WO2018166593A1

WO2018166593A1 - Method for bandwidth reservation and suitable grid element

Info

Publication number: WO2018166593A1
Application number: PCT/EP2017/056132
Authority: WO
Inventors: Feng Chen; Franz-Josef GÖTZ; Marcel Kiessling; An Ninh NGUYEN; Jürgen Schmitt
Original assignee: Siemens Aktiengesellschaft
Priority date: 2017-03-15
Filing date: 2017-03-15
Publication date: 2018-09-20

Abstract

The invention relates to the modifying of the reservation scheme of the Multiple Stream Reservation Protocol, MSRP. In industrial networks, the flat reservation of required network elements according to the invention is advantageous because said network elements are specifically designed to quickly provide an alternative route having sufficient resources in case of error. Automatic access protection occurs by means of the verification of bandwidth by means of a protocol. The network can no longer be overloaded by real-time data.

Description

description

Method for bandwidth reservation and suitable network element

Industrial applications are based on the data transmission of sensors and actuators with a central controller (Programmable Logic Controller, PLC). In the past, bus systems were mostly used for data transmission. The BUS had a known maximum bandwidth of all

Participants was shared. As long as the total bus bandwidth is not exceeded, communication is stable.

The best-known example is Profibus (Process Field Bus), the universal fieldbus that is widely used in manufacturing, process and building automation. Profibus was developed by Siemens and the Profibus user organization and standardized in the international standard series IEC 61158. Profibus allows communication of Ge ^¬ boards from different manufacturers without any special interface adjustments.

Profibus specifies the technical features of a serial field bus system ^¬, up to cell level may be in crosslinked form of distributed digital automation ^¬ apparatus of a field level with each other. Profibus is a multi-master system that enables the joint operation of several automation, engineering or visualization systems with the decentralized peripheral devices on one bus. The further development of network technology led to the use of Ethernet-based networks for data transmission. PROFINET {Process Field Network) is the open Industrial Ethernet standard of the PROFIBUS user organization e. V. (PNO) for automation. Profinet uses TCP / IP and IT standards, is real-time Ethernet-capable and allows the in ^¬ tegration of fieldbus systems. In the past, to ensure the required availability and latency guarantees, proprietary Ethernet additions were often used. Thereby the required quality of data transmission and latency requirements were met.

In industrial networks, the availability of the network plays a decisive role. Many proprietary processes and pro ^¬ protocols (eg Media Redundancy Protocol MRP, Hybrid Routing Protocol HRP enhanced Rapid Spanning Tree Protocol eRSTP, eRSTP +) were introduced in order to achieve a high availability of the network. All these mechanisms are based on the targeted activation of a redundant connection in the event of an error.

As the Ethernet technology evolved, the Networking TSN workgroup complemented mechanisms to meet the needs of industrial applications.

Time-sensitive networking (TSN) is a set of standards that the Time-Sensitive Networking Task Group (IEEE 802.1) uses. The standardization standards define mechanisms for the transmission of data over Ethernet networks. A large part of the projects defines extensions of the bridging standard

IEEE 802. IQ. These extensions mainly address the transmission with very low transmission latency and high availability. Possible application areas are convergent

Networks with real-time audio / video streams and in particular real-time control streams, which, for. B. in the automotive or industrial plants are used for control. The establishment of the connections in the Ethernet network is carried out automatically by a reservation protocol (Multiple Stream Reservation Protocol, MSRP). All required resources must be reserved to guarantee the forwarding. Guarantees can be given for each individual transmission link through the check; possible overload scenarios are automatically prevented. The reservation is Runaway ^¬ leads based on the current transmission through MSRP. So far, a reservation is made for each individual connection. The Automatic Resource Reservation Protocol, MSRP, considers the resources along the path to be used (given by Audio / Video Bridging, AVB through RSTP). By establishing the stream (for the periodic transmission of the real-time data) along the path, the transmission on the existing path is ensured.

Switching over the network to a backup path leads, depending ^¬ but to clear the connection. The end application must reserve the connection again after the switchover has ended in the network. Until the completion of the new reservation creates a communication disorder for the participants which can lead to problems especially in real time applications.

The Multiple Stream Reservation Protocol, MSRP, only considers the active transmission path. This is advantageous for general networks, in order to reserve only in the affected devices, the resources really needed. All unaffected devices receive the description of the data in the event that a new subscriber wants to receive the data. Only after the registration of a recipient on the network resources are reserved on the path to this.

Industrial networks are specially designed to use a planned replacement route in case of failure. The resources must be designed so that enough resources would remain available in the event of a network switch. This is typically accomplished by requiring each network device to provide enough resources - but reservation and verification of individual connections does not occur in legacy networks.

It is therefore an object of the invention to provide a method and an apparatus which makes it possible in the described environment, to carry out a changeover of the connection without the compound to dissolve completely and rebuild by Reservie ^¬ tion. This object is achieved by a method according to the Me paint of claim 1.

This object is likewise achieved by a device according to the features of patent claim 8.

The method for bandwidth reservation in a network, consisting of network elements that operate according to the Time Sensitive Net ^¬ work standard according to IEEE 802.1, and in the connection establishment between a transmitting and a receiving network element, a first transmission standard by default checks ^¬ ge and the resources required for the network elements are reserved, wherein in the connection establishment phase at least a second, alternative data transmission ge ^¬ checks and the resources are reserved in the network elements.

With a bus system, checking the bus bandwidth is sufficient, there were no other participants on the bus.

Why should the reservation scheme of Multiple Stream Reservation Protocol for industrial networks, changed MSRP ^¬ to.

At the boundary between the end device and the network (edge port), each new stream is checked against the maximum resources available in the entire network. If there are too few resources, the feed-in into the network is prevented. The resource check takes place at the edge port. This makes it possible to switch over in the network at any time - the examination of existing resources is already checked by the edge port.

Layering enables mapping of local automation cells and connection to a backbone network. This allows networks with a different performance to be combined. At the transition of the networks an edge port for the superimposed / neighboring Seg ^¬ ment arises. Local streams only need resources in the automation cell. Streams that have been transferred to another cell need resources in the local cells and additionally in the backbone (base network) network.

The new layering concept enables independent switching in every level / instance. The reserved but currently unused resources can be used with TSN for non-real-time data, so they are not wasted. They may only be used for the transmission of real-time data in the event of an error.

The bandwidth is no longer considered for the active forwarding path in each bridge, but per cell / network segment. So far, a reservation is made only along the currently active forwarding path.

This area reservation is advantageous in industrial networks - because they are designed specifically to error ^¬ falling rapidly an alternate route with sufficient resources to bie ^¬ th By testing the bandwidth with a protocol, an automatic access control takes place.. The network can no longer be overloaded by real-time data.

The switching of the forwarding path (transmission path of the data) takes place independently of the resource consideration. As a result, no reservation is necessary when changing the path. As long as a possible alternative path exists in the network, it can be used. This allows the existing methods for fast switching in industrial networks such. B. MRP continue to be used. Use of the new Shift redundancy concepts it is ^¬ enables (eg mechanisms such as Intermediate System to Intermediate System IS-IS SPB, Fast Reroute, ..) because the Reservie ^¬ tion has already been tested for all existing alternative routes.

In the following, the invention is illustrated by figures. It shows

FIG. 1 shows a ring topology with MRP as the redundancy method in the normal state, FIG. 2 shows a ring topology with MRP as a redundancy method in the event of a fault with interruption,

Figure 3 shows a typical ring topology as is common in industrial networks,

FIG. 4 shows a ring topology with area reservation, and FIG. 5 shows an example with several network segments.

Figure 1 shows a conventional ring topology in a network 1, which consists of network elements 11 to 18, and a redundancy manager RM. As a routing method is the media

Redundancy Protocol MRP or Hybrid Routing Protocol HRP used. The small X below the redundancy manager indicates a targeted interrupt in the ring topology to prevent redundant transmission of data packets on two redundant data paths by blocking the output port towards the network element 15.

In Figure 2, the interruption "x" is then represented as an error case in the ring, thus there is only one possible way to each network element in the ring.

Figure 3 shows a typical network for a industriel ^¬ len installation, wherein these are characterized in that a control element 3 (the PLC, Programmable Logic Controller) via a bridge 2 to the network elements 31 to 38 by the sensor data and measured values for controlling and control data is transmitted away from the controller. This is done according to the known principles described above according to TSN / AVB standard with data packets in time windows. The resources, thus in particular the required Speicherbe ^¬ may be used in the input and output queues and computing capacity for the switching and routing in the network elements 31 to 38 are reserved in advance during connection establishment. Figure 4 now shows a different view of the known network, here the path does not become a sequence of individual

Network elements on the way from the sender to the receiver (quasi one-dimensional) but as all possible paths from the sender to the receiver within a network or segment. fen (flat or two-dimensional, 2) and performed as a "sum reservation" in the individual network elements, so it is reserved in each network element a reservation for potential transmissions / possible connections between the sender and receiver - within the maximum available bandwidth of the network element.

FIG. 5 shows a connection of three networks N1, N2, N3 via a backbone BB, the networks in each case being networks such as that illustrated in FIG. The three networks Nl, N2, N3 are ver linked in each case via ^¬ "Edge ports" 20, 30, 29, 49 to each other via the backbone.

A reservation invention within a network ^¬ works Nl to N3 then has an impact on the Reservie ^¬ stakes in the resources of the backbone.

The reservations will look like this

Networks N1, N2 N3 comprise the reserved bandwidth in their own network plus one part of each external bandwidth reservations. The reservations in the backbone umfas ^¬ sen all attached networks.

Of course, this approach has the disadvantage that reservations be made in the network elements, which in the best case (namely, when happens in the transfer no error and the data packets can be worn over the optimal path through ^¬) are required before. This means that the transmission capacity of the network through "blind" reservations is lower than in the previously known methods.

However, in some applications it is more important that data transfer can be done safely and on time, disconnecting and rebuilding incl. new Reser ^¬ vierungsverhandlungen in the network elements is too time consuming and would result in the worst case to disturbing effects in the running for an industrial plant process, which is not acceptable for some applications.

Claims

claims

1. A method for bandwidth reservation in a network (1), consisting of network elements (31, 32, ... 38) which operate according to the Time Sensitive Network Standard according to IEEE 802.1, wherein when establishing a connection between a transmitting and ei ^¬ nem receiving network element ( 31, 32, ... 38), a first transmission standard and tested in accordance with this required Res ^¬ resources are reserved in the network elements,

characterized in that

at least a second, alterna ^¬ tiver data transmission path checked in the connection setup phase, and the resources are reserved in the network elements.

2. Method according to claim 1,

characterized in that

the procedure of the reservation according to the MSRP - Message Sessi ^¬ on Relay Protocol is executed.

3. The method according to any one of the preceding claims, characterized in that

the network (1) has a ring structure.

4. The method according to any one of the preceding claims, characterized in that

the network consists of several sub-networks (Nl, N2, N3), which are connected to each other via a backbone (BB).

5. The method according to any one of the preceding claims, characterized in that

a resource check in all available network elements (31, ... 38) through the edge port between terminal (2, 20, 30, 29, 39) and network (1) or a sub-network (Nl, N2, N3 ) and, if the test result is negative, transmission of the data to the network is prevented.

6. The method according to any one of the preceding claims, characterized in that

When changing the transmission path between a transmitting and a receiving network element (31, 32, ... 38) after connection establishment no re-examination of the resources neces ^¬ sary.

7. The method according to any one of the preceding claims, characterized in that

the reserved resources can be used elsewhere in the network element, unless they are required for Datenübertra ^¬ supply in the network.

8. suitable apparatus and adapted to Bandbreitenre ^¬ Serving in a network (1), which (31, 32, ... 38) consisting of network elements that operate in accordance with the Time Sensitive Network Standard IEEE 802.1,

wherein when establishing a connection between a sending and ei ^¬ nem receiving network element (31, 32, ... 38) by default tested a first transmission path through the apparatus and it needed to be reserved resources in the network elements, and in the connection setup phase, at least a two ^¬ ter, tested alternative data transmission path, and the resources are reserved in the network elements ^¬ .

9. Device suitable and arranged for performing ei ^¬ nes of the method according to the features of one of the claims 2 to 7.