Tiza Group

Home / Guides / Protection, control and communications

SCL Communication Section: Subnets and Addresses

What the SCL Communication section holds: subnets, connected access points, and the addresses of the server, GOOSE and sampled values.

By Shadai García, CEO and protection engineer at Tiza Group.

W1 · station (8-MMS)W2 · processPROTcontrol (router)protectionMU · fielddoes not crossConnectedAP · addressesServer (client-server)IP and transport selectorsGOOSE · multicast on W1MAC 01-0C-CD-01-00-01 · APPID · VLAN-PRIORITYSamples (SMV) · W2MAC 01-0C-CD-04-00-01 · APPID · VLAN-PRIORITYIED with router functionGOOSE, SV and time do not cross subnetsSectionSubNetworkConnectedAPAddressGSESMV© tizagrp.comW1 · station (8-MMS)W2 · processPROTcontrol (router)protectionMU · fielddoes not crossConnectedAP · addressesServer (client-server)IP and transport selectorsGOOSE · multicast on W1MAC 01-0C-CD-01-00-01 · APPID · VLAN-PRIORITYSamples (SMV) · W2MAC 01-0C-CD-04-00-01 · APPID · VLAN-PRIORITYIED with router functionGOOSE, SV and time do not cross subnetsSectionSubNetworkConnectedAPAddressGSESMV© tizagrp.com
SCL Communication section: subnets and addresses. Location in the system, operating principle and IEC 61850. Illustrative: actual values come from the nameplate and the manufacturer's manual.
Short answer

The SCL Communication section states which access points of which IEDs are connected to each subnet and with what addresses: the IP and selectors of the MMS server, and the multicast MAC, APPID and VLAN of every GOOSE or sampled values control block. It does not define the network plan; it records the engineering result so that every tool can configure its own device.

What the standard says and what it is for

Clauses 6.4 and 9.4 of IEC 61850-6 (edition 1, 2004) define the Communication section. The IED section already says which logical devices and nodes are reached through each access point; the Communication section completes the picture by saying which access points of which IEDs are connected to the same subnet and with which addresses. It is the place in the file where the network addresses live, not the device capabilities. How SCL describes the yard (voltage level, bay and equipment) is explained in SCL Substation Section.

In a project, this is the section consulted by whoever configures switches, VLANs or a traffic analyzer, and it must match what each IED actually publishes. The overview of the model is in the SCL object model, and the address plan of a real network is developed in IP and VLAN addressing plan.

Subnet: a logical concept

At this level the standard defines a subnet as a connection node for direct (data link layer) communication between access points. All the points on a subnet can talk to each other with the same protocol, without going through a router. There may be frame filtering at bridge level, but no network-layer routing.

Each subnet has:

Attribute or elementContent
nameUnique name within the file
desc, TextDescription
typeProtocol type, defined by the stack mappings. Those of part 8 start with “8-” and those of part 9 with “9-”. The one for IEC 61850-8-1 is “8-MMS”, and 9-2 uses the same
BitRateSpeed in Mbit/s (optional)
ConnectedAPList of connected access points

Because the subnet is logical, several of them with different upper-layer protocols can use the same physical bus. The physical structure can be inferred from good subnet and point naming, but its detailed description is outside the SCL core.

Connected access point (ConnectedAP)

Each ConnectedAP identifies the IED (iedName) and its access point (apName) and may optionally carry this information:

  • The server address (for client-server services).
  • One address for each GOOSE control block of the device.
  • One address for each sampled values control block.
  • The description of the physical connection (PhysConn).

If all three addresses are missing, the element only describes the connection topology, which is useful for network performance studies. A complete system file (SCD) must contain the server address or at least that of one control block. The address of access points with a server must be unique.

How an address is written

The address is a list of P parameters, each with a type that fixes its meaning. The meaning depends on the subnet protocol and is defined by the stack mappings. For part 8-1 (clause 25), these are the types:

ServiceAddress parameters
Client-server serverIP, subnet mask, gateway; OSI transport, session and presentation selectors (mandatory); other optional OSI parameters
GOOSEDestination MAC (mandatory), APPID (optional), VLAN priority (only if a VLAN is used) and VLAN-ID (optional)
Sampled valuesThe same as GOOSE (part 9-2, clause 10)

Format details imposed by 8-1: the MAC is written as six groups of two hexadecimal characters separated by hyphens; the APPID has four hexadecimal characters; the VLAN priority is a single character from 0 to 7; and the VLAN-ID has three hexadecimal characters. OSI selectors must have an even number of hexadecimal characters, with length limits.

A value can be declared with a restricted type so that an XML validator checks its format (for example, that an IP has the form of four numbers). Without that restriction, a badly written value would still be syntactically valid.

GOOSE and SV multicast addresses

Part 8-1 includes, in an informative annex, a recommendation for choosing destination addresses: the first three octets are 01-0C-CD, the fourth is 01 for GOOSE, 02 for GSSE and 04 for sampled values, and the last two are assigned within a recommended range. The reason is practical: it is convenient for the network card hardware to filter multicast traffic, and filtering algorithms vary between chips, so the integrator must evaluate their effect when assigning addresses. An example address from the standard: 01-0C-CD-01-00-01 for a GOOSE, with its APPID and VLAN priority.

GSE and SMV blocks

The GSE and SMV elements inside the ConnectedAP identify the control block by logical device (ldInst) and block name (cbName). They do not need to name the node because these blocks only exist in LLN0. Each one contains its address.

The GSE element also accepts two times:

  • MinTime: the maximum delay allowed for sending after a data change.
  • MaxTime: the source supervision time, the “heartbeat” cycle; within that time, a client must detect that the source has failed.

The standard clarifies that these times may influence stack mapping parameters (which ones and how is defined by each mapping). Their relation to GOOSE behavior is covered in GOOSE in IEC 61850.

Physical connection

The PhysConn element describes how the access point is connected: connection type (10BaseT electrical, optical or radio) and connector type (RJ45 or ST, for example). The specific values are defined by the stack mapping and may be extended for documentation purposes.

How it is implemented, in general terms

The system engineer connects each IED to its subnets: the station one, and the process bus ones if they exist. Then consistent addresses are assigned: unique IPs for the servers, a distinct multicast MAC and APPID for every control block, and VLANs and priorities in line with the network design.

What each party consults in the Communication section
  1. Network engineer. Subnets, VLANs, priorities and multicast MACs to configure the switches
  2. IED tool. The address of its own access point and of the blocks it publishes
  3. GOOSE subscriber. MAC, APPID and VLAN of the block it subscribes to
  4. Traffic analyst. Compares what should be on the wire with what was captured on the network

Typical errors are found by comparing: duplicate IPs, MACs or APPIDs when merging files from several projects, a VLAN that does not match the switches, or a change on a device that never made it back to the system file. You can review an SCD with our IEC 61850 SCL viewer, which shows IEDs, GOOSE and SV addresses, datasets and subscribers without uploading the file.

We can help with your files and your network

If you need to check that the SCD addresses match what actually runs on the network, or build the address plan for a new substation, at Tiza Group we review it through our remote relay configuration. For network design and full integration, see our IEC 61850 engineering and testing and commissioning.

References

  • IEC 61850-6, edition 1 (2004): clauses 6.4, 9.4 and the Communication section example.
  • IEC 61850-8-1, edition 1 (2004): clause 25 (address types) and informative annex B (multicast addresses).
  • IEC 61850-9-2, edition 1 (2004): clause 10 (sampled values addresses).

Frequently asked questions

What is the difference between the access point in the IED section and the ConnectedAP in the Communication section?

The access point in the IED section declares that the device has that access and what it contains. The ConnectedAP, in the Communication section, says which subnet it is connected to in this project and with which addresses. The first is a device capability; the second is a system engineering decision.

What addresses does a GOOSE control block carry in the file?

In edition 1, the destination multicast MAC address, the application identifier (APPID) and, if a VLAN is used, its priority and identifier. It may optionally carry the minimum and maximum retransmission times. The IP address plays no part because GOOSE does not use the network layer.

Are two subnets in the file two different physical networks?

Not necessarily. In SCL a subnet is a logical connection with a given protocol; several subnets with different protocols can share the same cable, and the physical structure is not part of the base description.

WhatsApp