maplain.github.io/index.html at master · maplain/maplain.github.io

Repository navigation

History

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569

570

571

572

573

574

575

576

577

578

579

580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

600

601

602

603

604

605

606

607

608

609

610

611

612

613

614

615

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

636

637

638

639

640

641

642

643

644

645

646

647

648

649

650

651

652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

667

668

669

670

671

672

673

674

675

676

677

678

679

680

681

682

683

684

685

686

687

688

689

690

691

692

693

694

695

696

697

698

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

718

719

720

721

722

723

724

725

726

727

728

729

730

731

732

733

734

735

736

737

738

739

740

741

742

743

744

745

746

747

748

749

750

751

752

753

754

755

756

757

758

759

760

761

762

763

764

765

766

767

768

769

770

771

772

773

774

775

776

777

778

779

780

781

782

783

784

785

786

787

788

789

790

791

792

793

794

795

796

797

798

799

800

801

802

803

804

805

806

807

808

809

810

811

812

813

814

815

816

817

818

819

820

821

822

823

824

825

826

827

828

829

830

831

832

833

834

835

836

837

838

839

840

841

842

843

844

845

846

847

848

849

850

851

852

853

854

855

856

857

858

859

860

861

862

863

864

865

866

867

868

869

870

871

872

873

874

875

876

877

878

879

880

881

882

883

884

885

886

887

888

889

890

891

892

893

894

895

896

897

898

899

900

901

902

903

904

905

906

907

908

909

910

911

912

913

914

915

916

917

918

919

920

921

922

923

924

925

926

927

928

929

930

931

932

933

934

935

936

937

938

939

940

941

942

943

944

945

946

947

948

949

950

951

952

953

954

955

956

957

958

959

960

961

962

963

964

965

966

967

968

969

970

971

972

973

974

975

976

977

978

979

980

981

982

983

984

985

986

987

988

989

990

991

992

993

994

995

996

997

998

999

1000

<!DOCTYPE html>

<head>

var NexT = window.NexT || {};

var CONFIG = {

root: '/',

scheme: 'Muse',

version: '7.3.0',

sidebar: {"position":"left","display":"post","offset":12,"onmobile":false},

back2top: {"enable":true,"sidebar":false,"scrollpercent":false},

copycode: {"enable":false,"show_result":false,"style":null},

fancybox: false,

mediumzoom: false,

lazyload: false,

pangu: false,

algolia: {

applicationID: '',

apiKey: '',

indexName: '',

hits: {"per_page":10},

labels: {"input_placeholder":"Search for Posts","hits_empty":"We didn't find any results for the search: ${query}","hits_stats":"${hits} results found in ${time} ms"}

localsearch: {"enable":false,"trigger":"auto","top_n_per_article":1,"unescape":false,"preload":false},

search: {

root: '/',

path: ''

tabs: true,

motion: {"enable":true,"async":false,"transition":{"post_block":"fadeIn","post_header":"slideDownIn","post_body":"slideDownIn","coll_header":"slideLeftIn","sidebar":"slideUpIn"}},

translation: {

copy_button: 'Copy',

copy_success: 'Copied',

copy_failure: 'Copy failed'

}

};

</script>

CONFIG.page = {

sidebar: "",

};

</script>

<title>Fangyuan Li's Blog</title>

<style>

.use-motion .motion-element,

.use-motion .brand,

.use-motion .menu-item,

.sidebar-inner,

.use-motion .post-block,

.use-motion .pagination,

.use-motion .comments,

.use-motion .post-header,

.use-motion .post-body,

.use-motion .collection-title { opacity: initial; }

.use-motion .logo,

.use-motion .site-title,

.use-motion .site-subtitle {

opacity: initial;

top: initial;

}

.use-motion .logo-line-before i { left: initial; }

.use-motion .logo-line-after i { right: initial; }

</style>

</noscript>

</head>

<span class="site-title">Fangyuan Li's Blog</span>

</a>

</div>

</button>

</div>

</li>

<a href="/archives/" rel="section">Archives</a>

</li>

</li>

</li>

</ul>

</nav>

</div>

</header>

</span>

</span>

<a href="/2021/07/16/What-makes-k8s-better/" class="post-title-link" itemprop="url">What makes k8s better</a>

</h1>

</span>

<span class="post-meta-item-text">Posted on</span>

</span>

</span>

<span class="post-meta-item-text">Edited on</span>

</span>

<br>

</div>

</header>

<p>It’s easy to create a process, but it’s hard to manage its lifecycle. What if it<br>dies, what if resources are insufficient, what if you want to create more<br>replicas, etc.</p>

<p>It’s easy to make a REST request, but it’s hard to manage its lifecycle. What if<br>it fails, what if you need to make many to achieve one logical goal, what if<br>some of them fail, what if the order matters and changes later, etc.</p>

<p>Therefore:<br>Q: What’s the difference between a Linux process and a K8s operator running in<br>Pod?</p>

<p>A: Roughly speaking, pod is a <em>managed</em> process whereas process is a pure compute unit, which could be managed by systemd etc.</p>

<p>Q: What’s the difference between managing some infra resources through making a<br>bunch of REST calls and creating a CRD for a k8s operator?</p>

<p>A: Roughly speaking, CRD is a <em>managed</em> request. Also yaml is a better user<br>interface.</p>

<p>Intention reads better than implementation. The system that helps you to focus<br>on intention wins eventually.</p>

</div>

</footer>

</div>

</article>

</span>

</span>

<a href="/2021/01/03/矛盾和统一/" class="post-title-link" itemprop="url">矛盾和统一</a>

</h1>

</span>

<span class="post-meta-item-text">Posted on</span>

</span>

<br>

</div>

</header>

<p>如果说人的异化以及”大理性“和”小理性“的失调是马克思主义学者们对现代资本主义的控诉，那么:</p>

<ol>

<li>是否存在纯粹建立在“大理性”或者“小理性”上的国家？</li>

<li>如果有，比如古希腊和古罗马的某些时期:)，那么由他们的发展历程是否可以发现所谓的“矛盾”恰恰构成了其他国家或时期社会稳定的基础？</li>

<li>如果这样的”矛盾“构成社会稳定的基础，那怎样保证”矛盾“的冲突始终在可控范围？比如：如果美国梦和宣扬禁欲伦理的新教是对资本主义财富积累的价值观的引导，但是如何平衡由之而来的种种泛民主的冲突和抗议，甚至暴力行为？话说回来，每个国家可能都有一个”美国梦“，比如”书中自有黄金屋“ :)</li>

</ol>

</div>

</footer>

</div>

</article>

</span>

</span>

</h1>

</span>

<span class="post-meta-item-text">Posted on</span>

</span>

<br>

</div>

</header>

<p>一旦人在科学研究中掺入了关于个人价值的判断，对事实的完整理解便将荡然无存。 –<br>Max Weber</p>

</blockquote>

<p>学术与科学的严肃性必然要求从知识沟通的机制中消除任何争论和攻击行为 – Max<br>Weber</p>

</blockquote>

<p>韦伯主张，知识分子在自己的生活中必须区分学者身份和行动者的身份。要是把两者所追求的判断混淆起来，就必然会危及到两种判断的有效性。</p>

</div>

</footer>

</div>

</article>

</span>

</span>

</h1>

</span>

<span class="post-meta-item-text">Posted on</span>

</span>

<br>

</div>

</header>

<p>Even though you’re building a To C Service, you should think about how could your service be<br>consumed one day by other companies.</p>

<p>It encourages modularity and the industry is shifting towards a more agile model<br>where more products and services could be easily built upon existing ones.</p>

<p>Try to make a building block instead of a silo.</p>

</div>

</footer>

</div>

</article>

</span>

</span>

<a href="/2020/10/12/Random-Thoughts-1/" class="post-title-link" itemprop="url">How to come up with good abstractions</a>

</h1>

</span>

<span class="post-meta-item-text">Posted on</span>

</span>

</span>

<span class="post-meta-item-text">Edited on</span>

</span>

<br>

</div>

</header>

<p>When thinking about abstraction, do not model objects, instead model intentions.<br>In other words, do not abstract nouns, abstract verbs.</p>

<p>Maybe that’s why composition is preferred in more places than inheritance.</p>

<p>And I think the former scales way better than the latter.</p>

</div>

</footer>

</div>

</article>

</span>

</span>

<a href="/2020/10/06/Random-Thoughts/" class="post-title-link" itemprop="url">Random Thoughts</a>

</h1>

</span>

<span class="post-meta-item-text">Posted on</span>

</span>

<br>

</div>

</header>

<p>Curiosity drives DFS, while entrepreneurship demands BFS</p>

</div>

</footer>

</div>

</article>

</span>

</span>

<a href="/2020/06/15/Cloud-Native-L4-Load-Balancer-MetalLB-NSX-T-and-Maglev/" class="post-title-link" itemprop="url">Cloud Native L4 Load Balancer: MetalLB, NSX-T and Maglev</a>

</h1>

</span>

<span class="post-meta-item-text">Posted on</span>

</span>

<br>

</div>

</header>

<p>Something magical happens when MetalLB is used in the following fashion:</p>

<ol>

<li>MetalLB is deployed in a dedicated LB cluster;</li>

<li>LB cluster is deployed in front of all workload clusters;</li>

<li>all Service type=LoadBalancer are projected into the LB cluster;</li>

</ol>

<p>Compare this setup with a traditional proprietary SDN, eg: NSX-T and Cloud LoadBalancer like Maglev used in GCP.</p>

<table>

<thead>

<tr>

<th>MetalLB in LB Cluster with Service Projected</th>

<th>Maglev</th>

</tr>

</thead>

<td>Control Plane</td>

<td>K8s API Server</td>

<td>NSX-T Manager</td>

<td>not mentioned, Borg?</td>

</tr>

<tr>

<td>Control Plane concurrency limit</td>

<td>1 million per second?</td>

<td>199 per second(NSX-T 2.5)</td>

<td>not mentioned, Borg?</td>

</tr>

<tr>

<td>CP Database</td>

<td>Corfu</td>

<td>not mentioned, Chubby?</td>

</tr>

<tr>

<td>Deployment Form</td>

<td>VM or most commonly Containers</td>

<td>unclear, mentioned Maglev deployment shares the same machines as other Applications, Borg?</td>

</tr>

<tr>

<td>South-North Data Plane</td>

<td>K8s Nodes</td>

<td>not mentioned, Borg nodes?</td>

</tr>

<tr>

<td>South-North <strong>Data Plane technology</strong></td>

<td>kube-proxy: iptables/ipvs</td>

<td>Nginx</td>

<td>optimized kernel-free datapath module</td>

</tr>

<tr>

<td>South-North Datapath</td>

<td>DNAT only; two hops in total: VIP->NodeIP→PodIP;</td>

<td>DNAT, DSR etc; one hop: VIP->PodIP;</td>

<td>DSR; hardware encapsulator between router and Maglev for fast overlay; one hop: VIP→Service EP;</td>

</tr>

<tr>

<td>Data Plane Programmability</td>

<td>K8s Controller + CR/Core Objects</td>

<td>NSX-T Data Model: LB + VirtualServer + ServerPool</td>

<td>MagLev Config objects which are committed atomically(implies a CP system like etcd or ZooKeeper(Google Chubby));</td>

</tr>

<tr>

<td>States management</td>

<td>Edge Active + Standby deployment</td>

<td>Maglev Consistent Hashing, minimize interruption yet optimize scale as much as possible, truly distributed; interruption rate is tunable by parameters in the consistent hashing algorithm;</td>

</tr>

<tr>

<td>cluster scalability</td>

<td>it doesn’t handle states, so unlimited</td>

<td>at most 10 nodes per edge cluster. in total, at most 160 edge nodes. one LB is only mapped to at most one pair of edge nodes though.</td>

<td>Maglev is stateless because it handles states in a stateless way(consistent hashing). so unlimited</td>

</tr>

</tbody></table>

<p>Clearly, the opportunity to build an enterprise-grade Distributed Software LB lies in the <strong>Dataplane</strong>.</p>

<ol>

<li><strong>Antrea</strong> serves as the lightweight version of NSX-T <strong>open vSwitch</strong> based dataplane agent;</li>

<li><strong>Cilium</strong> optimizes Dataplane using <strong>eBPF</strong> to replace vanilla Kube-proxy. That means we could potentially use Cilium in MetalLB Dedicated K8s Cluster to achieve better performance;</li>

</ol>

<p>Proposal:</p>

<ol>

<li>Use Cilium like <strong>eBPF-based</strong> module to optimize <strong>dataplane</strong><ol>

<li>could be deployed as Daemonset;</li>

<li>could be used to replace kube-proxy;</li>

</ol>

</li>

<li>Use Maglev <strong>Consistent Hashing</strong> to build truly distributed LB with <strong>states</strong> handled, meaning:<ol>

<li>connection <strong>stickiness</strong> is preserved as much as possible;</li>

<li><strong>scalable</strong> like Cloud; not traditional Active-standy or Active-Active model any more!</li>

</ol>

</li>

</ol>

<p>Besides,</p>

<p>based on Maglev paper, we need to add the following improvements <strong>based</strong> on <strong>MetalLB</strong> to implement something like Maglev:</p>

<ol>

<li>QoS: divide Services between multiple shards of LBs in the same cluster in order to achieve performance isolation;</li>

<li>Aggregation of VIP by a component like Routes Reflector sitting in front of all MetalLB BGP peers before all VIPs are published to the ToR Router/Gateway;</li>

</ol>

</div>

</footer>

</div>

</article>

</span>

</span>

<a href="/2020/01/13/Leveld-API-and-CAS/" class="post-title-link" itemprop="url">CAS, Level-based API and Agile</a>

</h1>

</span>

<span class="post-meta-item-text">Posted on</span>

</span>

</span>

<span class="post-meta-item-text">Edited on</span>

</span>

<br>

</div>

</header>

View remainder of file in raw view

FilesExpand file tree

index.html

Latest commit

History

index.html

File metadata and controls