[go: up one dir, main page]

TWI883861B - Data server system - Google Patents

Data server system Download PDF

Info

Publication number
TWI883861B
TWI883861B TW113109245A TW113109245A TWI883861B TW I883861 B TWI883861 B TW I883861B TW 113109245 A TW113109245 A TW 113109245A TW 113109245 A TW113109245 A TW 113109245A TW I883861 B TWI883861 B TW I883861B
Authority
TW
Taiwan
Prior art keywords
transmission
server system
data server
chip
chips
Prior art date
Application number
TW113109245A
Other languages
Chinese (zh)
Other versions
TW202536584A (en
Inventor
許麗
陳超
楊傑翰
林宏州
陳俞帆
Original Assignee
英業達股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 英業達股份有限公司 filed Critical 英業達股份有限公司
Priority to TW113109245A priority Critical patent/TWI883861B/en
Application granted granted Critical
Publication of TWI883861B publication Critical patent/TWI883861B/en
Publication of TW202536584A publication Critical patent/TW202536584A/en

Links

Images

Landscapes

  • Small-Scale Networks (AREA)
  • Multi Processors (AREA)

Abstract

A data server system includes a central processing unit and a baseboard management controller disposed on a motherboard, an Ethernet switch, a plurality of processor nodes, two buses and first optical transceiver disposed on a substrate, and second and third optical transceivers and a physical layer chip disposed on an optical module board. The baseboard management controller is connected to the central processing unit. The Ethernet switch is connected to the central processing unit and baseboard management controller. The processor nodes is connected to the Ethernet switch and the central processing unit. The two buses are connected to the central processing unit and the processor nodes. The first optical transceiver is connected to the Ethernet switch. The second optical transceiver is connected to the first optical transceiver. The physical layer chip is connected to the second optical transceiver. The third optical transceiver is connected to the physical layer chip.

Description

資料伺服器系統Data Server System

本發明係關於一種資料伺服器系統。 The present invention relates to a data server system.

為因應目前網路短影音平台延伸之商業模式,主要提供影片轉碼及遊戲處理需求。商業模式之一在於提供高畫質影音轉碼營銷產品,透過短影音的平台放送,吸引客戶端購買產品興趣;商業模式之二在於提供在線遊戲實時處理,讓在線用戶享受網路遊戲平台服務。為了滿足上述平台的需求,需要提供一種能夠處理大量數據的伺服器系統。 In response to the current business model of the extension of online short video platforms, the main needs are video transcoding and game processing. One business model is to provide high-definition video transcoding marketing products, broadcasting through the short video platform to attract customers to purchase product interest; the second business model is to provide real-time processing of online games, allowing online users to enjoy online game platform services. In order to meet the needs of the above platforms, it is necessary to provide a server system that can process large amounts of data.

鑒於上述,本發明提供一種資料伺服器系統。 In view of the above, the present invention provides a data server system.

依據本發明一實施例的資料伺服器系統,包含設置在主板上的中央處理器及基板管理控制器,設置在基板上的乙太網路交換器、多個處理器節點、第一匯流排、第二匯流排及第一光收發器,以及設置在光模組板上的第二、第三光收發器及實體層晶片。所述基板管理控制器連接於所述中央處理器。所述乙太網路交換器連接於所述中央處理器及基板管理控制器。所述處理器節點連接於所述乙太網路交換器及中央處理器。所述第一及第二匯流排連接於所述中央處理器及處理器節點。所述第一光收發器連接於所述乙太網路交換器。所述第二光收發器 連接於所述第一光收發器。所述實體層晶片連接於所述第二光收發器。所述第三光收發器連接於所述實體層晶片。 According to an embodiment of the present invention, a data server system includes a central processor and a baseboard management controller arranged on a mainboard, an Ethernet switch, a plurality of processor nodes, a first bus, a second bus and a first optical transceiver arranged on a baseboard, and a second and third optical transceivers and a physical layer chip arranged on an optical module board. The baseboard management controller is connected to the central processor. The Ethernet switch is connected to the central processor and the baseboard management controller. The processor node is connected to the Ethernet switch and the central processor. The first and second buses are connected to the central processor and the processor node. The first optical transceiver is connected to the Ethernet switch. The second optical transceiver is connected to the first optical transceiver. The physical layer chip is connected to the second optical transceiver. The third optical transceiver is connected to the physical layer chip.

藉由上述結構,本案所揭示的資料伺服器系統,透過光收發器接收與傳輸大量資料,並以乙太網路交換器及第一及第二匯流排作為中央處理器及多個處理器節點之間資料傳輸介面,可提高伺服器的資料儲存密度及處理能力,讓用戶即時享受網路短影音平台及網路遊戲平台的大數據服務。 Through the above structure, the data server system disclosed in this case receives and transmits large amounts of data through optical transceivers, and uses Ethernet switches and the first and second buses as data transmission interfaces between the central processor and multiple processor nodes, which can improve the data storage density and processing capabilities of the server, allowing users to enjoy the big data services of online short video platforms and online game platforms in real time.

以上之關於本揭露內容之說明及以下之實施方式之說明係用以示範與解釋本發明之精神與原理,並且提供本發明之專利申請範圍更進一步之解釋。 The above description of the disclosed content and the following description of the implementation method are used to demonstrate and explain the spirit and principle of the present invention, and provide a further explanation of the scope of the patent application of the present invention.

1:資料伺服器系統 1:Data server system

11:主板 11: Motherboard

12:基板 12:Substrate

13:光模組板 13: Optical module board

101:中央處理器 101:Central Processing Unit

102:基板管理控制器 102: Baseboard management controller

103:乙太網路交換器 103:Ethernet switch

104,104-1至104-n:處理器節點 104,104-1 to 104-n: Processor nodes

105:第一匯流排 105: First bus

106:第二匯流排 106: Second bus

107:第一光收發器 107: First optical transceiver

108:第二光收發器 108: Second optical transceiver

109:實體層晶片 109: Physical layer chip

110:第三光收發器 110: The third optical transceiver

111:乙太網路收發器 111:Ethernet transceiver

112:USB 3.0控制器 112:USB 3.0 controller

113:USB集線器 113:USB hub

114:USB 2.0控制器 114:USB 2.0 controller

115:USB接口晶片 115: USB interface chip

116:UART控制器 116:UART controller

117:複雜可程式化邏輯裝置 117: Complex programmable logic device

1031:第一傳輸晶片 1031: First transmission chip

1032:第二傳輸晶片 1032: Second transmission chip

1033:第三傳輸晶片 1033: The third transmission chip

1034:第四傳輸晶片 1034: The fourth transmission chip

T1:第一端 T1: First end

T2:第二端 T2: Second end

T3:第三端 T3: The third terminal

T4:第四端 T4: The fourth end

N1:第一節點 N1: First node

N2:第二節點 N2: Second node

N3:第三節點 N3: The third node

N4:第四節點 N4: The fourth node

N5:第五節點 N5: Fifth Node

N6:第六節點 N6: Node 6

G1:第一群組 G1: Group 1

G2:第二群組 G2: The second group

G3:第三群組 G3: The third group

圖1係依據本發明一實施例所繪示的資料伺服器系統的方塊圖。 FIG1 is a block diagram of a data server system according to an embodiment of the present invention.

圖2係依據本發明一實施例所繪示的資料伺服器系統的乙太網路交換器的內部架構的示意圖。 FIG2 is a schematic diagram of the internal structure of an Ethernet switch of a data server system according to an embodiment of the present invention.

圖3示出本發明一實施例的資料伺服器系統的第一及第二傳輸晶片在不同模式下的訊號連接關係。 FIG3 shows the signal connection relationship between the first and second transmission chips of the data server system of an embodiment of the present invention in different modes.

圖4係依據本發明一實施例所繪示的資料伺服器系統的管理架構的示意圖。 FIG4 is a schematic diagram of the management architecture of a data server system according to an embodiment of the present invention.

圖5係依據本發明另一實施例所繪示的資料伺服器系統的USB拓樸設計的示意圖。 FIG5 is a schematic diagram of a USB topology design of a data server system according to another embodiment of the present invention.

圖6係依據本發明另一實施例所繪示的資料伺服器系統的UART拓樸設計的示意圖。 FIG6 is a schematic diagram of a UART topology design of a data server system according to another embodiment of the present invention.

以下在實施方式中詳細敘述本發明之詳細特徵以及優點,其內容足以使任何熟習相關技藝者了解本發明之技術內容並據以實施,且根據本說明書所揭露之內容、申請專利範圍及圖式,任何熟習相關技藝者可輕易地理解本發明相關之目的及優點。以下之實施例係進一步詳細說明本發明之觀點,但非以任何觀點限制本發明之範疇。 The detailed features and advantages of the present invention are described in detail in the following implementation method. The content is sufficient for anyone familiar with the relevant technology to understand the technical content of the present invention and implement it accordingly. According to the content disclosed in this specification, the scope of the patent application and the drawings, anyone familiar with the relevant technology can easily understand the relevant purposes and advantages of the present invention. The following embodiments are to further illustrate the viewpoints of the present invention, but do not limit the scope of the present invention by any viewpoint.

請參考圖1,圖1係依據本發明一實施例所繪示的資料伺服器系統的方塊圖。如圖1所示,資料伺服器系統1包含設置在主板11上的中央處理器101及基板管理控制器102,設置在基板12上的乙太網路交換器103、多個處理器節點104、第一匯流排105、第二匯流排106及第一光收發器107,以及設置在光模組板13上的第二光收發器108、實體層晶片109及第三光收發器110。基板管理控制器102連接於中央處理器101。乙太網路交換器103連接於中央處理器101及基板管理控制器102。處理器節點104連接於乙太網路交換器103及中央處理器101。第一及第二匯流排105、106連接於中央處理器101及處理器節點104。第一光收發器107連接於乙太網路交換器103。第二光收發器108連接於第一光收發器107。實體層晶片109連接於第二光收發器108。第三光收發器110連接於實體層晶片109。 Please refer to FIG. 1, which is a block diagram of a data server system according to an embodiment of the present invention. As shown in FIG. 1, the data server system 1 includes a central processing unit 101 and a baseboard management controller 102 disposed on a mainboard 11, an Ethernet switch 103, a plurality of processor nodes 104, a first bus 105, a second bus 106 and a first optical transceiver 107 disposed on a substrate 12, and a second optical transceiver 108, a physical layer chip 109 and a third optical transceiver 110 disposed on an optical module board 13. The baseboard management controller 102 is connected to the central processing unit 101. The Ethernet switch 103 is connected to the central processing unit 101 and the baseboard management controller 102. The processor node 104 is connected to the Ethernet switch 103 and the central processing unit 101. The first and second buses 105 and 106 are connected to the central processor 101 and the processor node 104. The first optical transceiver 107 is connected to the Ethernet switch 103. The second optical transceiver 108 is connected to the first optical transceiver 107. The physical layer chip 109 is connected to the second optical transceiver 108. The third optical transceiver 110 is connected to the physical layer chip 109.

舉例而言,中央處理器101可例如為intel公司的D2775晶片,或者,資料伺服器系統1也可採用其他可處理大數據的高效能處理 器,本案不限於此。基板管理控制器102可例如為ASPEED公司的AST2600晶片,用於執行伺服器管理的處理流程,或者,本案的資料伺服器系統1也可採用其他的基板管理控制器(Board Management Controller,BMC)。中央處理器101與基板管理控制器102之間可透過多種連接器規格彼此連接,如PCIe、USB、UART、NCSI等。乙太網路交換器103(Ethernet switch)可包含多個資料傳輸晶片,用於作為主板11、基板12及光模組板13之間主要的資料傳輸介面。乙太網路交換器103與中央處理器101之間可透過多種連接器規格彼此連接,如PCIe、CAUI等。乙太網路交換器103與多個處理器節點104-1至104-n的每一者之間可透過2.5GBaseT的連接器規格彼此連接。乙太網路交換器103與第一光收發器107之間可透過100G CAUI的連接器規格彼此連接。 For example, the central processor 101 may be, for example, a D2775 chip of Intel Corporation, or the data server system 1 may also use other high-performance processors that can process large data, but the present invention is not limited thereto. The baseboard management controller 102 may be, for example, an AST2600 chip of ASPEED Corporation, which is used to execute the processing flow of server management, or the data server system 1 of the present invention may also use other baseboard management controllers (BMC). The central processor 101 and the baseboard management controller 102 may be connected to each other through a variety of connector specifications, such as PCIe, USB, UART, NCSI, etc. The Ethernet switch 103 may include a plurality of data transmission chips, which are used as the main data transmission interface between the mainboard 11, the baseboard 12 and the optical module board 13. The Ethernet switch 103 and the central processor 101 can be connected to each other through a variety of connector specifications, such as PCIe, CAUI, etc. The Ethernet switch 103 and each of the multiple processor nodes 104-1 to 104-n can be connected to each other through a 2.5GBaseT connector specification. The Ethernet switch 103 and the first optical transceiver 107 can be connected to each other through a 100G CAUI connector specification.

多個處理器節點104的每一者可為一個處理單元,例如Qualcomm公司的QCS8550晶片,或者,資料伺服器系統1也可採用其他可對大數據進行運算的高效能處理器,本案不限於此。在一實施例中,多個處理器節點104-1至104-n的數量可為240個,且這些處理器節點104-1至104-n的每四者可設置於一運算卡(POD card)上,且每一運算卡(共60張)可支持熱插拔功能。具體而言,每一張運算卡可安裝於運算卡底座的一插槽中,且透過金手指機構設計及相關線路保護機制,可以實現每一張運算卡的熱插拔功能,支持線上熱運行模式。 Each of the multiple processor nodes 104 can be a processing unit, such as Qualcomm's QCS8550 chip, or the data server system 1 can also use other high-performance processors that can perform operations on large data, but the present case is not limited thereto. In one embodiment, the number of the multiple processor nodes 104-1 to 104-n can be 240, and every four of these processor nodes 104-1 to 104-n can be set on a computing card (POD card), and each computing card (60 in total) can support hot-swap function. Specifically, each computing card can be installed in a slot of the computing card base, and through the gold finger mechanism design and related line protection mechanism, the hot-swap function of each computing card can be realized, supporting online hot operation mode.

第一匯流排105及第二匯流排106可例如為基於PCIe連接器規格的PI7C9X2G608GP交換器。第一匯流排105可透過PCIe轉UART的轉接器連接於每一個處理器節點104。第二匯流排106可透過 PCIe轉USB 3.0的轉接器連接於每一個處理器節點104。第一及第二光收發器107、108可例如為四通道小封裝可插拔(Quad Small Form-factor Pluggable,QSFP)的收發器,並支持100G的資料傳輸速率。第三光收發器110可包含雙通道小封裝可插拔(dual Small Form-factor Pluggable,DSFP)及四通道小封裝可插拔(QSFP)的收發器,並支持100G的資料傳輸速率。特別來說,一組光收發器可包含兩個光收發元件。實體層晶片109可例如為Broadcom公司的BCM81358晶片,其屬於一種低功耗元件,且具有重計時(retime)及均衡(equalize)之功能。 The first bus 105 and the second bus 106 may be, for example, PI7C9X2G608GP switches based on the PCIe connector specification. The first bus 105 may be connected to each processor node 104 via a PCIe to UART adapter. The second bus 106 may be connected to each processor node 104 via a PCIe to USB 3.0 adapter. The first and second optical transceivers 107 and 108 may be, for example, Quad Small Form-factor Pluggable (QSFP) transceivers, and support a data transmission rate of 100G. The third optical transceiver 110 may include a dual Small Form-factor Pluggable (DSFP) and a quad Small Form-factor Pluggable (QSFP) transceiver, and support a 100G data transmission rate. In particular, a set of optical transceivers may include two optical transceiver components. The physical layer chip 109 may be, for example, a BCM81358 chip from Broadcom, which is a low-power component and has the functions of retime and equalize.

透過上述配置,240個處理器節點可透過2.5G Base-T的網路接口進行資料傳輸,並將這240路2.5G Base-T通過兩級乙太網路交換器匯聚為2個100G上行光傳輸接口;支持兩路UART至各處理器節點,主UART採用USB轉UART,備用UART採用PCIe轉UART並經過複雜可程式邏輯元件多工器(CPLD MUX)切換,後者為備份設計方案,提高系統管理的可靠性;支持USB3.0至各處理器節點,以實現各個處理器節點的韌體升級。 Through the above configuration, 240 processor nodes can transmit data through the 2.5G Base-T network interface, and these 240 2.5G Base-Ts are aggregated into two 100G uplink optical transmission interfaces through a two-level Ethernet switch; it supports two UARTs to each processor node, the main UART uses USB to UART, and the backup UART uses PCIe to UART and switches through a complex programmable logic device multiplexer (CPLD MUX). The latter is a backup design solution to improve the reliability of system management; it supports USB3.0 to each processor node to realize the firmware upgrade of each processor node.

請結合圖1參考圖2,圖2係依據本發明一實施例所繪示的資料伺服器系統的乙太網路交換器的內部架構的示意圖。如圖2所示,本例的乙太網路交換器103可包含二第一傳輸晶片1031、一第二傳輸晶片1032、多個第三傳輸晶片1033及多個第四傳輸晶片1034。所述二第一傳輸晶片1031可透過第一端T1連接於中央處理器101,並透過第二端T2連接於第一光收發器107。第二傳輸晶片1032連接於所述二第一傳輸晶片1031。多個第三傳輸晶片1033連接於第二傳輸晶片1032。多個第四傳輸 晶片1034連接於多個第三傳輸晶片1033,並透過第三端T3連接於多個處理器節點104。 Please refer to FIG. 2 in conjunction with FIG. 1. FIG. 2 is a schematic diagram of the internal structure of an Ethernet switch of a data server system according to an embodiment of the present invention. As shown in FIG. 2, the Ethernet switch 103 of this example may include two first transmission chips 1031, a second transmission chip 1032, a plurality of third transmission chips 1033, and a plurality of fourth transmission chips 1034. The two first transmission chips 1031 may be connected to the central processor 101 through the first end T1, and connected to the first optical transceiver 107 through the second end T2. The second transmission chip 1032 is connected to the two first transmission chips 1031. A plurality of third transmission chips 1033 are connected to the second transmission chip 1032. Multiple fourth transmission chips 1034 are connected to multiple third transmission chips 1033, and are connected to multiple processor nodes 104 through the third terminal T3.

在本例中,第一傳輸晶片1031可為Broadcom公司的BCM81381晶片,第二傳輸晶片1032可為Broadcom公司的BCM56771晶片,第三傳輸晶片1033可為Broadcom公司的BCM56072晶片,第四傳輸晶片1034可為Broadcom公司的BCM54908E晶片。第一傳輸晶片1031與第二傳輸晶片1032之間可透過CAUI的連接器規格連接,且支持100G的資料傳輸速率。第二傳輸晶片1032與第三傳輸晶片1033之間可透過Base-KR4的連接器規格連接,且支持100G的資料傳輸速率。第三傳輸晶片1033與第四傳輸晶片1034之間可透過QXGMII的連接器規格連接,且支持100G的資料傳輸速率。進一步,本例的第三傳輸晶片1033的數量可為5個,第四傳輸晶片的數量可為30個,處理器節點104的數量可為240個。即,第二傳輸晶片1032可連接於5個第三傳輸晶片1033,每個第三傳輸晶片1033可連接於6個第四傳輸晶片1034,每個第四傳輸晶片1034可連接於8個處理器節點104。進一步,本例的資料伺服器系統可包含多個轉接元件。這些轉接元件透過Base-T連接埠連接於所述多個處理器節點104,且透過PCIe連接埠連接於乙太網路交換器103的第四傳輸晶片1034。 In this example, the first transmission chip 1031 may be a BCM81381 chip of Broadcom, the second transmission chip 1032 may be a BCM56771 chip of Broadcom, the third transmission chip 1033 may be a BCM56072 chip of Broadcom, and the fourth transmission chip 1034 may be a BCM54908E chip of Broadcom. The first transmission chip 1031 and the second transmission chip 1032 may be connected via a CAUI connector specification and support a data transmission rate of 100G. The second transmission chip 1032 and the third transmission chip 1033 may be connected via a Base-KR4 connector specification and support a data transmission rate of 100G. The third transmission chip 1033 and the fourth transmission chip 1034 may be connected via a QXGMII connector specification and support a data transmission rate of 100G. Furthermore, the number of third transmission chips 1033 in this example can be 5, the number of fourth transmission chips can be 30, and the number of processor nodes 104 can be 240. That is, the second transmission chip 1032 can be connected to 5 third transmission chips 1033, each third transmission chip 1033 can be connected to 6 fourth transmission chips 1034, and each fourth transmission chip 1034 can be connected to 8 processor nodes 104. Furthermore, the data server system in this example can include multiple switching components. These switching components are connected to the multiple processor nodes 104 through the Base-T connection port, and are connected to the fourth transmission chip 1034 of the Ethernet switch 103 through the PCIe connection port.

各處理器節點的PCIe透過Intel的乙太網路控制器I226轉為2.5GBase-T進入系統的乙太網鏈路;系統使用30個BCM54908E將每8個2.5GBase-T轉為2路10G QXGMII至5個BCM56072轉為5路100G Base-KR,再透過BCM56771彙聚為2個100G上行口;系統的2個 100G上行口透過主動式光纜連接至光模組板,透過BCM81358做重計時/適配(Retimer/Gearbox)以支持QSFP或DSFP;使用2個BCM81381執行一般模式(Normal Mode)實現100G雙上聯;使用2個BCM81381執行多工模式(MUX mode)實現雙口預啟動執行環境(Preboot eXecution Environment,PXE)。 The PCIe of each processor node is converted to 2.5GBase-T through Intel's Ethernet controller I226 and enters the system's Ethernet link; the system uses 30 BCM54908E to convert each 8 2.5GBase-T to 2 10G QXGMII, which are then converted to 5 100G Base-KR by 5 BCM56072, and then aggregated into 2 100G uplink ports through BCM56771; the system's 2 100G uplink ports are connected to the optical module board through active optical cables, and retimer/gearbox is used by BCM81358 to support QSFP or DSFP; 2 BCM81381 are used to run the normal mode (Normal Mode) to achieve 100G dual uplink; use two BCM81381s to run multiplexing mode (MUX mode) to achieve a dual-port preboot execution environment (Preboot eXecution Environment, PXE).

請結合圖1、圖2參考圖3,圖3示出本發明一實施例的資料伺服器系統的第一及第二傳輸晶片在不同模式下的訊號連接關係。如圖3所示,二第一傳輸晶片1031支持100G的資料傳輸速度,且用於受到基板管理控制器102的控制被切換為一第一模式或一第二模式。在第一模式下,中央處理器101可直接透過二第一傳輸晶片1031的第一節點N1及第二節點N2接收來自第一光收發器107的訊號。在第二模式下,中央處理器101透過二第一傳輸晶片1031的第一節點N1及第三節點N3和第二傳輸晶片1032的第五節點N5接收來自第一光收發器107的訊號。也就是說,在第一模式下,從第三光收發器110接收的訊號經過實體層晶片109、第二光收發器108傳輸至第一光收發器107,並經過第一傳輸晶片1031的第二節點N2及第一節點N1傳輸至中央處理器101。在第二模式下,從第三光收發器110接收的訊號經過實體層晶片109、第二光收發器108傳輸至第一光收發器107,並透過第一傳輸晶片1031的第二節點N2及第四節點N4傳輸至第二傳輸晶片1032(透過第六節點N6),再透過第二傳輸晶片1032的第五節點N5及第一傳輸晶片的第三節點N3及第一節點N1傳輸至中央處理器101。所述第一模式對應於上述的一般模式,第二模式對應於上述的多工模式。 Please refer to FIG. 3 in conjunction with FIG. 1 and FIG. 2. FIG. 3 shows the signal connection relationship between the first and second transmission chips of the data server system of an embodiment of the present invention in different modes. As shown in FIG. 3, the two first transmission chips 1031 support a data transmission speed of 100G and are used to be switched to a first mode or a second mode under the control of the baseboard management controller 102. In the first mode, the central processor 101 can directly receive the signal from the first optical transceiver 107 through the first node N1 and the second node N2 of the two first transmission chips 1031. In the second mode, the central processor 101 receives the signal from the first optical transceiver 107 through the first node N1 and the third node N3 of the two first transmission chips 1031 and the fifth node N5 of the second transmission chip 1032. That is, in the first mode, the signal received from the third optical transceiver 110 is transmitted to the first optical transceiver 107 through the physical layer chip 109 and the second optical transceiver 108, and is transmitted to the central processor 101 through the second node N2 and the first node N1 of the first transmission chip 1031. In the second mode, the signal received from the third optical transceiver 110 is transmitted to the first optical transceiver 107 through the physical layer chip 109 and the second optical transceiver 108, and is transmitted to the second transmission chip 1032 (through the sixth node N6) through the second node N2 and the fourth node N4 of the first transmission chip 1031, and then transmitted to the central processor 101 through the fifth node N5 of the second transmission chip 1032 and the third node N3 and the first node N1 of the first transmission chip. The first mode corresponds to the above-mentioned general mode, and the second mode corresponds to the above-mentioned multiplexing mode.

請結合圖1、圖2參考圖4,圖4係依據本發明一實施例所繪示的資料伺服器系統的管理架構的示意圖。如圖4所示,中央處理器101可用於管理第二傳輸晶片1032及多個第三傳輸晶片1033,且每個第三傳輸晶片1033可用於管理多個第四傳輸晶片1034。具體而言,中央處理器101可透過4個PCIeGen3管理第二傳輸晶片1032,並透過5個PCIeGen3分別管理5個第二傳輸晶片1032。基板管理控制器102可連接於一乙太網路收發器111,並用於管理實體層晶片109及兩個第一傳輸晶片1031。透過此配置,基板管理控制器102可透過MDC/MDIO接口完成2個BCM81381和BCM81358的初始化;中央處理器101透過PCIe接口來管理BCM56771和5個BCM56072;每個BCM56072透過2個MDC/MDIO接口管理6個BCM54908E。 Please refer to FIG. 4 in conjunction with FIG. 1 and FIG. 2. FIG. 4 is a schematic diagram of a management architecture of a data server system according to an embodiment of the present invention. As shown in FIG. 4, the central processor 101 can be used to manage the second transmission chip 1032 and multiple third transmission chips 1033, and each third transmission chip 1033 can be used to manage multiple fourth transmission chips 1034. Specifically, the central processor 101 can manage the second transmission chip 1032 through 4 PCIeGen3s, and manage 5 second transmission chips 1032 through 5 PCIeGen3s respectively. The baseboard management controller 102 can be connected to an Ethernet transceiver 111, and is used to manage the physical layer chip 109 and the two first transmission chips 1031. With this configuration, the baseboard management controller 102 can complete the initialization of two BCM81381s and BCM81358s through the MDC/MDIO interface; the central processor 101 manages the BCM56771 and five BCM56072s through the PCIe interface; and each BCM56072 manages six BCM54908Es through two MDC/MDIO interfaces.

請結合圖1參考圖5,圖5係依據本發明另一實施例所繪示的資料伺服器系統的USB拓樸設計的示意圖。本例的資料伺服器系統可更包含多個通用序列匯流排,連接於第二匯流排106與處理器節點104之間。如圖5所示,中央處理器101可透過1個PCIe2.0×4接口連接於第二匯流排106。第二匯流排106可透過4個PCIe2.0×1接口連接於4個第一群組G1的USB 3.0控制器112。中央處理器101可透過8個PCIe2.0×1接口連接於8個第一群組G1的USB 3.0控制器112。也就是說,中央處理器101可一共連接於12個第一群組G1的USB 3.0控制器112。中央處理器101還可透過1個PCIe2.0×1接口連接於1個第二群組G2的USB 3.0控制器112。在每個第一群組G1中,USB 3.0控制器112可分別透過3個USB 3.0接口連接於3個USB集線器113,3個USB集線器113再分別透過6個USB 3.0接口連接於一共18個處理器節點104。每個第一群組G1中,USB 3.0控制器112還可透過1個USB 3.0接口連接於1個處理器節點104。也就是說,在第一群組G1中,每個USB 3.0控制器112可連接於19個處理器節點104。 Please refer to FIG. 5 in conjunction with FIG. 1 , which is a schematic diagram of a USB topology design of a data server system according to another embodiment of the present invention. The data server system of this example may further include a plurality of universal serial buses connected between the second bus 106 and the processor node 104. As shown in FIG. 5 , the central processing unit 101 may be connected to the second bus 106 via a PCIe2.0×4 interface. The second bus 106 may be connected to the four USB 3.0 controllers 112 of the first group G1 via four PCIe2.0×1 interfaces. The central processing unit 101 may be connected to the eight USB 3.0 controllers 112 of the first group G1 via eight PCIe2.0×1 interfaces. That is, the central processor 101 can be connected to a total of 12 USB 3.0 controllers 112 of the first group G1. The central processor 101 can also be connected to a USB 3.0 controller 112 of the second group G2 through a PCIe2.0×1 interface. In each first group G1, the USB 3.0 controller 112 can be connected to three USB hubs 113 through three USB 3.0 interfaces, and the three USB hubs 113 are connected to a total of 18 processor nodes 104 through six USB 3.0 interfaces. In each first group G1, the USB 3.0 controller 112 can also be connected to one processor node 104 through one USB 3.0 interface. That is, in the first group G1, each USB 3.0 controller 112 can be connected to 19 processor nodes 104.

在第二群組G2中,USB 3.0控制器112可透過2個USB 3.0接口連接於2個USB集線器113,2個USB集線器113再分別透過6個USB 3.0接口連接於一共12個處理器節點104。也就是說,在第二群組G2中,每個USB 3.0控制器112可連接於12個處理器節點104。在此架構下,中央處理器101可透過第一群組G1及第二群組G2的USB 3.0控制器112以通用序列匯流排連接至12×19+12=240個處理器節點104。另外,中央處理器101還可透過USB連接埠連接於基板管理控制器102,且具有備用的USB連接埠以供其他裝置進行連接。 In the second group G2, the USB 3.0 controller 112 can be connected to two USB hubs 113 through two USB 3.0 interfaces, and the two USB hubs 113 are connected to a total of 12 processor nodes 104 through six USB 3.0 interfaces. In other words, in the second group G2, each USB 3.0 controller 112 can be connected to 12 processor nodes 104. Under this architecture, the central processor 101 can be connected to 12×19+12=240 processor nodes 104 through the USB 3.0 controllers 112 of the first group G1 and the second group G2 via a universal serial bus. In addition, the central processor 101 can also be connected to the baseboard management controller 102 through a USB port, and has a spare USB port for other devices to connect.

請參考圖6,圖6係依據本發明另一實施例所繪示的資料伺服器系統的UART拓樸設計的示意圖。本例的資料伺服器系統可更包含多個通用非同步收發傳輸器,連接於第一匯流排105與多個處理器節點104之間,這些通用非同步收發傳輸器包含多個主要傳輸器及多個備用傳輸器。如圖6所示,中央處理器101可透過1個PCIe2.0×1接口連接於第一匯流排105。第一匯流排105可透過4個PCIe2.0×1接口連接於4個第三群組G3的USB 3.0控制器112。也就是說,中央處理器101可一共連接於4個第三群組G3的USB 3.0控制器112。在每個第三群組G3中,USB 3.0控制器112可分別透過2個USB 2.0接口連接於2個USB 2.0控制器114,4個USB 2.0控制器114再分別透過7個USB 2.0接口連接於7個 USB接口晶片115。每個USB接口晶片115可透過UART連接埠連接於4個處理器節點104。在每個第三群組G3中,USB 3.0控制器112還可透過1個USB 2.0接口連接於1個USB接口晶片115以透過UART連接埠連接至4個處理器節點104。也就是說,在第三群組G3中,每個USB 3.0控制器112可連接於4×7×2+4=60個處理器節點104。另外,第一匯流排105還可透過1個PCIe1.0×1接口連接於UART控制器116。UART控制器116可透過8個UART連接埠連接至複雜可程式邏輯裝置(Complex Programmable Logic Device,CPLD)117。複雜可程式邏輯裝置117可透過240個UART連接埠連接至240個處理器節點104,以作為備用傳輸器。 Please refer to FIG. 6, which is a schematic diagram of a UART topology design of a data server system according to another embodiment of the present invention. The data server system of this example may further include a plurality of universal asynchronous transceivers connected between the first bus 105 and the plurality of processor nodes 104, and these universal asynchronous transceivers include a plurality of main transmitters and a plurality of backup transmitters. As shown in FIG. 6, the central processor 101 may be connected to the first bus 105 via a PCIe2.0×1 interface. The first bus 105 may be connected to the four USB 3.0 controllers 112 of the third group G3 via four PCIe2.0×1 interfaces. In other words, the central processor 101 may be connected to a total of four USB 3.0 controllers 112 of the third group G3. In each third group G3, the USB 3.0 controller 112 can be connected to two USB 2.0 controllers 114 through two USB 2.0 interfaces respectively, and the four USB 2.0 controllers 114 can be connected to seven USB interface chips 115 through seven USB 2.0 interfaces respectively. Each USB interface chip 115 can be connected to four processor nodes 104 through a UART port. In each third group G3, the USB 3.0 controller 112 can also be connected to one USB interface chip 115 through one USB 2.0 interface to connect to four processor nodes 104 through a UART port. That is, in the third group G3, each USB 3.0 controller 112 can be connected to 4×7×2+4=60 processor nodes 104. In addition, the first bus 105 can also be connected to the UART controller 116 through a PCIe1.0×1 interface. The UART controller 116 can be connected to the complex programmable logic device (CPLD) 117 through 8 UART ports. The complex programmable logic device 117 can be connected to 240 processor nodes 104 through 240 UART ports as a backup transmitter.

舉例而言,上述實施例的USB 3.0控制器可採用Renesas公司的UPD720201晶片實現。USB集線器113可採用Microchip公司的USB5806晶片實現。USB 2.0控制器114可採用Microchip公司的USB2517晶片實現。USB接口晶片115可採用FTDI公司的FT4232晶片實現。UART控制器116可採用MAXLinear公司的XR17V358晶片實現。然本案不限於選用上述元件。透過本實施例的配置,資料伺服器系統可支持2個前置USB3.0接口,CPU與BMC之間預留USB2.0接口用於通訊。本系統提供了240路USB3.0以實現對各節點中的QCS8550透過USB3.0進行韌體升級的功能。本方案中分配了CPU的PCIe口並通過PCIe Switch PI7C9X2G608GP擴展為13路PCIe2.0,再由USB控制器uPD720201搭配USB集線器擴展出240路USB3.0。其中受限於uPD720201支持的插槽及端點數量,一個uPD720201可擴展出19個 USB3.0,因此本方案中使用了12組1×uPD720201+3×USB5806的組合方案,外加1組1×uPD720201+2×USB5806的組合方案,實現了本系統中240路USB3.0的需求。 For example, the USB 3.0 controller of the above embodiment can be implemented using the UPD720201 chip of Renesas. The USB hub 113 can be implemented using the USB5806 chip of Microchip. The USB 2.0 controller 114 can be implemented using the USB2517 chip of Microchip. The USB interface chip 115 can be implemented using the FT4232 chip of FTDI. The UART controller 116 can be implemented using the XR17V358 chip of MAXLinear. However, the present invention is not limited to the above components. Through the configuration of this embodiment, the data server system can support 2 front USB3.0 interfaces, and a USB2.0 interface is reserved between the CPU and the BMC for communication. The system provides 240 USB3.0 channels to realize the function of upgrading the firmware of the QCS8550 in each node through USB3.0. In this solution, the CPU PCIe port is allocated and expanded to 13 PCIe2.0 through PCIe Switch PI7C9X2G608GP, and then expanded to 240 USB3.0 through USB controller uPD720201 and USB hub. Limited by the number of slots and endpoints supported by uPD720201, one uPD720201 can expand to 19 USB3.0, so this solution uses 12 sets of 1×uPD720201+3×USB5806 combination solutions, plus 1 set of 1×uPD720201+2×USB5806 combination solutions, to achieve the 240-channel USB3.0 requirement in this system.

透過本實施例的配置,本系統支持兩路UART至各節點,一路為主要UART,一路為備用UART。在本方案中分配CPU的1路X1PCIe2.0,通過PCIe交換器PI7C9X2G608GP擴展為5路PCIe,其中4路通過PCIe轉USB、USB轉UART的方案來擴展出240路主要UART,實現方案為:4組1×uPD720201+2×USB2517+7×FT4232;另一路PCIe透過XR17V358轉為8路URAT,再由CPLD實現240路備用UART的擴展及切換。 Through the configuration of this embodiment, this system supports two UARTs to each node, one for the main UART and one for the backup UART. In this solution, the CPU's 1-way X1PCIe2.0 is allocated and expanded to 5-way PCIe through the PCIe switch PI7C9X2G608GP, of which 4-way is expanded to 240-way main UART through the PCIe to USB and USB to UART solution. The implementation solution is: 4 sets of 1×uPD720201+2×USB2517+7×FT4232; the other PCIe is converted to 8-way URAT through XR17V358, and then the CPLD realizes the expansion and switching of 240-way backup UART.

關於PCIe的分配請參考以下描述,網路交換器BCM56771佔用一個PCIe3.0x4,5個BCM56072佔用5個PCIe3.0×1;UART:PCIe2.0x1透過交換器擴展4個PCIe2.0×1接4個USB控制器供USB-UART轉換晶片使用+擴展1路PCIe2.0×1接PCIe-UART控制器;USB:PCIe2.0×4透過交換器擴展4個PCIe2.0×1接4個USB控制器+HSIO引出9個PCIe2.0×1接9個USB控制器,共計13個;BMC:PCIe2.0×1;2個USB3.0×1;2個M.2 SATA。 For PCIe allocation, please refer to the following description. The network switch BCM56771 occupies one PCIe3.0x4, and five BCM56072s occupy five PCIe3.0×1s. UART: PCIe2.0x1 is expanded through the switch to four PCIe2.0×1s connected to four USB controllers for USB-UART conversion chips + one PCIe2.0×1 is expanded to connect to PCIe-UART controllers. USB: PCIe2.0×4 is expanded through the switch to four PCIe2.0×1s connected to four USB controllers + HSIO leads to nine PCIe2.0×1s connected to nine USB controllers, for a total of 13. BMC: PCIe2.0×1; two USB3.0×1s; two M.2 SATAs.

本系統亦支持4個電源供應器(2+2)之架構,系統電力分配板(Power Distribution Board,PDB)為各板卡供電。PDB通過4個2×12電力連接器給基板供電;PDB通過1個2×2電力連接器給光模組板供電;基板通過1個2X2電力連接器給主板供電。 This system also supports a 4-power supply (2+2) architecture, and the system power distribution board (PDB) supplies power to each board. The PDB supplies power to the baseboard through 4 2×12 power connectors; the PDB supplies power to the optical module board through 1 2×2 power connector; the baseboard supplies power to the mainboard through 1 2X2 power connector.

藉由上述結構,本案所揭示的資料伺服器系統,透過光收發器接收與傳輸大量資料,並以乙太網路交換器及第一及第二匯流排作為中央處理器及多個處理器節點之間資料傳輸介面,可提高伺服器的資料儲存密度及處理能力,讓用戶即時享受網路短影音平台及網路遊戲平台的大數據服務。另外,本系統支持240個節點的PCIe轉為2.5G Base-T,並通過系統的網絡彙聚為2個100G上行口,實現影音轉碼及遊戲數據的傳輸;系統支持240路主要UART用於個節點控制管理,同時支持240路備用UART作為冗餘設計,可提高系統管理的可靠性;系統支持前置100G雙上聯功能;系統支持PXE功能,通過BMC切換,支持透通及連通交換器晶片轉傳模式;每張POD上搭載4個節點,可節約壞卡維護成本,且支持熱插拔,透過金手指機構設計,及相關線路保護機制,支持在線熱運行模式。 Through the above structure, the data server system disclosed in this case receives and transmits large amounts of data through optical transceivers, and uses an Ethernet switch and the first and second buses as data transmission interfaces between a central processor and multiple processor nodes, thereby improving the data storage density and processing capabilities of the server, allowing users to enjoy the big data services of online short video platforms and online game platforms in real time. In addition, this system supports 240 nodes of PCIe to 2.5G Base-T, and aggregates them into 2 100G uplink ports through the system network to achieve video and audio transcoding and game data transmission; the system supports 240 main UARTs for individual node control and management, and supports 240 backup UARTs as a redundant design to improve the reliability of system management; the system supports front 100G dual uplink function; the system supports PXE function, through BMC switching, supports transparent and connected switch chip transfer mode; each POD is equipped with 4 nodes, which can save the maintenance cost of bad cards, and supports hot plugging, through the gold finger mechanism design, and related line protection mechanism, supports online hot operation mode.

雖然本發明以前述之實施例揭露如上,然其並非用以限定本發明。在不脫離本發明之精神和範圍內,所為之更動與潤飾,均屬本發明之專利保護範圍。關於本發明所界定之保護範圍請參考所附之申請專利範圍。 Although the present invention is disclosed as above by the aforementioned embodiments, it is not intended to limit the present invention. Any changes and modifications made within the spirit and scope of the present invention are within the scope of patent protection of the present invention. Please refer to the attached patent application for the scope of protection defined by the present invention.

1:資料伺服器系統 1:Data server system

11:主板 11: Motherboard

12:基板 12:Substrate

13:光模組板 13: Optical module board

101:中央處理器 101:Central Processing Unit

102:基板管理控制器 102: Baseboard management controller

103:乙太網路交換器 103:Ethernet switch

104,104-1至104-n:處理器節點 104,104-1 to 104-n: Processor nodes

105:第一匯流排 105: First bus

106:第二匯流排 106: Second bus

107:第一光收發器 107: First optical transceiver

108:第二光收發器 108: Second optical transceiver

109:實體層晶片 109: Physical layer chip

110:第三光收發器 110: The third optical transceiver

Claims (10)

一種資料伺服器系統,包含: 一主板; 一中央處理器,設置於該主板上; 一基板管理控制器,設置於該主板上,連接於該中央處理器; 一基板; 一乙太網路交換器,設置於該基板上,連接於該中央處理器及該基板管理控制器; 多個處理器節點,設置於該基板上,連接於該乙太網路交換器,且連接於該中央處理器; 一第一匯流排及一第二匯流排,設置於該基板上,連接於該中央處理器及該些處理器節點; 一組第一光收發器,設置於該基板上,連接於該乙太網路交換器; 一光模組板; 一組第二光收發器,設置於該光模組板上,連接於該組第一光收發器; 一實體層晶片,設置於該光模組板上,連接於該組第二光收發器;以及 一組第三光收發器,設置於該光模組板上,連接於該實體層晶片。 A data server system includes: a mainboard; a central processing unit, arranged on the mainboard; a baseboard management controller, arranged on the mainboard and connected to the central processing unit; a baseboard; an Ethernet switch, arranged on the baseboard and connected to the central processing unit and the baseboard management controller; a plurality of processor nodes, arranged on the baseboard, connected to the Ethernet switch and connected to the central processing unit; a first bus and a second bus, arranged on the baseboard and connected to the central processing unit and the processor nodes; a set of first optical transceivers, arranged on the baseboard and connected to the Ethernet switch; an optical module board; a set of second optical transceivers, arranged on the optical module board and connected to the set of first optical transceivers; A physical layer chip is disposed on the optical module board and connected to the second optical transceiver group; and a third optical transceiver group is disposed on the optical module board and connected to the physical layer chip. 如請求項1所述的資料伺服器系統,其中該乙太網路交換器包含二第一傳輸晶片、一第二傳輸晶片、多個第三傳輸晶片及多個第四傳輸晶片,該二第一傳輸晶片連接於該組第一光收發器,該第二傳輸晶片連接於該二第一傳輸晶片,該些第三傳輸晶片連接於該第二傳輸晶片,該些第四傳輸晶片連接於該些第三傳輸晶片及該些處理器節點。A data server system as described in claim 1, wherein the Ethernet switch includes two first transmission chips, a second transmission chip, multiple third transmission chips and multiple fourth transmission chips, the two first transmission chips are connected to the group of first optical transceivers, the second transmission chip is connected to the two first transmission chips, the third transmission chips are connected to the second transmission chip, and the fourth transmission chips are connected to the third transmission chips and the processor nodes. 如請求項2所述的資料伺服器系統,其中該些處理器節點的數量為240個,該些第三傳輸晶片的數量為5個,該些第四傳輸晶片的數量為30個。A data server system as described in claim 2, wherein the number of the processor nodes is 240, the number of the third transmission chips is 5, and the number of the fourth transmission chips is 30. 如請求項2所述的資料伺服器系統,其中該二第一傳輸晶片支持100G的資料傳輸速度,且用於受到該基板管理控制器的控制被切換為一第一模式或一第二模式,在該第一模式下該中央處理器直接透過該二第一傳輸晶片接收來自該組第一光收發器的訊號,在該第二模式下該中央處理器透過該二第一傳輸晶片及該第二傳輸晶片接收來自該組第一光收發器的訊號。A data server system as described in claim 2, wherein the two first transmission chips support a data transmission speed of 100G and are used to be switched to a first mode or a second mode under the control of the baseboard management controller, in which the central processing unit directly receives signals from the group of first optical transceivers through the two first transmission chips, and in which the central processing unit receives signals from the group of first optical transceivers through the two first transmission chips and the second transmission chip. 如請求項2所述的資料伺服器系統,其中該中央處理器用於管理該第二傳輸晶片及該些第三傳輸晶片,該些第三傳輸晶片用於管理該些第四傳輸晶片。A data server system as described in claim 2, wherein the central processor is used to manage the second transmission chip and the third transmission chips, and the third transmission chips are used to manage the fourth transmission chips. 如請求項1所述的資料伺服器系統,更包含多個轉接元件,該些轉接元件透過Base-T連接埠連接於該些處理器節點,且透過PCIe連接埠連接於該網路交換器。The data server system as described in claim 1 further includes multiple adapter components, which are connected to the processor nodes via Base-T ports and connected to the network switch via PCIe ports. 如請求項1所述的資料伺服器系統,更包含多個通用序列匯流排,連接於該第二匯流排與該些處理器節點之間。The data server system as described in claim 1 further includes a plurality of universal serial buses connected between the second bus and the processor nodes. 如請求項1所述的資料伺服器系統,更包含多個通用非同步收發傳輸器,連接於該第一匯流排與該些處理器節點之間,該些通用非同步收發傳輸器包含多個主要傳輸器及多個備用傳輸器。The data server system as described in claim 1 further includes a plurality of universal asynchronous receiver/transmitters connected between the first bus and the processor nodes, wherein the universal asynchronous receiver/transmitters include a plurality of main transmitters and a plurality of backup transmitters. 如請求項1所述的資料伺服器系統,其中該組第三光收發器包含DSFP連接埠及QSFP連接埠。A data server system as described in claim 1, wherein the set of third optical transceivers includes a DSFP port and a QSFP port. 如請求項1所述的資料伺服器系統,其中該些處理器節點的數量為240個,該些處理器節點的每四者設置於一運算卡上,且該運算卡支持熱插拔功能。A data server system as described in claim 1, wherein the number of the processor nodes is 240, every four of the processor nodes are arranged on a computing card, and the computing card supports hot-swap function.
TW113109245A 2024-03-13 2024-03-13 Data server system TWI883861B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW113109245A TWI883861B (en) 2024-03-13 2024-03-13 Data server system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW113109245A TWI883861B (en) 2024-03-13 2024-03-13 Data server system

Publications (2)

Publication Number Publication Date
TWI883861B true TWI883861B (en) 2025-05-11
TW202536584A TW202536584A (en) 2025-09-16

Family

ID=96581970

Family Applications (1)

Application Number Title Priority Date Filing Date
TW113109245A TWI883861B (en) 2024-03-13 2024-03-13 Data server system

Country Status (1)

Country Link
TW (1) TWI883861B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TW202331505A (en) * 2021-12-22 2023-08-01 美商賽發馥股份有限公司 Configuring a prefetcher associated with a processor core
US20240077781A1 (en) * 2022-09-06 2024-03-07 Luminous Computing, Inc. Computer architecture with disaggregated memory and high-bandwidth communication interconnects

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TW202331505A (en) * 2021-12-22 2023-08-01 美商賽發馥股份有限公司 Configuring a prefetcher associated with a processor core
US20240077781A1 (en) * 2022-09-06 2024-03-07 Luminous Computing, Inc. Computer architecture with disaggregated memory and high-bandwidth communication interconnects

Also Published As

Publication number Publication date
TW202536584A (en) 2025-09-16

Similar Documents

Publication Publication Date Title
WO2021169283A1 (en) Ocp adapter card and computer device
US7783818B1 (en) Modularized interconnect between root complexes and I/O modules
US6633946B1 (en) Flexible switch-based I/O system interconnect
CN109120624B (en) Multi-plane loose coupling high-bandwidth data exchange system
CN102473157A (en) Virtual hot inserting functions in a shared I/O environment
WO2007062593A1 (en) A kind of server
CN114721992B (en) Server and server management system thereof
CN104657317A (en) Server
WO2024212744A1 (en) Server, heterogeneous device, and data processing apparatus thereof
CN105099776A (en) Cloud server management system
CN101639821B (en) SMBUS interface chip of sever with content redundant link
WO2014067264A1 (en) Switch board and blade server
KR20250103762A (en) PCIe retimer providing failover to redundant endpoints using die-to-die data interface
US20080147924A1 (en) Chipset Agnostic Apparatus and Method for Serial Communication Bus Port Disablement
US20100257294A1 (en) Configurable provisioning of computer system resources
CN100484003C (en) Server
CN1901530B (en) a server system
TWI883861B (en) Data server system
CN119690685B (en) Memory pooling system and server system
US20250287127A1 (en) Data server system
WO2025157145A1 (en) Processor platform, circuit board, and server
CN216352292U (en) Server mainboard and server
CN207833500U (en) A kind of 40G rate network interface cards for supporting non-standard interface
CN117149674A (en) Multi-host PCIe (peripheral component interconnect express) interconnection structure for embedded real-time scene
CN101290556A (en) Disk Array Shared Device