JP2000148634A - Animation server system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To save a stream transmitting device by operating the connection processing of plural network streams to be transmitted from plural file servers on network equipment at reproducing streams. SOLUTION: This system is an animation server system for transmitting streams, in response to a request from a client 30. The system is provide with plural file servers 10 for storing plural stream files generated, by previously operating division processing by distributing them to plural file systems and network equipment 20 for operating connection processing to the previously processed stream group from the plural file servers 10, response to the request from the client 30, and for transmitting it as a single stream.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a moving picture server system for providing a plurality of compressed / uncompressed moving pictures and sounds.

[0002]

2. Description of the Related Art In order to use a plurality of video-on-demand (VOD) servers connected to a network to transmit video to a large number of users more than the stream transmission capacity of one VOD server to many users. In general, the following method is used. Generally, the following method is used as a method for transmitting video more than the stream transmission capability of one VOD server and responding to many users.

(1) In-server RAID system A system in which a disk array (RAID) composed of a plurality of hard disks is connected to a plurality of servers to provide information on the performance of a single hard disk or more as a whole system. Stream transmission exceeding the capacity of one server is realized by arranging the same stream in a plurality of servers.

(2) Server-to-server RAID system A system in which one stream is divided into a fixed length and arranged cyclically on a plurality of servers. This method is (1)
Compared to the in-server RAID method, since one server reads one stream in parallel, a method such as time slot is used so that concentration on a specific server does not occur. The required space of the hard disk can be reduced.
-205634).

When reading stream data from a file server, one file server forms a part of the file system storing the stream data. The stream sending unit accesses the logical block unit, and the stream sending device interprets the file system.

In the stream RAID system, the server area network is configured to use a device for switch connection between servers so as to avoid a bottleneck in the network in the shared path system.
A store-and-forward method is known as a method of realizing a switch-connected device used in this system. In this method, a received frame is placed on a RAM, a frame is transmitted from the RAM to a port to be transmitted after waiting for a port to be transmitted, and thereby a collision that causes a decrease in performance is prevented.

The present invention relates to the above-mentioned (2) inter-server RAID system.

[0008]

In the stream RAID system, there is a demand for a moving picture server system having a simpler structure, with a focus on improving the stream sending ability rather than the reliability.

It is an object of the present invention to provide a moving image server system which can further simplify a series of system configuration including a server network, a stream transmitting device, and a client network conventionally provided between a file server and a client. It is the purpose.

[0010]

In order to achieve the above object, a moving image server system according to the present invention has a file system having a file system for distributing and storing a plurality of stream files generated by performing division processing in advance. A file server and a group of pre-divided streams sent from a plurality of file servers using a protocol suitable for a stream in response to a request from a client to perform a concatenation process to form one stream And a network device for sending to the client.

[0011] By making the connection processing unit a packet unit, it is not necessary to wait for a stream and to divide an Ethernet frame, which are conventionally required, and a simplified moving image server system can be provided.

[0012]

FIG. 1 shows an example of the configuration of a moving image server system according to the present invention. The system shown in FIG. 1 includes a plurality of file servers 101, 102, and 103 collectively referred to as a file server 10, and a plurality of clients 301, 302, 303, and
File server 1 upon receiving a request from any of
0, and a network device 20 for extracting a desired stream and sending it to the client.

FIG. 2 shows a hardware configuration of the file server 10. File servers 101 to 103
Have a CPU 11 and a local bus 12
Via the RAM 13 and the hard disk drive (H
DD) A plurality of hard disk drives (HDDs) 151, 152, 1 via an interface (I / F) 14
An HDD 15 comprising a network interface (I / F) 16 is provided in a terminal unit for receiving a request from the network device 20 or transmitting the requested stream.

Each of the file servers 101, 102, 1
In the file system FS 03, a stream file generated by dividing one stream in advance is stored. The file server 10 is a POSIX
The read process RP operates as a user process in accordance with the compliant UNIX protocol. The stream file is opened by the open system call through the read process RP, read out by the read system call, temporarily stored in the FIFO buffer FB, and transmitted from the transmission process SP to the network device 20.

The FIFO buffer FB forms a ring buffer between the FIFO buffers FB in each of the file servers 101 to 103 in accordance with a distributed memory system which is one of the inter-process communication functions of the UNIX protocol. In this distributed memory system, since the OS is allocated from the RAM of the file server, it can be allocated with an optimum size according to the speed at which the stream flows and the speed at which the file is read. Read process RP on file server 10
Then, the reading order is scheduled based on the free ratio of the FIFO buffer FB, or
The scheduling can be performed based on the free space of B, the speed at which the stream flows, and the size read by the read process RP at a time. In that case, ones with different stream speeds can be mixed. In the read process RP on the file server 10, the data read by the read system call is stored in the FI corresponding to the scheduling result in accordance with the scheduling order.
Put into FO buffer FB. At this time, detection of a failure of the hard disk drive or the like can also be performed. The sending process SP on the file server 10 operates as a user process. This reads data from the FIFO buffer FB and sends a network stream to the network device 20. As the transmission rate control at the time of transmission to the network device 20, a first method in which the file server 10 performs control at a stream bandwidth reserved speed, a second method by flow control from the network device 20 or the client 30, and the like. There is. Among them,
As a second method, ie, a flow control method, TCP / I
There is a method of using a protocol such as P (the fourth layer of the OSI reference model), and a method of using the fifth layer session layer or the seventh application layer of the OSI reference model on the UDP / IP protocol.

FIG. 3 shows a hardware configuration of the network device 20. The network device 20 is a CP
U21 and a shared RAM via a shared bus 22.
23, a frame rewrite processing unit 24, a stream flow processing unit 25, and transmission / reception ports 26a, 26b,.
6n are connected. The transmission / reception ports 26a to 26a-2
6n is generally referred to by reference numeral 26. Each transmission / reception port 26 transmits
FIFO 27, receive FIFO 28, and network I / F 29
Contains. Incidentally, the conventional EtherSW does not include the frame rewriting processing unit 24 and the stream flow processing unit 25 as shown in FIG.

Transmission / reception port 26 of network device 20
Receives the Ethernet frame from the file server 10 and sends it via the FIFO buffer 28 to the shared RAM 2.
Write to 3. Thereafter, the packet is transferred to an appropriate port according to the destination MAC address of the Ethernet frame. When the network device 20 has a routing function, the destination MAC address of the Ethernet (Ethernet) frame is checked to determine whether or not the routing process is necessary. Forward to the appropriate port.

The network device 20 determines whether or not the stream is a network stream as follows, and if the stream is a network stream, performs an operation different from the above. 1) Whether the packet type is IP in the Ethernet frame. 2) In the IP datagram in the Ethernet frame, is the destination IP address reserved for the destination of the network stream? 3) UDP in IP datagram in Ethernet frame
In the packet or TCP packet, is the destination port reserved for the destination of the network stream?

Regarding the reservation, in the case of a UDP packet,
It can be set statically, or can be performed by the reservation protocol from the file server 10 before sending the stream. T
In the case of a CP packet, it can be performed by a method in which the network device 20 has an IP address and operates as a Proxy Server.

When the Ethernet frame received at the port 26 is recognized as a network stream, the network device 20 transmits the received data to the shared RAM via the FIFO buffer 28 of the receiving port, the frame rewriting processor 24, and the shared RAM.
23 is written. In the frame rewriting processing unit 24, an Ethernet frame header, an IP header, a UDP
Header, TCP header, CRC, and a protocol for stream transfer (for example, RTP (Real-time Transport Pr
otocol)) performs a partial rewrite process in the packet.
At this time, in the rewriting in the packet, the sequence number seq is also rewritten (see FIG. 5).

That is, in FIG. 3, the first port 2
When a network stream is input from 6a and the second port 26b and output from the nth port 26n to the client 30, if the sequence of the port 26a and the port 26b is arranged alternately, the sequence operation is as shown in FIG. .

The transfer from the shared RAM 23 to an appropriate port is the same as in the case of the conventional EtherSW (FIG. 4).

In the case of having a routing function, a routing process can be performed after the frame rewriting process in the frame rewriting processing unit 24.

When the network device 20 performs the Proxy Server operation, the network device 20
When the flow control packet is received as a flow control packet from the network port 0 and is recognized as a flow control packet of the network stream, the stream flow processing unit 25 receives the flow control packet via the FIFO buffer of the reception port.
Is written to the shared RAM 23 via The stream flow processing unit 25 determines a server that should receive the flow control packet, and generates a flow control packet for each server. Since this process is less frequent than stream data, without using dedicated hardware,
It can also be performed using the CPU 21.

The client 30 is connected to the network device 2
The stream data transmitted from 0 is received and a desired reproduction is realized.

[0026]

As described above, according to the present invention, it is possible to provide a system having a high stream transmission capability using stream data transmitted from a plurality of stream sources. By omitting the operations between streams and the queuing process, it is possible to reduce the number of functions added to existing network devices. In addition, shared RA
The management of M can be simplified.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the overall configuration of a server system according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a hardware configuration example of a file server in the server system of FIG. 1;

FIG. 3 is a block diagram showing a configuration example of a network device in the server system of FIG. 1;

FIG. 4 is a block diagram showing a configuration example of a conventional EtherSW.

FIG. 5 is a diagram showing an example of correspondence between a sequence number of a packet transmitted from a server and received at a port of a network device and a sequence number transmitted to a client after a frame rewriting process.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 File server 101,102,103 File server 20 Network equipment 21 CPU 22 Shared bus 23 Shared RAM 24 Frame rewrite processing part 25 Stream flow processing part 26a-26n Port 30 Client 301,302,303 Client

Continued on the front page F term (reference) 5B089 GA12 GA18 GA19 GA31 HA06 JA11 JB04 JB05 KA06 KA11 KD01 5C064 BA01 BB07 BC16 BC20 BD07 BD08

Claims (2)

[Claims] 1. A plurality of file servers having a file system for distributing and storing a plurality of stream files generated by performing division processing in advance, and suitable for streams from a plurality of file servers in response to a request from a client. A network device that performs concatenation processing on a group of streams that have been previously divided and transmitted using the above protocol, and transmits the stream to the client as one stream. 2. The moving image server system according to claim 1, wherein a unit of connection processing of the stream group is a packet unit.