JP2014186423A - Monitoring system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To eliminate processing delay in monitoring.SOLUTION: A monitoring system includes a plurality of sensors for detecting a predetermined state in a system, a controller for monitoring the state of the system on the basis of detection values detected by the sensors, and a sensor read circuit physically connected to each of the sensors. The sensor read circuit directly reads the detection values detected by the sensors from the sensors according to a sensor read instruction issued by the controller, and transmits the read detection values to the controller.

Description

  The present invention relates to a monitoring system, and more particularly to a monitoring system that monitors a state of an apparatus with a sensor. To monitor

  In order to manage the state of devices such as server systems, various types of abnormality monitoring such as temperature abnormality and voltage abnormality are performed. In such an abnormality monitoring, an abnormality determination is performed by using a sensor component and reading the sensor value, and cooperation such as cooling control by FAN is performed while accurately detecting the abnormal state. The abnormality monitoring control described above is performed by firmware control. For example, in Patent Document 1, vibration monitoring for a computer is performed.

JP 2012-48303 A

  When the monitoring target is a large-scale server system, it is necessary to perform sensor monitoring by mounting a large number of sensors, for example, several hundreds, in order to cope with environmental abnormalities accurately. However, such sensor monitoring requires reading out all the sensors within a fixed time, and speeding up the processing has been a problem.

  Here, generally, protocols such as I2C and SMBus interfaces are used for accessing a large number of sensors mounted on the server system. The sensor is accessed by firmware operating on a controller provided with a system management processor called BMC. Such a controller is provided with an I2C or SMBus interface, but generally there is only one interface. Therefore, access to the sensor by firmware control must access each sensor by serial processing. For this reason, in a server system having a configuration in which a large number of sensors are mounted, processing delay occurs as the number of sensors increases.

  On the other hand, when a large number of sensors cannot be read out within a certain time, there is a problem that the mounting accuracy of the number of sensors must be set, and the monitoring accuracy is lowered.

  For this reason, an object of the present invention is to solve the above-described problems of processing delay and deterioration of accuracy during monitoring.

A monitoring system according to one aspect of the present invention is:
A plurality of sensors for detecting a predetermined state in the system;
A controller for monitoring a state in the system based on a detection value by each sensor;
A sensor readout circuit physically connected to each of the sensors,
The sensor readout circuit directly reads out the detection value of each sensor from each of the sensors in response to a sensor readout instruction from the controller, and transmits the readout detection value to the controller.
Take the configuration.

A server system according to another embodiment of the present invention is
A server system equipped with predetermined components inside,
A plurality of sensors arranged in the server system for detecting a predetermined state in the server system;
A controller for monitoring a state in the server system based on a detection value by each sensor;
A sensor readout circuit physically connected to each of the sensors,
The sensor readout circuit directly reads out the detection value of each sensor from each of the sensors in response to a sensor readout instruction from the controller, and transmits the readout detection value to the controller.
Take the configuration.

Moreover, the monitoring method which is the other form of this invention is the following.
A plurality of sensors for detecting a predetermined state in the system;
A controller for monitoring a state in the system based on a detection value by each sensor;
A method of monitoring a system comprising a sensor readout circuit physically connected to each of the sensors,
In response to a sensor read instruction from the controller, the sensor read circuit directly reads the detection value from each sensor from each sensor, and transmits the read detection value to the controller.
Take the configuration.

  With the configuration as described above, the present invention can increase the monitoring accuracy while improving the monitoring accuracy.

It is a block diagram which shows the structure of the server system in Embodiment 1 of this invention. It is a flowchart which shows the monitoring operation | movement by the server system disclosed in FIG. It is a timing chart which shows the mode of the reading process of the sensor value by the server system disclosed in FIG. It is a timing chart which shows the mode of the reading process of the sensor value by the method different from this invention.

<Embodiment 1>
A first embodiment of the present invention will be described with reference to FIGS. FIG. 1 is a diagram showing a configuration of a server system, and FIG. 2 is a flowchart showing an operation. FIG. 3 is a timing chart showing how the sensor value is read out in the server system.

  As shown in FIG. 1, the server system 10 according to the present embodiment includes a plurality of CPUs (Central Processing Units) 1 to 4, a plurality of memories 1 to 4 (for example, DIMMs), and a cooling system for the server system. A plurality of FANs 1 to 4 for performing the above are provided as components. In the server system 10, a plurality of groups G1 to G4 are formed inside. Each of the groups G1 to G4 has the same configuration including one CPU and one memory, and FANs are arranged correspondingly. Here, each group G1-G4 is a computer module, such as one blade server. In addition, the component with which the server system 10 is provided is not limited to what was mentioned above, Moreover, each said group does not necessarily need to be formed.

  The server system 10 includes a plurality of sensors S1 and the like that detect the state of the server system 10 therein. The sensor S1 or the like is, for example, a temperature sensor that measures the temperature of the CPU or memory, a voltage sensor that detects the input voltage of the CPU or memory, a temperature sensor that detects the temperature in the server system housing, and the like. These sensors S1 and the like are arranged around the CPU 1 and the memory 1 as indicated by black circles S1 to S8 in the group G1 in FIG. Similarly, sensors are arranged in the other groups G2 to G4.

  Here, availability and reliability are required for the operation of the server system 10. On the other hand, in the server system 10, if an abnormality such as a temperature rise or a voltage drop occurs, there is a risk of causing a malfunction (garbled data) or failure of the CPU or memory. If such an abnormal state occurs, the server system cannot be operated in a healthy state. In order to monitor such an abnormal state, as described above, a large number of sensors, such as S1 to Sn, for measuring the temperature state of the CPU and memory, the state of the input voltage, and the like are provided. In particular, in a large server system, the number of parts constituting the system increases, and the number of sensors increases accordingly.

  In the above description, the sensor S1 and the like are mounted around the CPU and memory. However, temperature abnormalities and voltage abnormalities are not limited to the CPU and memory, so the status of other components is detected. A sensor may be provided. For example, sensors for monitoring the temperature around an HDD (Hard Disk Drive), the intake air temperature, the exhaust gas temperature in the server system, and the like may be arranged at an appropriate location.

  The server system 10 also includes a sensor read control circuit 30 (read circuit) that reads the detection values from the sensors S1 and the like described above from the sensors. The sensor readout control circuit 30 is physically directly connected to each of the sensors S1 and the like described above by a bus (signal line). Further, as shown in FIG. 1, the sensor read control circuit 30 includes a read command control unit 31 and a sensor value holding register 32.

  When receiving the sensor monitoring control command from the BMC 20 described later, the read command control unit 31 directly reads the sensor value (detected value) detected by each sensor via the bus. Then, the read command control unit 31 holds the read sensor values in the sensor value holding register 32. At this time, since the read command control unit 31 is directly connected to all the sensors of the server system 10 via the bus, the sensor values of all the sensors in each of the groups G1 to G4 are obtained for each of the groups G1 to G4. , They can be read together at a single timing. That is, the sensor values of all the sensors are read out at the same time at the same time and for each group. The sensor values read out for each of the groups G1 to G4 are stored in the sensor value holding register 32 for each of the groups G1 to G4.

  As described above, the sensor readout control circuit 30 is directly connected to the plurality of sensors by the bus, and thus can access the plurality of sensors collectively in parallel and read the sensor value. Therefore, with such a hardware configuration, it is possible to speed up the process of reading the sensor value. A processing unit for accessing a plurality of sensors in parallel is defined as a group.

  The sensor readout control circuit 30 transmits the sensor value held in the sensor value holding register 32 to the BMC 20. At this time, the sensor readout control circuit 30 sequentially transmits sensor values to the BMC 20 for each group.

  The server system 10 also includes a BMC (Baseboard Management Controller) 20 that is a diagnostic control unit that performs sensor monitoring control. The BMC 20 incorporates firmware that executes sensor monitoring control. Specifically, the firmware issues a sensor value reading instruction to the sensor reading control circuit 30. Then, the firmware receives the sensor value transmitted from the sensor readout control circuit 30, and monitors the abnormal state of temperature and voltage based on the sensor value.

  Next, the monitoring operation in the server system 10 having the above-described configuration will be described with reference to FIGS.

  First, the BMC 20 requests a sensor readout command from the sensor readout control circuit 30 in order to perform sensor monitoring of the server system 10 (step A1). In the present embodiment, the number of sensor readout commands from the BMC 20 need only be one.

  When the sensor read control circuit 30 receives a sensor read command from the BMC 20, the read command control unit 31 starts reading all sensor values classified into the groups G1 to G4 (step A2). This reading is performed simultaneously for each of the groups G1 to G4. That is, reading of sensor values is performed simultaneously from all sensors at once. The read sensor value is held in the sensor value holding register 32 for each group. The sensor readout control circuit 30 performs the readout again when one readout is completed, and repeats this operation (see FIG. 3). Therefore, when the sensor value is read from the BMC 20, the latest sensor value is always reflected (see the arrow in FIG. 3).

  After requesting the sensor read command, the BMC 20 sequentially reads the sensor values held in the sensor value holding register 32 for each group of G1 to G4 (step A3). Reading by the BMC 20 is polled at a certain interval. Specifically, as shown in FIG. 3, the BMC 20 sequentially reads the sensor values read by the sensor read control circuit 30 in the order of the groups G1, G2, G3, and G4.

  As described above, in the server system according to the present embodiment, sensor values are automatically read at a granularity classified into a certain group by hardware such as the sensor read control circuit 30 directly connected to each sensor S1 or the like. Is called. Therefore, the subsequent reading of the BMC 20 can be performed in the sensor value reading processing time from the sensor reading control circuit 30 for the number of groups. That is, the sensor readout processing time in this embodiment is “time a × G” at the longest.

  On the other hand, in another technique related to the present invention, as described above, the sensor value is read serially from each sensor by the BMC. For this reason, in such another technique, as shown in FIG. 4, if the processing time for reading the sensor value from one sensor is a, processing time for the number of sensors Sn is required.

That is, the sensor readout processing time in the present invention shown in FIG. 3 and the sensor readout processing time in another technique shown in FIG. 4 can be compared by the following expression.
That is, the following processing times can be compared.
Time a × Sn (different technology)> Time a × G (invention)
For example, when the number of sensors Sn is 256 and the number of groups G is 4, in the present invention, the sensor processing time is 1/64, and high speed can be realized.

  As described above, according to the present invention, sensor values are not read serially from a large number of sensors, but sensor values are read in parallel for each group of G1 to G4 by the sensor read control circuit 30. For this reason, in the present invention, the number of readings of sensor values can be reduced, and reading of a large number of sensor values can be speeded up. Accordingly, the sensor value can be read out in a state close to real time, and an abnormal state in the sensor monitoring control can be immediately detected. For example, when the temperature state of the server system becomes high, an abnormal state is detected immediately, so that rapid cooling control is possible. Furthermore, since the number of read commands from the BMC 20 is reduced, the load state of the firmware is also reduced.

  In the above-described embodiment, the sensors are classified into four groups G1 to G4 as an example. However, such a method is arbitrary, and the sensors of S1 to Sn are not classified into groups. May be read in order. Even in this case, since reading of all the sensors is automatically performed by the sensor reading control circuit 30, the number of reading commands issued by the BMC 20 is not changed.

<Appendix>
Part or all of the above-described embodiment can be described as in the following supplementary notes. The outline of the configuration of the monitoring system, server system, and monitoring method in the present invention will be described below. However, the present invention is not limited to the following configuration.

(Appendix 1)
A plurality of sensors for detecting a predetermined state in the system;
A controller for monitoring a state in the system based on a detection value by each sensor;
A sensor readout circuit physically connected to each of the sensors,
The sensor readout circuit directly reads out the detection value of each sensor from each of the sensors in response to a sensor readout instruction from the controller, and transmits the readout detection value to the controller.
Monitoring system.

(Appendix 2)
The monitoring system according to attachment 1, wherein
The sensor readout circuit reads out the detection values of the plurality of sensors collectively from the sensors at a single timing in response to a sensor readout instruction from the controller.
Monitoring system.

(Appendix 3)
The monitoring system according to attachment 2, wherein
The sensor readout circuit reads the detection values of the plurality of sensors belonging to a preset group collectively in the group unit, and transmits the detection values read out in the group unit to the controller.
Monitoring system.

(Appendix 4)
The monitoring system according to attachment 3, wherein
The sensor readout circuit reads out the detection value by the group unit simultaneously for a plurality of the groups.
Monitoring system.

(Appendix 5)
The monitoring system according to appendix 3 or 4,
The group is formed in units of modules that have the same configuration and are built in the system.
Monitoring system.

(Appendix 6)
The monitoring system according to any one of appendices 1 to 5,
The sensor readout circuit includes a register that temporarily stores detection values read from the sensors, and sequentially transmits the detection values stored in the registers to the controller.
Monitoring system.

(Appendix 7)
A server system equipped with predetermined components inside,
A plurality of sensors arranged in the server system for detecting a predetermined state in the server system;
A controller for monitoring a state in the server system based on a detection value by each sensor;
A sensor readout circuit physically connected to each of the sensors,
The sensor readout circuit directly reads out the detection value of each sensor from each of the sensors in response to a sensor readout instruction from the controller, and transmits the readout detection value to the controller.
Server system.

(Appendix 8)
The server system according to appendix 7,
The sensor readout circuit reads out the detection values of the plurality of sensors collectively from the sensors at a single timing in response to a sensor readout instruction from the controller.
Server system.

(Appendix 9)
A plurality of sensors for detecting a predetermined state in the system;
A controller for monitoring a state in the system based on a detection value by each sensor;
A method of monitoring a system comprising a sensor readout circuit physically connected to each of the sensors,
In response to a sensor read instruction from the controller, the sensor read circuit directly reads the detection value from each sensor from each sensor, and transmits the read detection value to the controller.
Monitoring method.

(Appendix 10)
The monitoring method according to attachment 9, wherein
The sensor readout circuit reads out the detection values of the plurality of sensors collectively from the sensors at a single timing in response to a sensor readout instruction from the controller.
Monitoring method.

  Although the present invention has been described with reference to the above-described embodiment and the like, the present invention is not limited to the above-described embodiment. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

10 server system 20 BMC
30 Sensor Read Control Circuit 31 Read Command Control Unit 32 Sensor Value Holding Registers S1 to S8 Sensor

Claims (10)

A plurality of sensors for detecting a predetermined state in the system;
A controller for monitoring a state in the system based on a detection value by each sensor;
A sensor readout circuit physically connected to each of the sensors,
The sensor readout circuit directly reads out the detection value of each sensor from each of the sensors in response to a sensor readout instruction from the controller, and transmits the readout detection value to the controller.
Monitoring system. The monitoring system according to claim 1,
The sensor readout circuit reads out the detection values of the plurality of sensors collectively from the sensors at a single timing in response to a sensor readout instruction from the controller.
Monitoring system. The monitoring system according to claim 2,
The sensor readout circuit reads the detection values of the plurality of sensors belonging to a preset group collectively in the group unit, and transmits the detection values read out in the group unit to the controller.
Monitoring system. The monitoring system according to claim 3,
The sensor readout circuit reads out the detection value by the group unit simultaneously for a plurality of the groups.
Monitoring system. The monitoring system according to claim 3 or 4,
The group is formed in units of modules that have the same configuration and are built in the system.
Monitoring system. The monitoring system according to any one of claims 1 to 5,
The sensor readout circuit includes a register that temporarily stores detection values read from the sensors, and sequentially transmits the detection values stored in the registers to the controller.
Monitoring system. A server system equipped with predetermined components inside,
A plurality of sensors arranged in the server system for detecting a predetermined state in the server system;
A controller for monitoring a state in the server system based on a detection value by each sensor;
A sensor readout circuit physically connected to each of the sensors,
The sensor readout circuit directly reads out the detection value of each sensor from each of the sensors in response to a sensor readout instruction from the controller, and transmits the readout detection value to the controller.
Server system. The server system according to claim 7,
The sensor readout circuit reads out the detection values of the plurality of sensors collectively from the sensors at a single timing in response to a sensor readout instruction from the controller.
Server system. A plurality of sensors for detecting a predetermined state in the system;
A controller for monitoring a state in the system based on a detection value by each sensor;
A method of monitoring a system comprising a sensor readout circuit physically connected to each of the sensors,
In response to a sensor read instruction from the controller, the sensor read circuit directly reads the detection value from each sensor from each sensor, and transmits the read detection value to the controller.
Monitoring method. The monitoring method according to claim 9, comprising:
The sensor readout circuit reads out the detection values of the plurality of sensors collectively from the sensors at a single timing in response to a sensor readout instruction from the controller.
Monitoring method.

Images