The Many Faces of Struct Tags Kyle Erf and Sam Helman Why are we here? ● To showcase and explain an underused feature of the language ● To galvanize you to consider struct tags as a potential Priorities []string `stated_priorities` } x := Candidate{} y := VocalCandidate(x) // <-- not allowed Why? ● Attach arbitrary values to your fields ● Provided without a specific use case Namespacing What github.com/mitchellh/mapstructure Library for converting between structs and Go's builtin map. Why? "If you have configuration or an encoding that changes slightly depending on specific fields" In

Chuang (jojochuang) Ritesh Shukla (kerneltime) Agenda • Overview of how Ozone and how it scales • Why NVME is important for Ozone for scaling • Benefits of using NVME • Impala performance results from • Write path improvements results from NVME clusters • Summary • Questions Ozone Architecture Why does Ozone Scale? Separation of concerns Ozone Manager Ozone Client Datanode Storage Container background load Datanode Datanode Datanode hadoop-hdds hadoop-ozone Scale out No foreground load Why does Ozone Scale? Aggregation via containers Chunks Chunks Chunks Blocks User Chunks Chunks Blocks

the following features: Multi-tenancy Kyuubi supports the end-to-end multi-tenancy, and this is why we want to create this project despite that the Spark Thrift JDBC/ODBC server [http://spark.apache Spark Adaptive Query Execution (AQE) in Kyuubi Engines 1. The Share Level Of Kyuubi Engines 1.1. Why do we need this feature? 1.2. The current supported share levels 1.3. Related Configurations 1.4. Conclusion mode. In other words, an engine is cluster widely shared by all Kyuubi server peers if could. 1.1. Why do we need this feature? Apache Spark is a unified engine for large-scale data analytics. Using Spark

the following features: Multi-tenancy Kyuubi supports the end-to-end multi-tenancy, and this is why we want to create this project despite that the Spark Thrift JDBC/ODBC server [http://spark.apache Spark Adaptive Query Execution (AQE) in Kyuubi Engines 1. The Share Level Of Kyuubi Engines 1.1. Why do we need this feature? 1.2. The current supported share levels 1.3. Related Configurations 1.4. Conclusion mode. In other words, an engine is cluster widely shared by all Kyuubi server peers if could. 1.1. Why do we need this feature? Apache Spark is a unified engine for large-scale data analytics. Using Spark

Service Mesh for VM Native, Chris Crall, Jianfei Hu, Google Cloud Next ‘19 #IstioCon Why Add VMs to the Mesh? ● = Why Service Mesh? ○ More services = more complexity ○ Need consistent policy enforcement Observability ○ See VM metrics alongside containers ● Extensibility #IstioCon Why Should Istio Support VMs ● ≈ Why VMs? ○ Technical reasons ■ Better known security controls ■ Better isolation (of seems closer to be production ready? Should we expect more? And what do we need else? #IstioCon Why We Expect More? A Closer Look… ● Example use case: Telco & Edge computing ○ where VMs play a crucial

USAGE GUIDE CHAPTER ONE MULTI-TENANCY Kyuubi supports the end-to-end multi-tenancy, and this is why we want to create this project despite that the Spark Thrift JDBC/ODBC server already exists. 1. Supports node mode. In other words, an engine is cluster widely shared by all Kyuubi server peers if could. Why do we need this feature? Apache Spark is a unified engine for large-scale data analytics. Using Spark missing on this page about Apache Kudu background knowledge, you can refer to its official website. Why Kyuubi on Kudu Basically, Kyuubi can take place of HiveServer2 as a multi tenant ad-hoc SQL on Hadoop

USAGE GUIDE CHAPTER ONE MULTI-TENANCY Kyuubi supports the end-to-end multi-tenancy, and this is why we want to create this project despite that the Spark Thrift JDBC/ODBC server already exists. 1. Supports node mode. In other words, an engine is cluster widely shared by all Kyuubi server peers if could. Why do we need this feature? Apache Spark is a unified engine for large-scale data analytics. Using Spark missing on this page about Apache Kudu background knowledge, you can refer to its official website. Why Kyuubi on Kudu Basically, Kyuubi can take place of HiveServer2 as a multi tenant ad-hoc SQL on Hadoop

the following features: Multi-tenancy Kyuubi supports the end-to-end multi-tenancy, and this is why we want to create this project despite that the Spark Thrift JDBC/ODBC server [http://spark.apache transport. transportMode=http Integrations 1. Kyuubi On Apache Kudu 1.1. What is Apache Kudu 1.2. Why Kyuubi on Kudu 1.3. Kudu Integration with Apache Spark 1.4. Kudu Integration with Kyuubi 1.5. References on this page about Apache Kudu background knowledge, you can refer to its official website. 1.2. Why Kyuubi on Kudu Basically, Kyuubi can take place of HiveServer2 as a multi tenant ad-hoc SQL on Hadoop

the following features: Multi-tenancy Kyuubi supports the end-to-end multi-tenancy, and this is why we want to create this project despite that the Spark Thrift JDBC/ODBC server [http://spark.apache transport. transportMode=http Integrations 1. Kyuubi On Apache Kudu 1.1. What is Apache Kudu 1.2. Why Kyuubi on Kudu 1.3. Kudu Integration with Apache Spark 1.4. Kudu Integration with Kyuubi 1.5. References on this page about Apache Kudu background knowledge, you can refer to its official website. 1.2. Why Kyuubi on Kudu Basically, Kyuubi can take place of HiveServer2 as a multi tenant ad-hoc SQL on Hadoop

Stability and more functionalities Alternative solutions? • Why not use DPDK? • DPDK performs busy polling even when network is idle. • Why not use a pure eBPF service? • Not mature enough eBPF brief kernel • Attach to network tc hooks • Triggered by ingress/egress packets IPVS bypass conntrack • Why IPVS depends on conntrack? • Iptables/conntrack SNAT • How IPVS bypasses conntrack? • Ingress • Pre-route IPVS entry BPF SNAT IPVS mode data path IPVS-eBPF mode data path How eBPF does SNAT • Why does SNAT with eBPF • eBPF program is easy to deploy • How to do SNAT in eBPF • Do SNAT in TC egress