KR0168947B1 - Method for booting node without disk in real-time distributing system - Google Patents

Abstract

The present invention relates to a method of booting a node without a disk in a real-time distributed system. First, a method for booting a node without a disk in a distributed system is provided. Second, the first method does not have a disk according to the effect. Simplify the structure of the hardware by reducing the hardware structure by configuring a distributed system with nodes, and third, by having a structure that can separate the communication hardware devices and drivers into one module (interworking) There is an effect that can be applied to hardware.

Description

How to boot a node without a disk in a real-time distributed system

1 is a schematic structural diagram of hardware to which the present invention is applied.

2 (a) to 2 (c) are block diagrams for explaining the schematic operation of the present invention.

3 is a diagram illustrating a booting protocol provided by the present invention.

4 is a flowchart illustrating an embodiment of a booting method according to the present invention.

The present invention relates to a node booting method having no disk in a real time distributed system.

The control system of an Asynchronous Transfer Mode (ATM) exchange is a large distributed architecture in which numerous nodes and controllers are interconnected through an ATM switch. The control system performs the control and management of all resources and hardware devices of the ATM switch. This function is distributed to each node and the various software (operating system and application) necessary for this are nodes with hard disk. (Node With Disk: hereinafter referred to as NWD). When the ATM switch is powered on, the software on the disk is loaded into the main memory of each node. The control system of the ATM switch has a structure in which only one node among the nodes constituting the system has a disk in order to reduce the cost of hardware. Therefore, a special boot procedure is required for nodes without disks (NWOD).

Accordingly, the present invention was devised in response to the above-described demands, and is not booted by remotely loading an operating system for booting nodes without a disk in an ATM switching system having a distributed structure, but by each node as well as an operating system. Its purpose is to provide a way to boot a node that does not have a disk with requirements that uniquely needed applications need to be loaded.

In order to achieve the above object, the present invention provides a booting method for a node having no disk in a real-time distributed system, comprising: a first step of checking a type of a head when a head frame is received; A second step of obtaining a starting address and a size value of a memory to be loaded when receiving an operating system head, and then changing the operating system code to a state to receive and returning the operating system code; In the first step, after receiving the application program loader head, obtaining a start address and a size value of the memory to be loaded, and then changing and returning the application program loader code to a state to be received; Receiving a code frame, checking a type of code to be loaded; A fifth step of checking whether or not to load each code according to the reception of the operating system code or the application program loader code, if not, and if it is loaded, checking whether it is in a normal state; A sixth step of loading a boot image into main memory if not normal, notifying and returning the result to the system loader, and notifying and returning to the system loader not normal if not normal; A seventh step of checking whether an application loader end frame is received when receiving the end frame; An eighth step of returning if it is not an application loader end frame and an operating system end frame, otherwise returning and returning the application loader end frame to a state to receive; And in the seventh step, in the case of the application program loader end frame, informing the system loader of the end of the loading process, executing a boot image to transition to the starting state.

Hereinafter, with reference to the accompanying drawings will be described in detail the present invention.

FIG. 1 is a schematic structural example of hardware to which the present invention is applied. Each node has a central processing unit (CPU), a memory, and an ATM switch interface as a basic component, and only a node having a disk (NWD) As a control device for access, it has a mass storage interface board assembly (MSIA) and a disk. The disk stores an operating system and applications.

2 (a) to 2 (c) are block diagrams for explaining the schematic operation of the present invention.

Referring to Figure 2 (a) the overall operation of the present invention will be described.

When power is supplied, each node constituting the distributed control system needs to load and execute necessary software (operating system and application program) in its main memory in order to perform its own function. This process is commonly referred to as booting, which is divided into two types depending on the presence of auxiliary storage. A node with a disk (hereinafter referred to as NWD (Node With Disk)) directly accesses its own disk, loads the necessary software into main memory, executes and boots it, and has no disk (hereinafter referred to as NWOD (Node WithOut Disk). ) Uses the Inter Process Communication (IPC) communication function unique to the ATM exchanger's control system to load and execute the software on the disk into its main memory, which is called a remote booter.

From the remote (hereinafter referred to as RB), a system loader (SL) that is mounted and operated in NWD and an application loader for loading a unique application for each operating system and node according to a predefined boot protocol for booting Loader: Loads and executes APL).

Booting from the remote is classified into two cases: NWD's own judgment and NWOD's request for loading.

First, self-judgment is the case when the NWD is judged by the necessity of periodically loading the NWODs or when there is a need to replace it with a new operating system, and when the NWOD requires loading, When the power is supplied and operated for the first time, it occurs when it is down and regenerated by any malfunction. When the necessity of loading occurs due to the above two cases, the loading of the APL loading the operating system and the application program is performed according to the predetermined boot protocol, and when the loading is completed, the operating system and the APL are executed remotely. . Next, node-specific application programs are loaded by IPC communication between APL of NWOD and SL of NWD. In this way, the execution from the remote is divided into a necessity stage for loading, a loading stage, and a starting stage for execution.

FIG. 2 (b) shows a process in which the necessity of loading is generated by the NWD during the necessity of loading, and the loading is required. FIG. 2 (c) shows a process in which necessity of loading is generated by the NWOD and the loading is required. .

The process of generating the necessity of loading shown in FIG. 2 (b) is as follows.

When an NWD asks each NWOD their status via IPC (PrCheck), each NWOD that receives this IPC responds to its status via IPC (PrStatus). The status of the NWOD notified by this response (PrStatus) is two: the status indicating that the operating system is currently being loaded, and the operating system should be loaded. First, if the operating system needs to be loaded, that is, the first loading state, the SL of the NWD performs a loading step for the NWODs that have reported the state.

The loading necessity generation process shown in FIG. 2 (c) is as follows.

If the NWOD is powered up, or if there is a need to load the operating system, operating system, or APL due to other problems, remotely send the NWD's SL an IPC (LdRequest) to request loading, and the SL will allow the IPC (LdGrant to load). ) To perform the loading step.

3 is a procedure for the process of performing the boot protocol provided by the present invention.

When the remote requests a loading (LdRequest), the SL accepts the loading request and sends a response (PrCheck) to notify the result of the receipt. Remotely responds to this (PrStatus) and the SL, which receives this response, is ready to be loaded. On the other hand, upon receiving the response (PrStatus), the SL informs the head address (H_frame) of the address value of the main memory device and the size of the data from which the data will start loading from the remote and starts sending data to be loaded. The data sent for loading is sent in the basic message unit of the IPC. From the remote, it notifies the SL whether or not it is normally received for each message. At this time, if the normal reception is notified to the SL to receive the message again. If even the final data is normally received, ie, loaded, the SL notifies that all data is loaded normally and executes the loaded data.

4 is a flowchart illustrating an embodiment of a NWOD booting method according to the present invention.

The need for loading from the remote end The operation from the remote is started in the state of finishing the process of waiting for loading.

In the reception state of the operating system, when the operating system head related to the operating system and the PL head related to the APL loading the application program are inputted (2, 3) (respectively, the start address and size of the main memory to load the operating system and the APL are inputted). Check whether the current state is to receive the head frame (4, 7), and if it is to be loaded, obtain the address of the main memory to be loaded and the size of the data to be loaded. Ignore IPC messages (5,8). In each case, the code frame for the operating system or the code frame for the APL is received and returned (6, 9).

When a code frame is received (10), it is classified as a code frame corresponding to an operating system or a code frame corresponding to an APL (11) and checked whether it is to be loaded for each case (12, 13). To check the data integrity of one IPC message (each corresponding code frame), the sum check checks for errors (14). At this time, if there is an error, it sends a NAK signal to the SL of the NWD to notify the error (17), if there is no error, loads it into the main memory (15), sends an ACK signal to the SL of the NWD to notify the normal reception result (16) and returns it. .

Receiving an end frame (18) checks whether it is an end frame for APL loading (19), checks whether it is an end frame for an operating system if it is not an end frame for APL (20), and if it is an end frame for an operating system Change the head frame to receive (21), or ignore the received IPC message. When receiving the end frame for the APL, a signal (PLOK) is transmitted to the SL of the NWD indicating that the loading step is completed (22), and the booting operation is terminated by executing the operating system and the APL (23). Transitions to the starting state, which is the state of loading the application.

The present invention described above is not limited to the above-described embodiments and the accompanying drawings, and various substitutions, modifications, and changes are possible in the art without departing from the technical spirit of the present invention. It will be clear to those of ordinary knowledge.

The effects of the present invention as described above are as follows.

First, unlike the conventional network booting, only the application programs suitable for the unique functions to be performed including the operating system for each node distributed to each node can be loaded so that an optimal distributed system can be configured at the booting stage.

Second, we propose a simpler and more reliable boot protocol for ATM switches than the protocol defined for conventional network booting.

Third, by allowing the operating systems and applications necessary for each node constituting the exchange system to be concentrated on one node having a disk, the other nodes do not need to have a disk, thereby reducing the cost of hardware.

Claims (1)

A booting method for a node having no disk in a real time distributed system, comprising: a first step of checking a type of a head when a head frame is received; A second step of obtaining a start address and a size value of a memory to be loaded when receiving an operating system head, and then changing the operating system code to a state to receive and returning the operating system code; In the first step, after receiving the application program loader head, obtaining a start address and a size value of the memory to be loaded, and then changing and returning the application program loader code to a state to be received; Receiving a code frame, checking a type of code to be loaded; A fifth step of checking whether or not to load each code according to the reception of the operating system code or the application program loader code, if not, and if it is loaded, checking whether it is in a normal state; A sixth step of loading a boot image into main memory if not normal, notifying and returning the result to the system loader, and notifying and returning to the system loader not normal if not normal; A seventh step of checking whether an application loader end frame is received when receiving the end frame; An eighth step of returning if it is not an application loader end frame and an operating system end frame, otherwise returning and returning the application loader end frame to a state to receive; And in the seventh step, in the case of the application loader end frame, informing the system loader of the end of the loading process, executing a boot image, and transitioning to a starting state. .