算力網絡詳解 捲1：算網大腦

目 錄

 

第1章 算力網絡概述 ··············1

 

1.1　算力網絡概念······················2

 

1.1.1　算力網絡的定義 ····················2

 

1.1.2　算力網絡與電力網絡 ··············3

 

1.2　新型數字化經濟需求·············4

 

1.2.1　人工智能產業化需求 ··············4

 

1.2.2　分佈式多級算力需求 ··············5

 

1.3　算力網絡發展的基礎·············8

 

1.3.1　算力網絡的技術底座 ··············8

 

1.3.2　“雙碳”社會的國際共識 ·······10

 

1.3.3　算力網絡發展的國家政策 ·······11

 

1.4　算力網絡的新應用形態介紹 ·······12

 

1.4.1　新型算力組合 ······················12

 

1.4.2　新型算力模式 ······················12

 

1.5　算力網絡產業規劃···············13

 

1.5.1　算力網絡的標準規劃 ·············13

 

1.5.2　中國運營商的算力網絡規劃 ····15

 

第2章 算力網絡與雲網融合的關系 ·············20

 

2.1　主流運營商與雲商的雲網融合發展戰略 ·····20

 

2.2　雲網融合技術特點···············24

 

2.2.1　雲基礎設施技術 ···················24

 

2.2.2　網基礎設施技術 ···················28

 

2.2.3　雲網融合管理編排技術 ··········29

 

2.3　雲網融合典型場景···············33

 

2.4　從雲網融合到算力網絡·········35

 

2.5　算力網絡目標·····················37

 

第3章算力網絡技術體系 ······40

 

3.1　算力網絡標準體系進展·········40

 

3.1.1　ITU標準中的算力網絡體系框架 ·······40

 

3.1.2　CCSA標準中算力網絡體系架構 ···········43

 

3.2　亞信科技算力網絡系統參考體系··············44

 

3.2.1　系統功能架構 ······················45

 

3.2.2　系統接口設計 ······················46

 

3.3　算網運營交易中心···············50

 

3.3.1　中心簡介 ····························50

 

3.3.2　功能介紹 ····························50

 

3.3.3　集成關系 ····························54

 

3.4　算力網絡編排中心···············54

 

3.4.1　中心簡介 ····························54

 

3.4.2　功能介紹 ····························55

 

3.4.3　集成關系 ····························57

 

3.5　網絡管理調度中心···············57

 

3.5.1　中心簡介 ····························57

 

3.5.2　功能介紹 ····························57

 

3.5.3　集成關系 ····························59

 

3.6　算力管理調度中心···············59

 

3.6.1　中心簡介 ····························59

 

3.6.2　功能介紹 ····························60

 

3.6.3　集成關系 ····························61

 

3.7　算力網絡智能引擎 ··············62

 

3.7.1　簡介 ··································62

 

3.7.2　功能介紹 ····························62

 

3.7.3　集成關系 ····························63

 

3.8　算力網絡數字孿生中心·········64

 

3.8.1　中心簡介 ····························64

 

3.8.2　功能介紹 ····························64

 

3.8.3　集成關系 ····························66

 

3.9　算力網絡基礎設施···············66

 

3.9.1　網絡基礎設施 ······················66

 

3.9.2　算力基礎設施 ······················68

 

3.10　業務流程介紹 ···················70

 

3.10.1　算力網絡業務開通 ···············70

 

3.10.2　算力網絡業務變更 ···············72

 

3.10.3　算力網絡業務撤銷 ···············74

 

3.10.4　算力註冊並網 ·····················75

 

3.10.5　應用服務部署 ·····················76

 

3.10.6　應用服務自智保證 ···············77

 

第4章 算力網絡運營交易關鍵技術 ·········79

 

4.1　多量綱的模型體系催生多形態的算力網絡商品 ········79

 

4.1.1　三要素體系定義 ···················79

 

4.1.2　多量綱模型體系 ···················81

 

4.1.3　多形態算力網絡商品 ·············81

 

4.2　多樣化的商業模式促成雙驅動的算力網絡受理 ················82

 

4.2.1　商業模式定義 ······················82

 

4.2.2　雙驅動算力網絡受理 ·············84

 

4.3　區塊鏈的技術底座實現可信任的算力網絡交易 ················85

 

4.3.1　算力網絡存證 ······················86

 

4.3.2　算力網絡通證 ······················87

 

4.3.3　智能合約 ····························88

 

4.3.4　交易溯源 ····························89

 

4.4　沈浸式的用戶體驗發掘算力網絡的新價值 ······90

 

第5章算力網絡一體編排關鍵技術··················91

 

5.1　算力網絡節點評估···············91

 

5.1.1　算力網絡節點定義 ················91

 

5.1.2　算力網絡節點性能指標 ··········93

 

5.1.3　算力網絡節點資源指標 ··········98

 

5.1.4　性能與資源關系模型 ···········103

 

5.2　算力網絡融合資源視圖·······107

 

5.2.1　算力網絡鏈路 ····················107

 

5.2.2　算力節點感知 ····················110

 

5.2.3　算力網絡拓撲 ····················113

 

5.3　算網融合感知···················114

 

5.3.1　數據統一採集 ····················115

 

5.3.2　數據統一存儲 ····················117

 

5.3.3　數據時間粒度 ····················120

 

5.3.4　跨域數據關聯 ····················121

 

5.3.5　數據開放使用 ····················123

 

5.4　算力網絡一體化編排流程····126

 

5.4.1　業務SLA ··························127

 

5.4.2　節點編排決策 ····················129

 

5.4.3　路徑編排決策 ····················133

 

5.4.4　算網編排方案 ····················135

 

第6章算力網絡管理調度關鍵技術 ··············138

 

6.1　SRv6網絡管理調度 ···········138

 

6.1.1　技術架構 ··························138

 

6.1.2　調度機制 ··························140

 

6.1.3　能力開放 ··························148

 

6.2　SD-OTN網絡管理調度·······150

 

 

6.2.1　技術架構 ··························150

 

6.2.2　能力開放 ··························154

 

6.3　中心雲管理調度················156

 

6.3.1　多算力雲管理 ····················157

 

6.3.2　異構算力資源納管 ··············158

 

6.3.3　算力資源管理 ····················160

 

6.3.4　算力資源調度 ····················161

 

6.4　MEC管理調度 ·················162

 

6.4.1　MEC組成架構 ···················163

 

6.4.2　MEC服務 ·························168

 

6.4.3　MEC能力開放 ···················169

 

6.5　5G切片管理調度 ··············171

 

6.5.1　技術簡介 ··························171

 

6.5.2　切片架構 ··························173

 

6.5.3　關鍵能力 ··························175

 

6.6　O-RAN管理調度 ··············180

 

6.6.1　O-RAN 架構 ······················181

 

6.6.2　O-RAN 資源控制調度機制 ····182

 

6.6.3　O-RAN 能力開放 ················183

 

6.7　5GC網絡管理調度 ············185

 

6.7.1　5GC系統架構 ····················185

 

6.7.2　5GC 管理調度機制 ··············186

 

6.7.3　5GC能力開放 ····················189

 

6.8　算力資源彈性調度·············191

 

6.8.1　資源適配 ··························191

 

6.8.2　資源管理 ··························193

 

6.8.3　應用管理 ··························195

 

第7章 算力網絡智能運維關鍵技術 ···············198

 

7.1　算力網絡自智運維管理·······198

 

7.1.1　算力網絡資源監控 ··············198

 

7.1.2　算力網絡運行智能預警 ········203

 

7.1.3　算力網絡業務故障自處理 ·····205

 

7.2　算力網絡業務質量保障·······210

 

7.2.1　端到端保障體系 ·················210

 

7.2.2　用戶體驗模型 ····················211

 

7.2.3　服務質量模型 ····················212

 

第8章算力網絡智能引擎關鍵技術 ·············214

 

8.1　算力網絡智能引擎的基本概念··················214

 

8.2　算力網絡智能引擎的系統組件·················215

 

8.3　算力網絡引擎的研究現狀····216

 

8.4　算力網絡智能引擎關鍵技術················217

 

8.4.1　算力網絡節點評估算法 ········217

 

8.4.2　算力網絡鏈路評估技術 ········224

 

8.4.3　算力網絡綜合代價函數評估 ·················226

 

8.5　算力網絡智能引擎應用場景示例···················228

 

第9章算力網絡數字孿生關鍵技術 ······················231

 

9.1　應用價值·························231

 

9.2　問題現狀·························233

 

9.3　業務目標·························234

 

9.4　核心算法·························236

 

9.4.1　分級多層佈局 ····················236

 

9.4.2　混合空間佈局 ····················236

 

9.4.3　縱向多級穿透 ····················237

 

9.5　關鍵技術·························238

 

9.5.1　3D地理可視化技術 ·············238

 

9.5.2　網絡拓撲可視化技術 ···········239

 

9.5.3　遊戲引擎技術 ····················239

 

9.6　應用場景·························240

 

第10章 應用案例分析與設計 ······················241

 

10.1　雲—邊—端算力協同 ········242

 

10.1.1　AR/VR業務算網方案 ·········242

 

10.1.2　公安機關嫌疑犯識別 ··········245

 

10.2　多雲算力協同 ·················248

 

10.2.1　基於聯邦學習的醫療推薦業務場景 ···················248

 

10.2.2　“東數西算”數據遷移場景 ······················252

 

10.3　雲邊算力協同 ·················255

 

第11章 未來展望 ···············259