mirror of
https://github.com/OPCFoundation/UA-.NETStandard.git
synced 2026-10-18 14:43:08 +08:00
romanett
28f1401324
Profiling the reference server under load showed most request-path CPU spent spinning on contended locks, plus several avoidable allocations per request or per node. This PR removes them, switches the reference server to Server GC, and adds the load harness used to measure it. ## Results Paired A/B against `master` (3 alternating rounds, medians, 20 sessions, workstation GC on both sides, 24-core machine): | Scenario | Throughput | p50 | p90 | Server CPU / op | Server alloc / op | Lock contention | |---|---|---|---|---|---|---| | Read, 1 value | **+45%** (94k → 137k req/s) | −20% | −46% | −37% | −39% | −94% | | Read, 100 values | **+40%** | −26% | −17% | −24% | −25% | +130% ¹ | | Read, 100 values, SignAndEncrypt | **+48%** | −36% | −20% | −30% | −26% | +106% ¹ | | Browse (Server object) | **+33%** | −31% | −12% | −13% | −14% | −40% | | Write, 100 values | +13% | −12% | −4% | −10% | −19% | −23% | | Data-change subscriptions | ±0 ² | ≈ | ≈ | −14% | −17% | −29% | ¹ The faster server spends more time in the remaining per-node locks. ² Throughput is bounded by the writer rate in this scenario. Separately, Server GC instead of workstation GC doubles Read-100 throughput (21k → 44k req/s) and cuts p90 from 2.1 ms to 0.6 ms, because workstation GC paused the server for about half the wall time under load. The working set grows from ~210 MB to ~430 MB. ## Changes - **`ObjectPool<T>`**: lock-free slot array. `ConcurrentBag.Count` took every per-thread lock on each `Return` of the pooled response value-task source, once per request. - **`FairRequestQueue`**: - drops redundant gate entries: the fail-fast pre-checks and the stopped check read volatile fields and are re-checked under the gate; the owner count is folded into selection; - coalesces wake pulses with a pending flag, so the bounded signal channel's lock is only taken when no token is outstanding. Wake-count semantics in the tests are unchanged. - **`DefaultServerResourceIsolationProvider`**: - compares peer addresses without `GetAddressBytes()` allocations; - builds the unsecured verification context only when a classifier is configured. - **TCP listener / channel**: caches the certificate DER copies per channel, keyed on the certificate instance. `X509Certificate2.RawData` returns a fresh copy on every access, and the listener read it for every request. - **`EndpointBase`**: resolves the request `ActivitySource` once instead of `Assembly.GetCallingAssembly()` per request. - **Node managers**: - one `ServerSystemContext` copy per request (cached on the `OperationContext`) for per-node callbacks: permission metadata, node metadata, view membership, method lookup; - removes a discarded copy in `GetReferenceDescriptionAsync`; - per-request operation caches are plain `Dictionary` instead of a `ConcurrentDictionary` with 24 lock objects; - the validation cache is built in one pass; - fixes zero-capacity error lists. - **`ResourceManager`**: translation lookups read an immutable snapshot that is rebuilt after writes. Browse serialised on the global lock for every display name. - **`MonitoredItem`**: shares the default `DataChangeFilter` instead of allocating one per sampled value. - **Reference server**: `ServerGarbageCollection=true`; new "Garbage collector mode" section in `docs/ServerScalability.md`. - **`perf/`**: `ServerLoadHarness` (out-of-process server plus load driver, server-side CPU/GC/alloc/contention deltas, optional `dotnet-trace` capture), `TraceAnalyzer` (TraceEvent-based CPU/allocation/contention aggregation), and paired A/B scripts. Not part of `UA.slnx`. ### Behaviour notes for custom node managers - The `IDictionary<NodeId, NodeState>` operation cache handed to overrides is now a `Dictionary`. It was a `ConcurrentDictionary`, but only one request flow uses it. - Within one request, per-node callbacks of a node manager share one `ServerSystemContext` copy. ### Later commits on this PR - **Permission checks without `NodeMetadata`** (`9068684c3`): the role-permission and access-restriction validators run on an internal `PermissionMetadata` struct; `AsyncCustomNodeManager` fills it on the stack from the same attributes and namespace defaults. Browse's reference filter reads 3 attributes per target instead of 13. Read 100 values: 120.6 → 95.7 KB/request, ~21k → 26k req/s; Browse 67.6 → 51.8 KB, 26.7k → 29.8k req/s. - **Read-only channel evidence** (`936b28d76`, `dc638ab9f`): `SecureChannelContext.ClientChannelCertificate`, `ServerChannelCertificate` and `ChannelThumbprint` are `ByteString`, so the cached per-channel certificate buffer is shared without a copy and cannot be altered by handlers or classifiers. **Public API change**, see the new section in `docs/MigrationGuide.md`. The isolation provider's revalidation snapshot still copies. - **Session lookup** (`936b28d76`): `SessionManager.TryGetSessionContext` takes the candidate snapshot without the bindings lock; the validating check under the lock is unchanged (2 instead of 4 global lock acquisitions per request). - Unit tests for `ObjectPool`, the per-request system context cache, the translation snapshot, `SecureChannelContext` null semantics and a stopped `FairRequestQueue`; `perf/` follows the repository conventions. ## Testing - `Opc.Ua.Server.Tests` (net10.0): 8386 passed before the `ByteString` change. After it, 13 secure-session and isolation tests failed; both causes are fixed in `dc638ab9f` and those tests (25 with their variants) pass. A full rerun of the Server, Sessions, Core, Redundancy, LDS and GDS suites on the final commit is in progress. - Full solution builds for net10.0 and net48; `dotnet format whitespace` clean. ### o6-automation opcua-benchmarks (sampling, deep pipelines) Replicated both tests of the [o6-automation benchmark](https://github.com/o6-automation/opcua-benchmarks) in the harness (`--scenario counters`, and `read` with 10 clients × N outstanding), paired against `origin/master`: | Test | master | this PR | |---|---|---| | Subscription sampling, 100 Hz, every item ≥ 99 % of samples, 1 item | FAIL (75–94 %) | PASS (100 %) | | same, 1024 items | FAIL (64–73 %) | PASS (99.9 %) | | same, capacity | — | between 4096 and 6144 items (~410–520k samples/s) | | Read, 10 clients × 100 outstanding, None | stalls (0–1.7k req/s) | 121–131k req/s | | Read, 10 clients × 100 outstanding, SignAndEncrypt | stalls (0 req/s) | 89–91k req/s | | Read, 10 clients × 1000 outstanding | channels faulted (`BadTcpNotEnoughResources`) | 0 errors, ~45–54k req/s | - **Fixed-rate sampling** (`30b769d13`): the sampling loop waited one interval after each sample, adding timer overshoot and sampling time to every cycle; samples now have absolute deadlines. - **Async response send** (`cdac5c5ca`): `SendResponse` blocked thread-pool workers on the channel gate; with many requests in flight the receive loop that was handed the gate waited for a pool thread and the channel stalled. The TCP, Kestrel TCP and WebSocket listeners now use `SendResponseAsync`. - **Backpressure instead of channel fault** (`191cd2ba7`): server channels stop reading while more than 100 responses are queued and resume at 50, instead of faulting the channel. Client channels keep reading. At 10,000 outstanding requests throughput is ~50k req/s with or without the limit engaged, so the remaining gap to the 100-per-client case is not the backpressure. ## Not changed - `SentMessageQueue` deep-clones every `NotificationMessage`. Retained messages return their pooled payloads on acknowledgement while the response copy may still be queued for send, and a client can acknowledge a sequence number before receiving it; the clone keeps a recycled payload from reaching another session's response. Removing it needs a send-completion hook from the transport. - Generated responses allocate a default `ResponseHeader` that is replaced (~2-3 % of per-request allocation). Fixing it means making structure fields of every generated type lazily initialized, a generator-wide change to default values, equality and encoding. 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
OPC Foundation UA .NET Standard Reference Server
Additional samples
- OPC 10000-21 onboarding registrar/client — loads the generated Onboarding companion model, binds the standard registrar administration methods to an injected ticket store, and drives a secure register/unregister round trip with a PowerShell script.
Introduction
This document is referring to the CTT version 1.04.9.398.
This OPC Server is designed to be the default OPC UA Server when opening the OPC UA Compliance Test Tool and it uses an address-space that matches the design of the UACTT and the requirements for OPC UA compliance testing.
It uses the OPC Foundation UA .NET Standard Library. Therefore it supports both the opc.tcp and https transports. There is a .NET framework based server with UI and a .NET Core console version of the server which runs on any OS supporting .NET 8.0 and later.
How to build and run the Windows OPC UA Reference Server with UACTT
- Open the solution UA.slnx with Visual Studio 2026.
- Choose the project
Reference Serverin the Solution Explorer and set it with a right click asStartup Project. - Hit
F5to build and execute the sample.
How to build and run the console OPC UA Reference Server on Windows, Linux and iOS
This section describes how to run the ConsoleReferenceServer.
Please follow instructions in this article to setup the dotnet command line environment for your platform.
Start the server
- Open a command prompt.
- Navigate to the folder samples/Reference/ConsoleReferenceServer.
- To run the server sample type
dotnet run --project ConsoleReferenceServer.csproj. The server is now running and waiting for the connection of the UACTT.
UACTT test certificates
The reference server always rejects new client certificates and requires that the UACTT certificates are in appropriate folders.
- The console server certificates are stored in %LocalApplicationData%/OPC Foundation/pki.
- The Windows .NET framework server stores the certificates in %CommonApplicationData%\OPC Foundation\pki.
- %CommonApplicationData% maps to the path set by the environment variable ProgramData on Windows.
- %LocalApplicationData% maps to a hidden location in the user home folder and depends on the target platform. See Certificates for more information.
Certificate stores
Under pki, the following stores contain certificates under certs, CRLs under crl or private keys under private.
- own contains the reference server public certificate and private key.
- rejected contains the rejected client certificates. To trust a client certificate, copy the rejected certificate to the trusted/certs folder.
- trusted contains trusted client and CAs certificates and CRLs.
- issuer contains CAs certificates and CRLs needed for validation.
- trustedUser contains trusted user and user CA certificates and CRLs.
- issuerUser contains user CA certificates and CRLs needed for validation of the certificate chains.
Placing the UACTT certificates
CTT creates a PKI folder in a new project. There is a subfolder copyToServer which contains the application and user identity certificates. Copy the ApplicationInstance_PKI certificates for testing with the UACTT to the following stores:
- trusted/certs: trusted certificates
- trusted/crl: revocation lists of trusted certificates
- issuer/certs: issuer certificates
- issuer/crl: revocation lists of issuer certificates
Copy the X509UserIdentity_PKI certificates for testing with the UACTT to the following stores:
- trustedUser/certs: trusted user certificates
- trustedUser/crl: revocation lists of trusted user certificates
- issuerUser/certs: user issuer certificates
- issuerUser/crl: revocation lists of user issuer certificates
UACTT Testing
Download and install the OPC UA Compliance Test Tool.
Note: Access to the UACTT is granted to OPC Foundation Corporate Members.
UACTT sample configuration
A sample configuration for the UACTT Version 1.04.9.398 can be found in UAReferenceServer.ctt.xml. Originally the reference server was certified only against the Standard UA Server profile, the Method Server Facet and the DataAccess Server Facet, in addition to all security related profiles. During release testing all Conformance Units should be enabled / checked.
Enabling Alarms & Conditions
The address space for Alarms contains a Start method which needs to be called to produce alarms for CTT testing. The method is called from a connected client.
The ConsoleReferenceServer has also a command line option -ctt which starts the Alarms when the server is started. The following Alarm instances are available for monitoring:
AlarmConditionTypeDiscrepancyAlarmTypeLimitAlarmTypeExclusiveLimitAlarmTypeExclusiveDeviationAlarmTypeExclusiveRateOfChangeAlarmTypeNonExclusiveLimitAlarmTypeNonExclusiveDeviationAlarmTypeNonExclusiveRateOfChangeAlarmTypeDiscreteAlarmTypeSystemOffNormalAlarmTypeExclusiveLevelAlarmTypeNonExclusiveLevelAlarmTypeOffNormalAlarmType
Known missing Tests and Issues
Depending on the setup in the provided project configuration, unsupported profiles should be skipped. At the time of writing (1.4.368) the following tests, profiles or facets are not yet covered or not configured due to issues:
- no support for events, historian and only a few basic Alarms & Conditions.
- Decimal type causes crash of CTT (needs investigation)
- no Image / Enumeration scalar test cases
- Arrays may not include Variant (needs investigation)
- no ArrayItemType implemented
- multiple hierarchical references test node missing
- deprecated security profiles support for SHA1/1k RSA keys implemented not according to certification. SHA1/1k keys can only be turned on/off for all profiles, not just for the deprecated profiles.
- Monitored Items with array are failing a few test cases.
It is recommended to run the server as retail build with disabled logging, to avoid side effects due to timing artifacts when log entries are written to a disk drive.
Running the CTT against a Docker container
Please refer to the Docker container of the Reference Server documentation page. Once setup to map the certificate store and the configuration to a local folder, follow the instructions above to copy the certificates to the PKI and get started.
Finding the Address Space Configuration Code
- Project: Reference Server
- File: ReferenceNodeManager.cs
- Method: CreateAddressSpace
Finding the UA Services
- Project: Opc.Ua.Server
- File: StandardServer.cs