Add artifact-evaluation

copy-file.js

@@ -4,7 +4,7 @@ const path = require('path')
 
 const TARGET_FILE = path.resolve('./docs/')
 const AIM_FILE = path.resolve('./dist/')
-const AIM_FILE_TYPE = ['.json', '.txt', '.png', '.jpg', '.jpeg', '.svg']
+const AIM_FILE_TYPE = ['.json', '.txt', '.png', '.jpg', '.jpeg', '.svg', '.mp4']
 const IGNORE_FILE = ['.vuepress']
 
 handleFile(TARGET_FILE)

docs/04-auto-tracing/04-artifact-evaluation.md

@@ -0,0 +1,55 @@
+---
+title: Artifact Evaluation for DeepFlow
+permalink: /auto-tracing/artifact-evaluation
+---
+
+# DeepFlow paper
+
+We attach our submitted paper for the committee's reference.
+[Network-Centric Distributed Tracing with DeepFlow: Troubleshooting Your Microservices in Zero Code](https://github.com/deepflowio/deepflow/blob/AEC/docs/deepflow_sigcomm2023.pdf)
+
+# Source code
+
+[https://github.com/deepflowio/deepflow/tree/AEC](https://github.com/deepflowio/deepflow/tree/AEC)
+
+Currentlyource code includes the following core modules:
+
+- agent
+
+  It is used to capture trace data using pre-defined eBPF instrumentation hooks and instrumentation extensions. In addition, the Agent is responsible for integrating metrics and tags from third-party frameworks or cloud platforms and transmitting them to the Server.
+  Among them, `agent/src/ebpf` is the source code related to eBPF, which is divided into two parts:
+
+  - `kernel`: The source code of the eBPF program, will be compiled into bytecode and loaded into the kernel.
+  - `user`: Used for load eBPF bytecode, and receiving tracing data.
+
+- server
+
+  It is responsible for storing spans in the database and assembling them into traces when users query.
+
+# How to install and run artifact
+
+We have prepared a demo of a microservice for your review, sourced from [this GitHub repository](https://github.com/istio/istio/tree/master/samples/bookinfo).
+Please follow the instructions in the document below to install and run artifact：
+
+- [Deploying Istio Bookinfo Demo](https://deepflow.io/docs/auto-tracing/istio-bookinfo-demo/).
+- [Deploying DeepFlow to monitor a single K8s cluster](https://deepflow.io/docs/install/single-k8s/).
+
+We have set up an online accessible environment for you. Please [click here](https://ce-demo.deepflow.yunshan.net/d/Distributed_Tracing/distributed-tracing?var-namespace=deepflow-ebpf-istio-demo&from=deepflow-doc) to access it.
+
+# Artifact Operating Instructions
+
+According to the introduction in Section 3, you can build and run the artifact on your own, or you can [click here](https://ce-demo.deepflow.yunshan.net/d/Distributed_Tracing/distributed-tracing?var-namespace=deepflow-ebpf-istio-demo&from=deepflow-doc) to access it directly online.
+Go to Grafana, open the `Distributed Tracing` Dashboard, and after selecting `namespace` = `deepflow-ebpf-istio-demo`, you can choose a specific call to trace.
+
+The following video demonstrates the procedure:
+
+<video width="100%" height="100%" controls autoplay muted>
+  <source src="./imgs/artifact-evaluation.mp4" type="video/mp4">
+</video>
+
+In the `Flame Graph`, you will observe associations among all spans. Here,`S` represents system spans, which are collected by attaching eBPF programs to application service system call interfaces. `N` represents network spans, and the data is collected using the `AF_PACKET` method. These data are correlated using the following four field values:
+
+- `syscall_trace_id_request` and `syscall_trace_id_responseare` trace IDs generated based on the same thread or coroutine, enabling intra-microservice data tracing (association of spans within microservices).
+- `req_tcp_seq` and `resp_tcp_seq` are TCP sequence numbers used for associating client and server spans.
+
+You can click on `Related Data` in order to view the connections between all spans.