WO2016073005A1

WO2016073005A1 - Operationally independent reset

Info

Publication number: WO2016073005A1
Application number: PCT/US2014/064645
Authority: WO
Inventors: Nilashis DEY; Binu John BINU; Albert H CHANG
Original assignee: Hewlett Packard Enterprise Development Lp
Priority date: 2014-11-07
Filing date: 2014-11-07
Publication date: 2016-05-12

Abstract

Techniques for resetting operationally independent devices, for example, computers, are described. In one aspect, a request may be received at a first computer. The request may be a request to reset a second computer. The second computer may be reset from the first computer. The first computer and the second computer may be operationally independent.

Description

OPERATIONALLY INDEPENDENT RESET BACKGROUND

[0001] Physical access to the computer systems located in a data center is becoming more and more difficult. In many cases, for security reasons, the operator of a data center may restrict physical access to a limited number of trusted personnel. Thus, persons not on the trusted list may be locked out of physically accessing the devices in the data center. In addition, many data center operators are locating their data centers in remote locations. For example, some data center operators may locate their data centers in areas that are naturally cold (e.g. above the Arctic Circle) such that the lower ambient temperature may be used to aid in cooling the data center, thus reducing the expense incurred from the operation of air conditioning units. The remote location may make physical access to the data center difficult.

BRIEF DESCRIPTION OF THE DRAWINGS

[0002] FIG. 1 is a block diagram of an example device that may utilize the operationally independent reset techniques described herein.

[0003] FIG. 2 is an example of a circuit that may be used in the example device described in FIG. 1 .

[0004] FIG. 3 is an example of a system including a non-transitory medium containing processor executable instructions that may be used in an implementation of the operationally independent reset techniques described herein.

[0005] FIG. 4 is example of a flow diagram of an implementation of the operationally independent reset techniques described herein. [0006] FIG. 5 is another example of a flow diagram of an implementation of the operationally independent reset techniques described herein.

DETAILED DESCRIPTION

[0007] The lack of physical access to the computers in a data center typically does not present a problem. The computers in the data center are typically connected to a network which allows for remote access to the computers. Maintenance or administration of the computer may typically involve an administrator logging into the computer remotely over the network, and performing any needed actions. As such, the administrator may typically have no need to physically access the computer.

[0008] Although the administrator typically need not have physical access to the computers being maintained, there are certain failure scenarios in which physical access is beneficial. For example, in some cases, the computer may have completely locked up, such that the processor on the computer is no longer executing instructions. Because the processor is no longer executing, it may not be possible for the administrator to remotely log into the computer, because the remote log in process may utilize the processor. In some cases, the network interface between the computer and the network may have failed, thus preventing the computer from communicating over the network. If the computer cannot communicate over the network, the administrator cannot remotely log into the computer.

[0009] Often times, a hard reset of the computer may cause the computer to restart. The process of restarting may cause the computer to reinitialize its hardware and software, which may possible resolve the issue that was causing the computer to become unresponsive. A hard reset, or simply a reset, of the computer may be a simple matter of physically turning the computer off and back on through a physical switch or button.

[0010] Although a reset may resolve the issues of an unresponsive computer, such a reset may be difficult in cases where a data center is secured. Assuming the administrator is not one of the trusted individuals, the administrator must first find a trusted individual. The administrator may then have to plead his case as to why he should be allowed physical access to the data center. In the case of a remotely located data center, physical access to the data center may be even more difficult, as it may require travel to the data center. For example, in the case of a "lights out" data center, there may be no human presence in the data center, such that there may not even be a person who can be given instructions to reset the computer.

[0011] The techniques described herein overcome the issue of the inability to perform a hard reset of a computer remotely. A device, such as a computer, may have a connection to a network, over which requests to reset an external device may be received. The device and the external device may be coupled with a circuit. The device and the external device may be operationally independent. The device may reset the external device using the circuit upon receipt of the request to reset the external device.

[0012] FIG. 1 is a block diagram of an example device that may utilize the operationally independent reset techniques described herein. FIG. 1 depicts a system 100 that may include a network 1 10, a device 120, a circuit 130, and an external device 140. The network 1 10 may be any suitable network that allows for communications between computing devices. For example, the network may be an Ethernet or wireless based network, an intranet, an internet, or the Internet. The network may be a public or a private network. The particular form of the network is not important. Rather, it should be understood that network 1 10 may allow communication between a remote user (not shown) and device 120.

[0013] Device 120 may be any device that is capable of being connected to a network for remote management. For example, device 120 may be a computer, such as a server computer. However, it should be understood that device 120 is not limited to computers. Device 120 may be a computer, a tablet, a phone, a wearable computing device, or any other type of device. The techniques described herein are not dependent on any particular form factor for device 120. [0014] Device 120 may be coupled to a circuit 130. An example of an implementation of circuit 130 is depicted in FIG. 2. The circuit 130 may operatively couple device 120 with an external device 140. The circuit 130 may couple the device 120 and external device 140 while at the same time allowing both of those devices to remain operationally independent. For purposes of this description, operationally independent means that the devices 120 and 140 are not dependent on each other for operation.

[0015] For example, each device may operate even when the other device is in an unresponsive or powered off state. Furthermore, faults on one device may not affect the other device. In addition, the operational

independence means that the two devices do not share resources, such as power rails. As such, the power supplies of the device and the external device may be reset without affecting the operation of the other device.

[0016] In operation, a request to reset the external device 140 may be received by the device 120. The request may be received by the device 120 over the network 1 10. The request may have been sent from a remote administrator (not shown). The device 120 may receive the request and utilizing circuit 130 may cause the external device 140 to reset. For example, the external device may be reset by resetting the power supply of the external device or by resetting a processor of the external device. As mentioned above, because the device and external device are operationally independent, a reset of the external device may have no impact on the device.

[0017] Although FIG. 1 has been described with respect to the device 120 using the circuit 130 to reset the external device 140, it should be understood that the techniques described herein are not so limited. For example, the external device 140 may also be coupled to the network 1 10 (not shown) and may receive requests to reset the device 120. A second circuit, similar to circuit 130 (not shown) may be provided to allow for the external device 140 to cause the device 120 to reset.

[0018] FIG. 2 is an example of a circuit 200 that may be used in the example device described in FIG. 1 . It should be understood that the example circuit depicted in FIG. 2 is one possible implementation of a circuit that may implement the techniques described herein. Circuit 200 may be divided into two halves, as depicted by the dashed line 210. For purposes of this description, the left half of circuit 200 is described as belonging to the device that will be used to reset the external device. The right half of the circuit will be described as belonging to the external device that is going to be reset. In other words, the left half of the circuit corresponds to device 120, while the right half of the circuit corresponds to the external device 140. Although not depicted, it should be understood that a mirror circuit may also exist, in which the external device is able to reset the device.

[0019] The left half of the circuit may include transistors Q1 , Q2, and Q3, as well as resistors R1 , R2, and R3. In addition, the circuit may include a reset in signal 220. The reset in signal may be a signal received from the processor of the device 120 indicating that external device 140 is to be reset. For example, the signal may be a signal generated by a general purpose input / output (GPIO) pin of a processor in device 120. Processors are described in further detail with respect to FIG. 3. What should be understood is that when the device wishes to reset the external device, the reset in signal 220 may be utilized.

[0020] In normal operation, the reset in signal will be pulled to ground through resistor R1 . As such, Q1 will be in the off state and the gate of Q2 will be pulled to V1 , through R2. Since the gate of Q2 is high, Q2 will be in the on state, resulting in the voltage V1 being dropped across R3, which results in the gate of Q3 being pulled to ground, causing Q3 to remain in the off state. AS such, no current will flow from the drain to source of Q3.

[0021] Because no current flows through Q3, the photodiode of opto couple 230 will remain off, resulting in no current flowing through the photo detector of the opto couple. As such, the reset out 240 will be pulled to V2 through resistor R5. As will be described below, the reset out signal may be used to reset the external device. For example, the signal may be used to reset the processor or power supply of the external device. It should also be noted that the left hand and right hand portions of the circuit share no power or ground rails. The only connection between the two halves of the circuit are depicted along the dashed line 210. As such, the device and the external device remain operationally independent, as the operation of each system is independent of the other.

[0022] For example, if the power supply providing V1 to the left hand portion of the circuit were to fail, there would be no impact to operation of the external device, which is associated with the right hand portion of the circuit. Likewise a failure of the power supply providing V2 to the right hand portion of the circuit would have no impact on the device, which is associated with the left hand portion of the circuit.

[0023] When the device wishes to reset the external device, the processor of the device may cause the reset in signal to be asserted. This results in Q1 turning on, which in turn results in the gate voltage of Q2 being pulled to ground through R2, which in turn results in Q2 turning off. Accordingly, the gate voltage of Q3 will be pulled to V1 through resistor R3, causing Q3 to turn on.

[0024] Because Q3 has turned on, current will flow from V2, through R4, and cause the photodiode of opto couple 230 to turn on. This in turn causes the photo detector of the opto couple to turn on, causing current to flow from V2, through R5, resulting in reset out 240 being pulled to ground. The external device may be designed such that a ground signal on reset out 240 causes the external device to reset.

[0025] FIG. 3 is an example of a system 300 including a non-transitory medium containing processor executable instructions that may be used in an implementation of the operationally independent reset techniques described herein. The system shown in FIG. 3 may include a computer 310, a computer 330, and a network 380 coupling the computers. The system may also include a circuit 360 coupling the computers 310, 330. In addition, the network 380 may couple the computers to an administrator 390.

[0026] The computer 310 may include a processor 31 1 . The processor may be of any type suitable for executing processor executable instructions. For example, the processor may be a processor that executes x86 instructions, ARM instructions, or any other suitable instruction set. Furthermore, the processor may be the main processor, such as the Central Processing Unit (CPU) of the computer 310. The processor may also be an auxiliary processor. For example, the processor may be a management processor, which in some cases may be referred to as a baseboard management processor. In some cases, the processor may be another processor on the computer 310, such as a graphics processing unit (GPU). The particular form of the processor is unimportant. What should be understood is that the processor is able to execute instructions in order to implement the techniques described herein.

[0027] The computer may also include a non-transitory processor readable medium 312 coupled to the processor. The medium may store instructions, such as reset instructions 313, which when executed by the processor cause the processor to implement the techniques described herein. For example, the reset instructions may cause the processor to implement the operationally independent reset techniques described above. The reset techniques are described further below.

[0028] The computer 310 may also include a network interface 314. The network interface may couple the computer to network 380. The computer may communicate with external entities over network 380. For example, the computer 310 may communicate with administrator 390 in order to receive reset commands as described above. For example, a command to reset another computer, such as computer 330, may be sent form the administrator 390 over the network 380 to the computer 310. The computer 310 may execute the reset instructions 313 using the processor 31 1 to reset the second computer. For example, the reset instructions may cause the processor to utilize circuit 360 to reset the second computer 330.

[0029] Circuit 360 may be a circuit that couples computers 310, 330. However, as described above, the computer remains operationally independent. Such, operation of each computer is independent of the other computer. As explained above, the computer may share no power rails. One example of circuit 360 is shown with respect to FIG. 2. Using circuit 360, the computer 310 may cause a reset of the computer 330. [0030] Computer 330 may have a structure that is similar to computer 310. For ease of description, certain components, such as the non-transitory medium containing instructions have been omitted. It should be noted that the operation of the system described in FIG. 3 is not dependent on the processor 331 executing instructions. As described above, one of the issues that the techniques described herein resolve is a situation in which the computer 330 is unresponsive. This may occur, for example, when the processor 331 has stopped executing instructions, and is thus incapable of executing remote administration commands.

[0031] The computer 331 may include a network interface 334. However, it should be noted the techniques described herein are not dependent on the proper functioning of the network interface. For example, in one possible scenario, the computer 330 may be unresponsive to remote administration commands because the network interface 334 has stopped functioning. In some cases, a hard reset of the computer 330 may cause the network interface to reinitialize and restart, thus recovering from the fault that caused the network interface to cease functioning in the first place.

[0032] The computer 330 may also include a power supply 335. As mentioned above, the computers 310,330 may be operationally independent, meaning that they share no resources such as power rails. Thus a reset of the power supply of computer 330 may have no impact on the operation of computer 310. The circuit 360 may be used to cause the power supply 335 of computer 330 to reset, which may also be referred to as power cycling the power supply. Resetting the power supply may cause the computer 330 to restart, potentially resolving the issue that caused computer 330 to become unresponsive.

[0033] In another example implementation, the circuit 360 may be coupled directly to the processor 331 of the computer 330. The circuit 360 may be utilized to directly cause a reset of processor 331 . A reset of processor 331 may have the same result as power cycling the power supply 335. In other words, computer 330 may restart and the restart may clear the error condition that cased computer 330 to become unresponsive in the first place. [0034] FIG. 4 is example of a flow diagram of an implementation of the operationally independent reset techniques described herein. For example, the reset instructions 313 depicted in FIG. 3 may cause the computer 310 to implement the flow diagram described in FIG. 4. In block 410, a request to reset a second computer may be received at a first computer. As explained above, in some cases, the second computer may be inaccessible remotely, thus the second computer cannot be reset directly. For example, the second computer may no longer be responsive to external administrative commands. However, the techniques described are not limited to cases in which the second computer is unresponsive.

[0035] In block 420, the second computer may be reset from the first computer. The first and second computer may be operationally independent. As described above, because the two computers are operationally independent, the reset operation is not dependent on both computers being up and responsive. As long as the first computer is able to respond to requests received from a remote administrator, the first computer may be able to utilize the circuit described above to reset the second computer. The reset of the second computer is not dependent on the second computer being responsive to commands received over the network. Rather, the reset behavior may be that which would be expected had the second computer been reset locally.

[0036] FIG. 5 is another example of a flow diagram of an implementation of the operationally independent reset techniques described herein. As described with respect to FIG. 4, the flow diagram in FIG. 5 may be executed by the computer 310 depicted in FIG. 3. In block 510, a request to reset the second computer may be received over a network by a first computer. As described previously, techniques described herein may be used to remotely manage computers in a data center. As such, administration commands may be received form a remote administrator over a network.

[0037] In block 520, just as in block 420, the second computer may be reset from the first computer. The first and second computer may be

operationally independent. In block 530, the second computer may be power cycled. As described above, in one implementation, a reset of the second computer may occur by power cycling the power supply that provides power to the second computer. As explained above, because the first and second computer are operationally independent, power cycling the power supply of the second computer may have no effect on the operation of the first computer. In some cases, the power cycling of the second computer may cause the second computer to restart, which may in turn cause the second computer to recover from the condition that had originally caused the second computer to become unresponsive.

Claims

We Claim:

1 . A method comprising:

receiving, at a first computer, a request to reset a second computer; and resetting the second computer from the first computer, wherein the first and second computer are operationally independent.

2. The method of claim 1 further comprising:

receiving, at the first computer, the request to reset the second computer over a network.

3. The method of claim 1 wherein the second computer is unresponsive to external communications.

4. The method of claim 1 further comprising:

power cycling the second computer.

5. The method of claim 1 wherein a reset of the first computer does not cause a reset of the second computer.

6. The method of claim 1 wherein the first and second computer do not share power rails.

7. A device comprising:

a connection to a network to receive a request to reset an external device; and

a circuit coupling the device to the external device, the device and the external device being operationally independent,

wherein the device resets the external device using the circuit upon receipt of the request to reset the external device.

8. The device of claim 7 wherein a reset of the device does not cause a reset of the external device.

9. The device of claim 7 wherein operationally independent further comprises the device and external device having no power rails in common.

10. The device of claim 7 further comprising:

the external device including a connection to a network to receive a request to reset the device; and

a circuit coupling the external device to the device, the circuit sharing no power rails with the device,

wherein the external device resets the device using the circuit upon receipt of the request to reset the device.

1 1 . The device of claim 10 wherein a reset of the external device does not cause a reset of the device.

12. A non-transitory processor readable medium containing a set of instructions thereon, which when executed by the processor cause processor to:

receive a request to reset an operationally independent computer; and resetting the operationally independent computer using a circuit that operatively couples the processor to the operationally independent computer.

13. The medium of claim 12 wherein the processor and the operationally independent computer share no power rails.

14. The medium of claim 12 wherein the request to reset the operationally independent computer is received over a network.

15. The medium of claim 12 wherein the instructions to reset the operationally independent computer includes instructions to reset a power supply of the operationally independent computer.