layers of AI infrastructure: agentic interfaces, enterprise copilots, real-world autonomous systems, and sovereign models. Rapid advances in artificial intelligence, compute infrastructure, and global ‘AI Index: Mapping the $4 Trillion Enterprise Impact’ via Morgan Stanley (10/23) Enabling Infrastructure CPUs Big Data / Cloud GPUs Computing Cycles Over Time – 1960s-2020s, per Morgan Stanley we access and move information is happening much faster… …AI is a compounder – on internet infrastructure, which allows for wicked-fast adoption of easy-to-use broad-interest services. AI Technology

为什么Istio选择Mixer和Proxy分离的架构？ Part 1：ServiceMesh灵魂拷问一：要架构还是要性能？ Proxy Mixer In- Process Adapter Infrastructure Backend Proxy 优点： • 架构优雅，职责分明，边界清晰 • Mixer的变动不影响Proxy • Proxy无需和Adapter耦合 • 读取配置 -> 连接k8s/Galley Adapter? Part 1：ServiceMesh灵魂拷问一：要架构还是要性能？ Proxy Mixer Bypass Adapter Out-of- Process Adapter Infrastructure Backend ？ Adapters Configurations 优点： • 架构更优雅，职责更分明，边界更清晰 • Out-of-Process Adapter不再是Istio的组成部分 Proxy Mixer In- Process Adapter Infrastructure Backend Proxy Mixer Bypass Adapter Out-of- Process Adapter Infrastructure Backend Proxy Infrastructure Backend In- Process AdapterPart 1

the application layer while using existing infrastructure. Program Scope Program spans core capabilities and underlying platform or infrastructure. The government is responsible for primary Complexity Many contractors are required to develop program elements. Program can leverage test infrastructure and automated tests, and testers are active throughout development. Test Environment and have a foundational understanding of the desired operational capabilities and technical infrastructure required. 8.1 Documentation The government can develop capstone documents at the larger

high performers had MTTR 168x faster than low performers b. CREATE OUR CENTRALIZED TELEMETRY INFRASTRUCTURE i. Remove the silos of information – Developers don’t just log what’s interesting to development fact-based problem-solving e. ENABLE CREATION OF PRODUCTION METRICS AS PART OF DAILY WORK i. Create infrastructure to make it easy for Dev or Ops to create telemetry for functions built ii. Generate graphs problems g. FIND AND FILL ANY TELEMETRY GAPS i. Expand metrics from business, application, infrastructure, client software, and deployment pipeline levels 1. With every production incident identify

risks of the underlying infrastructure must be measured and quantified, so that the total risks and impacts to software applications are understood.  Deploy immutable infrastructure, such as containers concept of immutable infrastructure is an IT strategy in which deployed components are replaced in their entirety, rather than being updated in place. Deploying immutable infrastructure requires standardization standardization and emulation of common infrastructure components to achieve consistent and predictable resultConceptual Model DevSecOps LifecycleDevSecOps Pillars DevSecOps EcosystemDevSecOps Software Factory

build support for improvements to our codebase CB Bailey Designs and maintains widely used infrastructure tools and processes underpinning thousands of applications and shared C++ libraries© 2018 Bloomberg Finance L.P. All rights reserved. 10 We are, however, both part of Bloomberg’s Technology Infrastructure (TI) team where shared the concern that that our huge codebase is outgrowing critical container Bloomberg has ~50,000 libraries, and the company has been around for 40 years. In addition to infrastructure teams whose job it is to look at library architecture, there are dozens of engineers who have

share local discoveries to the entire organization 1. Configuration standards for libraries, infrastructure, and environments 2. Deployment tools 3. Testing standards and tools, including security 4 ability to not be responsible for unsupported platforms ii. Systematically review production infrastructure and services for items that are causing disproportionate amounts of failure and unplanned work; other security/auditing relevant requirements iv. Reduce the friction by providing application/infrastructure stacks that are pre- approved and appropriately configured and secured. f. INTEGRATE SECURITY

Apache (incubating) TVM project. A goal is to nurture the TVM community and contribute new infrastructure and features. octom|.ai @octoml Q octoML Founding Team - The Octonauts - 四人全外日 Luis Ceze employees to contribute to TVML. ee Today we'ltouch on a few of those contribution areas: o Core Infrastructure Improvements to TVM o_uTVM: support for microcontrollers in TVM o_ Virtual Machine and dynamic o_ Improved NLP support, with focus on transformers QQ octoML Core Infrastructure Refactors ee New Integer Analysis Infrastructure o_ Supports the ability to handle nested division and modulus o_ Improves

Segregated technical and business functions  Development  Operations (Operational Waterfall)  Infrastructure Ops  Product Ops  Product Management  Every technology under the sun  Solaris, Windows iterations, planning and demos  Put everything into a repository (configurations, scripts, etc.) Infrastructure as Code  Turned over repeatable and automatable tasks to the product team wherever possible  How we need to improve  Sharing ideas, code, Community of Practice, etc.  Test Driven Infrastructure  Blue – green deployments  Combining DevOps Scrum – planning, standups, boards, and backlogs

Compliance: e. Myth—DevOps Means Eliminating IT Operations, or “NoOps” f. Myth—DevOps is Just “Infrastructure as Code” or Automation: g. Myth—DevOps is Only for Open Source Software: 2. Foreword xix can update independently without having to update other services. iii. Decouple from shared infrastructure – shared databases iv. Bounded context – compartmentalized and well-defined interfaces k. KEEP PRODUCTIVITY i. Enable Dev teams to spend more time building functionality, as opposed to created infrastructure ii. Automation – across the board iii. Create known, good environments that are ready for production