMA implementation; it standardizes the behavior of the C-SMA that is observable by other CATS components, in particular the C-PS.¶

1.2. Requirements Language

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶

3.1. C-SMA Position in CATS Framework

The C-SMA sits between the service site (SCIs) and the CATS control plane (C-PS). Its relationships with other CATS components are:¶

• The C-SMA receives metric reports from one or more SCIs within the service site.¶

• The C-SMA validates, and caches these metrics.¶

• The C-SMA advertises metrics to the C-PS via the control-plane protocol.¶

• The C-SMA does not directly interact with clients, C-TCs, or CATS-Forwarders.¶

• The C-SMA may be co-located with an SCI, an Egress CATS-Forwarder, or deployed as a standalone component within the service site.¶

The C-SMA is the authoritative source of computing metrics for the service site. The C-PS relies on the C-SMA's reports to build the Computing Service Table.¶

4.1. Collection Interface

Push Mode (Recommended):

SCIs proactively push metric reports to the C-SMA at configurable intervals or upon significant changes. This reduces polling overhead and enables near-real-time updates. The SI MUST support at least one of the following push transports: gRPC streaming (bidirectional), HTTP/2 with Server-Sent Events (SSE), MQTT publish/subscribe, or Internal message bus (e.g., Kafka, RabbitMQ).¶

Pull Mode:

The C-SMA periodically queries SCIs for metrics. This is useful when SCIs cannot initiate connections (e.g., behind NAT or firewall). The SI MUST support HTTP GET or gRPC unary calls for pull mode.¶

Hybrid Mode:

A combination of push and pull. SCIs push routine updates, and the C-SMA pulls on-demand during recovery or anomaly detection.¶

The SI MUST support multiple concurrent SCI connections. Each connection is identified by the CSCI-ID and authenticated.¶

4.2. Metric Validation and Sanitization

The Validator & Sanitizer (VS) inspects every incoming SCI report before it enters the Metric Cache. The purpose is to prevent erroneous or malicious reports from influencing traffic steering decisions. The validation rules below are normative in the sense that a C-SMA MUST apply them; the exact thresholds are example values and MUST be configurable. The semantics and units of the validated metrics are those defined in [I-D.zhangb-cats-service-metrics-op] . The VS performs the following checks:¶

Range Validation:

Each metric value MUST fall within a valid range consistent with its field definition in [I-D.zhangb-cats-service-metrics-op]: GAS: 0 to MAX_UINT32; Computing Time: 0 to 3600000 ms (1 hour, as an example ceiling); Price: 0 to MAX_UINT32; Reputation: 0 to 10 (as an example scale); Security Level: 0 to 10 (as an example scale). Values outside these ranges are rejected. The valid ranges MUST be derived from the registered field definitions; the values above are examples.¶

Rate-of-Change Validation:

The VS detects anomalous changes: GAS change > 50% within 1 second without a corresponding health change; Computing Time change > 100% within 1 second; Health status flapping (> 3 transitions within 10 seconds). These thresholds are example configurations. Anomalous reports MAY be quarantined for manual review or accepted with a warning flag.¶

Sequence Validation:

The VS checks the sequence number or timestamp to detect duplicate, out-of-order, or replayed reports. Old reports are discarded.¶

Cross-Metric Consistency:

The VS checks consistency across metrics: if all SCIs are UNHEALTHY, Site_GAS MUST be 0; if Site_GAS > 0, at least one SCI MUST be HEALTHY; Computing Time SHOULD be > 0 if GAS > 0.¶

Sanitization:

The VS MAY apply smoothing filters (e.g., moving average) to reduce noise in frequently fluctuating metrics.¶

5.1. Local Metric Cache

Each entry includes: CSCI-ID and CS-ID; all reported metrics (GAS, Computing Time, optional metrics); Health status; Timestamp of last update; Sequence number of last report; TTL (time-to-live) for soft-state aging.¶

The MC provides O(1) lookup by CSCI-ID and supports range queries by CS-ID for aggregation.¶

5.2. Soft-State and Aging

The MC implements soft-state aging to handle SCI failures or network partitions. The behavior in this section addresses the CATS charter item on the appropriate frequency and scope of metric distribution: aging and withdrawal semantics determine how a C-PS can safely rely on previously advertised metrics, and what the C-SMA advertises when a source of metrics disappears.¶

TTL-Based Expiration:

Each cache entry has a TTL (e.g., 2x the expected reporting interval). If no update is received within the TTL, the entry is marked as STALE. The TTL for each metric class MUST be configurable; the example multiplier is informative.¶

STALE Handling:

When an entry becomes STALE, the C-SMA MAY send a pull request to the SCI. If no response is received within a grace period, the SCI is marked as UNHEALTHY and its GAS contribution is set to 0. The C-SMA sends an update to the C-PS reflecting the reduced capacity.¶

Hard-State vs. Soft-State:

The C-SMA maintains hard-state for static information (e.g., CS-ID to CSCI-ID mappings) and soft-state for dynamic metrics (e.g., GAS, Computing Time).¶

Cache Size Management:

The MC MUST implement a maximum cache size. When the limit is reached, the least-recently-used (LRU) entries are evicted. Eviction of an entry triggers a withdrawal advertisement to the C-PS.¶

5.3. Per-SCI State Tracking

The C-SMA maintains a state machine for each SCI. This state machine operates at the granularity of an SCI (a reporting endpoint at a service site) and tracks the reliability of the SCI as a source of metric reports. It is distinct from, and complementary to, the instance-level health state machine defined in [I-D.zhangb-cats-sci-implementation] (HEALTHY, DEGRADED, UNHEALTHY, UNKNOWN), which tracks the operational state of a service instance: an SCI in the ACTIVE state may report instances in any health state, and an SCI that is FAILED as a reporting source implies that the health of the instances behind it is unknown to the C-PS.¶

• ACTIVE: The SCI has been discovered but not yet confirmed healthy.¶

• HEALTHY: The SCI is reporting metrics normally.¶

• STALE: No report received within TTL. The C-SMA attempts recovery.¶

• FAILED: The SCI is considered unreachable. Its metrics are withdrawn.¶

State transitions trigger updates to the C-PS as appropriate.¶

6.1. Centralized Model: Northbound Interface

In the centralized model (e.g., SDN controller, see [RFC7149] and [RFC7426]), the C-SMA acts as an application that reports metrics to a centralized C-PS via a northbound interface. The PA adapts the metric cache to a northbound API whose concrete design is out of scope for this document; the following are common transport choices discussed in [I-D.yxl-cats-protocols-applicability]:¶

RESTful API (HTTP/HTTPS):

The C-SMA publishes metric reports to the C-PS. The specific resource paths and payload schemas are defined by the applicable northbound specification, not by this document. The PA MUST preserve the semantics of the operations: publish a new or updated metric set, update metrics for a specific CSCI-ID, withdraw metrics for a failed SCI, and retrieve cached metrics for reconciliation.¶

gRPC:

The C-SMA may establish a gRPC stream to the C-PS for bidirectional communication, enabling streaming metric updates from C-SMA to C-PS and streaming policy updates or queries from C-PS to C-SMA. The service definition is out of scope here.¶

NETCONF/YANG-Push:

The C-SMA publishes metrics as YANG-modeled data via NETCONF notifications [RFC6241] or YANG-Push [RFC8639], using the data model of [I-D.ietf-cats-data-model] where applicable. This is suitable for deployments that already use NETCONF for network management.¶

The PA MUST support transport security (e.g., TLS) for all northbound connections.¶

6.2. Distributed Model: Routing Protocol Extensions

In the distributed model, the C-SMA distributes metrics via routing protocol extensions, as analyzed in [I-D.yxl-cats-protocols-applicability]. Whether and how CATS metrics are carried in BGP-LS, GRASP, or IGP TE extensions is a protocol design question that belongs to the corresponding protocol documents. This document does not specify such extensions.¶

BGP-LS with CATS Extensions:

If BGP-LS [RFC8571] (an application of the multiprotocol extensions of BGP-4 [RFC4760]) is selected as the distribution protocol, the encoding and registration of CATS metric TLVs are to be defined in the protocol document identified by [I-D.yxl-cats-protocols-applicability] (see Section 6.4). The PA is responsible for mapping internal metric values to the TLV fields once such an encoding exists.¶

GRASP:

If GRASP [RFC8990] is selected, the definition of CATS metric objectives is likewise a protocol design matter for the corresponding document. The PA maps internal metrics to the objective payload once defined.¶

OSPF/IS-IS TE Extensions:

Using IGP TE extensions [RFC7471][RFC8570] to carry CATS metrics requires a scoping analysis (e.g., whether the CATS metric set is suitable for link-state distribution and how often it changes). This document does not take a position on the suitability; the question is left to [I-D.yxl-cats-protocols-applicability].¶

The PA MUST implement protocol-specific rate limiting and dampening to prevent routing instability.¶

6.3. Hybrid Model: Combined Approach

In the hybrid model, the C-SMA uses a combination of centralized and distributed mechanisms:¶

• Static/Quasi-Static Metrics via Distributed Protocols: Infrequently changing metrics (e.g., maximum capacity, security levels) may be distributed via BGP-LS or GRASP where such encoding exists.¶

• Dynamic/Status Metrics via Centralized Interface: Frequently changing metrics (e.g., GAS, Computing Time) are reported via the RESTful or gRPC northbound interface to avoid flooding the routing protocol.¶

The PA MUST ensure consistency between the two channels. If a conflict is detected (e.g., a distributed advertisement reports GAS=500 while the northbound interface reports GAS=400), the more recent update takes precedence.¶

6.4. Open Items for Protocol Documents

The following items are intentionally not resolved in this document and need to be addressed by the protocol documents identified in [I-D.yxl-cats-protocols-applicability]:¶

• The wire encoding of CATS metrics in BGP-LS, including TLV codepoints and their IANA registration.¶

• The GRASP objective name and payload schema for CATS metrics, if GRASP is selected.¶

• The applicability analysis for OSPF/IS-IS TE extensions.¶

• The YANG data model objects for the thresholds and policies referenced in Section 7 (see [I-D.ietf-cats-data-model]).¶

7.1. Local Policy Engine

Rule Format: IF <condition> THEN <action>¶

Examples (values are examples, not defaults): IF Site_GAS < 100 THEN SET Security_Level = 5; IF SCI_Health = DEGRADED THEN REDUCE SCI_GAS BY 50%; IF Time IN Maintenance_Window THEN SUPPRESS_UPDATES; IF Price > 100 THEN SET Price = 100.¶

The PE evaluates policies in order of priority before metrics are sent to the C-PS.¶

7.2. Metric Transformation Policies

The PE applies the following transformation policies:¶

Safety Margin:

Reduce reported GAS by a configured percentage to prevent over-subscription. Example (informative): Reported_GAS = Site_GAS * 0.8.¶

Ceiling/Floor:

Enforce upper and lower bounds on metrics. Example (informative): Computing Time MUST be >= 1 ms and <= 10000 ms.¶

Smoothing:

Apply exponential moving average (EMA) to reduce noise: Smoothed_GAS(t) = alpha * Raw_GAS(t) + (1 - alpha) * Smoothed_GAS(t-1), where alpha is a configurable smoothing factor (example default 0.3).¶

7.3. Update Control Policies

The PE controls when updates are sent to the C-PS:¶

Threshold Policies:

GAS delta > 10% or absolute delta > 50; Computing Time delta > 20%; Health status change; Security level change. Thresholds are example values and MUST be configurable.¶

Time-Based Policies:

Maximum update interval: force an update at least every 60 seconds (heartbeat); Minimum update interval: suppress updates more frequent than 5 seconds (rate limiting); Maintenance window: suppress all non-critical updates during scheduled maintenance. The intervals are example values and MUST be configurable.¶

Burst Control:

Token bucket algorithm to limit update rate; Backoff algorithm for repeated failures.¶

8.1. SCI Authentication

Mutual TLS (mTLS):

The SCI and C-SMA exchange X.509 certificates during connection establishment. The C-SMA validates the SCI's certificate against a trusted CA and checks that the certificate subject matches the expected CSCI-ID.¶

Token-Based:

The SCI includes a signed token (e.g., JWT) in each report. The C-SMA validates the token signature, expiration, and issuer.¶

IPsec:

For deployments where SCIs and C-SMA communicate over an untrusted network, IPsec MAY be used to protect the entire communication channel.¶

The C-SMA MUST maintain a certificate revocation list (CRL) or support OCSP to check for revoked certificates.¶

8.2. C-PS Authentication

The C-SMA MUST authenticate the C-PS before sending updates. Supported methods:¶

Mutual TLS (mTLS):

The C-SMA validates the C-PS certificate during northbound connection establishment.¶

Pre-Shared Key (PSK):

For lightweight deployments, the C-SMA and C-PS MAY use a pre-shared key for message authentication (e.g., HMAC-SHA256).¶

The C-SMA MUST verify that the C-PS is authorized to receive the metrics for the advertised CS-IDs. This MAY be enforced via certificate attributes or policy configuration.¶

8.3. Metric Integrity and Freshness

The C-SMA MUST protect the integrity of metric reports:¶

Digital Signature:

Each report sent to the C-PS is signed using the C-SMA's private key. The C-PS verifies the signature using the C-SMA's public key.¶

Message Authentication Code (MAC):

For deployments using PSK, a MAC (e.g., HMAC-SHA256) is appended to each report.¶

Timestamp and Nonce:

Each report includes a timestamp and a monotonically increasing nonce to prevent replay attacks. The C-PS MUST reject reports with timestamps outside a reasonable window (e.g., +/- 30 seconds) or duplicate nonces.¶

Freshness Indicator:

The C-SMA includes a freshness indicator (e.g., sequence number or epoch) in each report. The C-PS uses this to detect stale or out-of-order updates.¶

9.1. Trigger-Based Updates

The UC sends updates to the C-PS based on the following triggers. The behavior in this section, together with Section 9.2 and Section 9.3, defines the synchronization semantics that determine how fresh the C-PS's view of a service site is; the required update frequency and scope depend on the service class and are out of scope for this document (see the CATS charter and [I-D.zhangb-cats-service-metrics-op] for the corresponding discussion).¶

• Metric Change Trigger: When an aggregated metric crosses a configured threshold, the UC sends a delta update.¶

• Health Change Trigger: When an SCI changes health state (HEALTHY -> DEGRADED, DEGRADED -> UNHEALTHY, etc.), the UC sends an immediate update.¶

• SCI Discovery Trigger: When a new SCI is discovered or an existing SCI is withdrawn, the UC sends a full update.¶

• Policy Change Trigger: When a local policy is updated (e.g., new safety margin), the UC re-evaluates all metrics and sends updates as needed.¶

The UC MUST prioritize triggers. Critical triggers (e.g., health change) take precedence over routine triggers (e.g., metric change).¶

9.2. Periodic Synchronization

The UC sends periodic updates to maintain soft state:¶

Heartbeat:

A full metric snapshot is sent at a fixed interval (e.g., every 60 seconds) even if no triggers have fired. This ensures the C-PS does not age out the C-SMA's entries.¶

Refresh:

If the C-PS supports it, the C-SMA MAY use a lightweight refresh message (e.g., keepalive) to extend the C-PS's cache TTL without retransmitting all metrics.¶

Synchronization Interval:

The heartbeat interval SHOULD be configurable and MAY be adjusted dynamically based on network conditions or C-PS load.¶

9.3. Recovery and Reconciliation

After a C-PS restart or network partition, the C-SMA and C-PS MUST reconcile their state:¶

Full Sync:

The C-SMA sends a complete snapshot of all metrics to the C-PS upon reconnection.¶

Incremental Sync:

If the C-PS has persisted some state, the C-SMA MAY send only the changes since the last acknowledged sequence number.¶

Reconciliation Protocol:

The C-PS MAY request a full sync by sending a synchronization request to the C-SMA. The C-SMA responds with all current metrics.¶

Conflict Resolution:

If the C-PS and C-SMA have divergent state (e.g., due to message loss), the C-SMA's view takes precedence as the authoritative source of service metrics. This document proposes this precedence rule because the framework does not define a conflict-resolution procedure between the C-SMA and the C-PS; the rule is subject to confirmation by the working group.¶

10.1. SCI Failure Detection

Missing Reports:

If an SCI fails to send reports within the TTL, the C-SMA marks it as STALE and then FAILED.¶

Health Monitor Integration:

The C-SMA MAY integrate with an external health monitor (e.g., the SCI's own health checks or a site-wide monitoring system) to detect failures faster than TTL-based detection.¶

Active Probing:

The C-SMA MAY send active probes (e.g., ping, HTTP GET) to unresponsive SCIs to distinguish between SCI failure and network partition.¶

Upon detecting an SCI failure, the C-SMA:¶

1. Sets the SCI's GAS contribution to 0.¶

2. Updates the aggregated site metrics.¶

3. Sends an immediate update to the C-PS.¶

4. Logs the failure for operational analysis.¶

10.2. C-PS Unreachability

If the C-SMA cannot reach the C-PS:¶

Retry with Backoff:

The C-SMA retries with exponential backoff (e.g., 1s, 2s, 4s, 8s, ... up to a maximum of 60s).¶

Buffering:

The C-SMA buffers updates in a local queue. When the C-PS becomes reachable, the buffered updates are sent in batch. The buffer MUST have a maximum size; old updates are dropped when the buffer is full.¶

Fallback:

In distributed deployments, the C-SMA MAY fall back to an alternative C-PS or advertise metrics via a routing protocol if the primary C-PS is unreachable.¶

Local Action:

The C-SMA MAY notify local operators (e.g., via syslog, SNMP trap) of the C-PS unreachability.¶

10.3. Partitioned Operation

During a network partition between the service site and the C-PS:¶

Read-Only Mode:

The C-SMA continues to collect and cache metrics from SCIs but stops sending updates to the C-PS.¶

Local Steering:

If the C-SMA is co-located with an Egress CATS-Forwarder, it MAY provide local steering decisions using cached metrics until the partition heals.¶

Partition Healing:

Upon reconnection, the C-SMA performs a full sync to reconcile state.¶