Difference between revisions of "Archived-DMM"

From fd.io
Jump to: navigation, search
(What is DMM)
m (Valderrv moved page DMM to Archived-DMM)
 
(66 intermediate revisions by 6 users not shown)
Line 1: Line 1:
 +
 
{{Project Facts
 
{{Project Facts
 
|name=DMM
 
|name=DMM
 
|shortname=dmm
 
|shortname=dmm
 
|jiraName=DMM
 
|jiraName=DMM
|projectLead=George Zhao
+
|projectLead=George Zhao ''<George.Y.Zhao@huawei.com>''
 
|committers=
 
|committers=
* Kai Zheng
+
* Florin Coras  ''<fcoras@cisco.com>''
* George Zhao
+
* Ray Kinsella  ''<ray.kinsella@intel.com>''
* Yalei Wang
+
* Junfeng Li  ''<hotjunfeng@163.com>''
* Liuxiaoxu
+
* Hailong Wang  ''<whl739@163.com>''
* Swarup Nayak
+
* Kai Zheng  ''<kai.zheng@huawei.com>''
* Satish Karunanithi
+
* George Zhao  ''<George.Y.Zhao@huawei.com>''
* Qing Chang
+
* Qing Chang ''<Qing.Chang1@huawei.com>''
* Fanhongwei
+
* Liuxiaoxu  ''<liuxiaoxu1@huawei.com>''
* Florin Coras
+
* Yalei Wang  ''<william.wangyalei@huawei.com>''
* Ray Kinsella
+
* Swarup Nayak  ''<swarup.nayak1@huawei.com>''
* Junfeng Li
+
* Satish Karunanithi  ''<satishk@huawei.com>''
* Hailong Wang
+
 
}}
 
}}
  
Line 27: Line 27:
  
 
[[Project_Proposals/DMM|DMM Project Proposal]]
 
[[Project_Proposals/DMM|DMM Project Proposal]]
 
<!--[[File:TLDK - Network Sample Code - TCP plan.pptx|thumb|TLDK TCP Plan]] -->
 
  
 
== History ==
 
== History ==
  
Emerging applications are bringing extremely high-performance requirements to the network system. Eg. AR/VR, IOT etc. And Many of them
+
Emerging applications are bringing extremely high-performance requirements to their network systems. Eg. AR/VR, IoT etc. Many of these applications also
coming with their unique demand of QOS/SLA. Some application need low latency network, some need high reliability etc. Though such performance  
+
have unique demands for QOS/SLA. Some application need a low latency network, some need high reliability, etc. Though such performance  
targets should be required for the complete communication system, the transport layer protocols play a key role and encountered a relatively higher  
+
targets should be required for the complete communication system, the transport layer protocols play a key role and encounter a relatively higher  
challenge, because traditionally the TCP-based transport layer exploits the “best-effort” principle and provides no performance guarantees in its nature.
+
challenge, because traditionally the TCP-based transport layer exploits the “best-effort” principle and inherently provides no performance guarantees.
However, as Internet applications rapidly grow and diversify, an all-powerful or one-fits-all protocol or algorithm becomes less feasible. Thus,
+
However, as internet applications rapidly grow and diversify, an all-powerful or one-size-fits-all protocol or algorithm becomes less feasible. Thus  
the traditional single-instance TCP-based network stack bears great challenges when serving many applications with different QoS/SLA requirements  
+
the traditional single-instance TCP-based network stack faces many challenges when serving applications with different QoS/SLA requirements  
 
simultaneously on the same platform.  
 
simultaneously on the same platform.  
  
Also Moving the networking stack out of the kernel is an obvious trend in both the industry and literature.
+
Moving the networking stack out of the kernel is an occurring trend in both industry implementations and academic literature.
Technolgies like DPDK, etc improving performance of network stack, by bypassing the kernel, avoiding context-switching and data copies, as well  
+
Technologies like DPDK (and others) are improving performance of the network stack, bypassing the kernel and avoiding context-switching and data copies, as well as providing a complete set of packet-processing acceleration libraries.
as providing a complete set of packet-processing acceleration libraries.
+
  
Keeping above trends in mind the DMM/nStack provides a framework where, system operators can plug in dedicated types of networking stack  
+
Keeping the above trends in mind the DMM/nStack provides a framework where, system operators can plug in dedicated types of networking stack  
instances according to performance and/or functional requirements from the user space applications. Application doesnt have to worry about
+
instances according to performance and/or functional requirements from the user space applications. The application need not consider changes to the transport layer API. A lightweight nStack management daemon is responsible for maintaining the stack instances and the  
change their transport layer API. A lightweight nStack management daemon is responsible for maintaining the stack instances and the  
+
 
app/socket-to-stack-mappings, which are provided via the orchestration/management interface.
 
app/socket-to-stack-mappings, which are provided via the orchestration/management interface.
 +
 +
== DMM Architecture ==
 +
 +
[[File:Dmm-framework-arc.png|center|450px|DMM Simple Arch]]
 +
<br>
 +
 +
The DMM framework provides posix socket APIs to the application. A protocol stack could be plugged into the DMM. DMM will choose the most suitable stack according to the RD policy.
 +
 +
== Quick Start ==
 +
With this quick start, users can have a simple understanding of DMM.
 +
Following the quick start, it will set environment, build DMM and test apps automatically on Vagrant. <br>
 +
The test topology will typically look like this. It uses DMM + kernel stack now.
 +
Make sure vagrant and virtual box have been installed first.
 +
 +
[[File:Dmm_simple_test_topology.png|center|450px|DMM Simple Test topology]]
 +
<br>
 +
1. Download dmm code
 +
 +
    # git clone https://gerrit.fd.io/r/dmm
 +
 +
2. Setup Vagrant VMs
 +
 
 +
  # cd dmm/resources/extras
 +
 +
If users want to change the DISTRO, NICS, VMCPU or VMRAM settings, modified the env.sh and run source env.sh.
 +
 +
  # vagrant up
 +
 +
This steps will take a while. Until vagrant automatically set the environment and compile dpdk, glog and DMM.
 +
 +
3. Log in to server/client in two different terminals.
 +
 +
  # vagrant ssh dmm-test-server
 +
            and
 +
  # vagrant ssh dmm-test-client
 +
   
 +
4. Users can test sample client/server app in /dmm/release/bin.
 +
 +
Before testing apps, users can update configuration files as needed.<br>
 +
Copy two configuration files from /dmm/config/app-test to /dmm/release/bin
 +
 +
    #sudo cp /dmm/config/app_test/module_config.json /dmm/release/bin
 +
    #sudo cp /dmm/config/app_test/rd_config.json /dmm/release/bin
 +
 +
After building the DMM, sample test apps will be generated like below, '''vs_epoll''' and '''vc_common''' <br>
 +
 +
Note:  Currently we '''only''' support '''dmm + kernel stack''', export NSTACK_LOG_ON=DBG
 +
 +
Examples:
 +
 +
server vbox:
 +
 +
    #./vs_epoll -p 20000 -d 172.16.25.125 -a 10000 -s 172.16.25.126 -l 200 -t 5000000 -i 0 -f 1 -r 20000 -n 1 -w 10 -u 10000 -e 10 -x 1
 +
 +
client vbox:
 +
 +
    #./vc_common -p 20000 -d 172.16.25.126 -a 10000 -s 172.16.25.125 -l 200 -t 5000000 -i 0 -f 1 -r 20000 -n 1 -w 10 -u 10000 -e 10 -x 1
 +
 +
 +
If users want to test their own apps, copy program executable in app-test directory. Run program using LD_PRELOAD from app-test directory
 +
 
 +
  # sudo LD_PRELOAD=../../release/lib64/libnStackAPI.so ./your_app   
 +
 +
5. Post testing cleanup. After completing testing, issue the following command to clean up the vagrant VM's:
 +
 +
  # vagrant destroy -f
  
 
== Get Involved ==
 
== Get Involved ==
Line 55: Line 117:
 
* [[IRC | Join fdio-dmm IRC channel]].
 
* [[IRC | Join fdio-dmm IRC channel]].
 
* [https://git.fd.io/dmm/tree/ Browse the code].
 
* [https://git.fd.io/dmm/tree/ Browse the code].
 +
* [https://wiki.fd.io/view/DMM/workshop 2018/6/26 DMM China Seminar].
 
* [[Projects/dmm/Release_Plans/Release_Plan_18.04 | 18.04 Release Plan]]
 
* [[Projects/dmm/Release_Plans/Release_Plan_18.04 | 18.04 Release Plan]]
 +
* [[Projects/dmm/Release_Plans/Release_Plan_18.07 | 18.07 Release Plan]]
 +
* [[Projects/dmm/Release_Plans/Release_Plan_18.10 | 18.10 Release Plan]]
 +
* [[Projects/dmm/Release_Plans/Release_Plan_19.01 | 19.01 Release Plan]]
 +
* [[Projects/dmm/Release_Plans/Release_Plan_19.04 | 19.04 Release Plan]]
  
==Start Here==
+
==Reference==
  
How to submit a patch to DMM - [https://wiki.fd.io/view/VPP/Pulling,_Building,_Running,_Hacking_and_Pushing_VPP_Code#Pushing Pushing Code with git review]
+
If you are using Chrome, you can install extension "'''Markdown Reader'''" to read the MD file.
 
+
==Reference==
+
  
Enabling “Protocol Routing”: Revisiting Transport Layer Protocol Design in Internet Communications - [http://ieeexplore.ieee.org/document/8114687/]
+
[http://ieeexplore.ieee.org/document/8114687/] Enabling “Protocol Routing”: Revisiting Transport Layer Protocol Design in Internet Communications, By Dr. '''Kai Zheng''' November 2017 <br>
 +
[https://wiki.fd.io/view/File:DMM_-_2018_FDIO_Mini_Summit_@LA_v2.pdf] Empower Diverse Open Transport Layer Protocol in the Cloud Networking, By '''Qing Chang''' At FD.io ONS2018 mini summit <br>
 +
[https://www.youtube.com/watch?v=Iu-O8e6Q6-E] Video presentation, By '''Qing Chang''' At FD.io Mini-Summit at KubeCon + CloudNativeCon EU 2018 <br>
 +
[https://git.fd.io/dmm/plain/doc/DMM_DeveloperManual.md] Doc: DMM Developer Manual <br>
 +
[https://git.fd.io/dmm/plain/doc/Build_DMM.md] Doc: How to build DMM <br>
 +
[https://git.fd.io/dmm/plain/doc/TestAppUserGuide.md] Doc: How to test APP linked with DMM <br>
 +
[https://wiki.fd.io/view/File:FD.io_-_The_Universal_Dataplane_v2.pdf] Ray kinsella introduce fd.io and DMM in NJ summit <br>
 +
[https://wiki.fd.io/view/File:02_ld_Light_A_scalable_High_Performance_and_Fully_compatible_TCP_Stack.pdf] Li Dan introduce Light and DMM in LC3 DMM workshop <br>
 +
[https://wiki.fd.io/view/File:F-Stack.pptx] Wang Hailong introduce F-stack and DMM in LC3 DMM workshop <br>

Latest revision as of 18:13, 15 December 2020

DMM Facts

Project Lead: George Zhao <George.Y.Zhao@huawei.com>
Committers:

  • Florin Coras <fcoras@cisco.com>
  • Ray Kinsella <ray.kinsella@intel.com>
  • Junfeng Li <hotjunfeng@163.com>
  • Hailong Wang <whl739@163.com>
  • Kai Zheng <kai.zheng@huawei.com>
  • George Zhao <George.Y.Zhao@huawei.com>
  • Qing Chang <Qing.Chang1@huawei.com>
  • Liuxiaoxu <liuxiaoxu1@huawei.com>
  • Yalei Wang <william.wangyalei@huawei.com>
  • Swarup Nayak <swarup.nayak1@huawei.com>
  • Satish Karunanithi <satishk@huawei.com>

Repository: git clone https://gerrit.fd.io/r/dmm
Mailing List: dmm-dev@lists.fd.io
Jenkins: jenkins silo
Gerrit Patches: code patches/reviews
Bugs: DMM bugs

What is DMM

DMM (Dual Mode, Multi-protocol, Multi-instance) is to implement a transport agnostic framework for network applications that can

  • Work with both user space and kernel space network stacks
  • Use different network protocol stacks based on their functional and performance requirements (QOS)
  • Work with multiple instances of a transport protocol stack

Use and engage or adopt a new protocol stack dynamically as applicable.

DMM Project Proposal

History

Emerging applications are bringing extremely high-performance requirements to their network systems. Eg. AR/VR, IoT etc. Many of these applications also have unique demands for QOS/SLA. Some application need a low latency network, some need high reliability, etc. Though such performance targets should be required for the complete communication system, the transport layer protocols play a key role and encounter a relatively higher challenge, because traditionally the TCP-based transport layer exploits the “best-effort” principle and inherently provides no performance guarantees. However, as internet applications rapidly grow and diversify, an all-powerful or one-size-fits-all protocol or algorithm becomes less feasible. Thus the traditional single-instance TCP-based network stack faces many challenges when serving applications with different QoS/SLA requirements simultaneously on the same platform.

Moving the networking stack out of the kernel is an occurring trend in both industry implementations and academic literature. Technologies like DPDK (and others) are improving performance of the network stack, bypassing the kernel and avoiding context-switching and data copies, as well as providing a complete set of packet-processing acceleration libraries.

Keeping the above trends in mind the DMM/nStack provides a framework where, system operators can plug in dedicated types of networking stack instances according to performance and/or functional requirements from the user space applications. The application need not consider changes to the transport layer API. A lightweight nStack management daemon is responsible for maintaining the stack instances and the app/socket-to-stack-mappings, which are provided via the orchestration/management interface.

DMM Architecture

DMM Simple Arch


The DMM framework provides posix socket APIs to the application. A protocol stack could be plugged into the DMM. DMM will choose the most suitable stack according to the RD policy.

Quick Start

With this quick start, users can have a simple understanding of DMM. Following the quick start, it will set environment, build DMM and test apps automatically on Vagrant.
The test topology will typically look like this. It uses DMM + kernel stack now. Make sure vagrant and virtual box have been installed first.

DMM Simple Test topology


1. Download dmm code

   # git clone https://gerrit.fd.io/r/dmm

2. Setup Vagrant VMs

  # cd dmm/resources/extras

If users want to change the DISTRO, NICS, VMCPU or VMRAM settings, modified the env.sh and run source env.sh.

  # vagrant up

This steps will take a while. Until vagrant automatically set the environment and compile dpdk, glog and DMM.

3. Log in to server/client in two different terminals.

  # vagrant ssh dmm-test-server
           and
  # vagrant ssh dmm-test-client
    

4. Users can test sample client/server app in /dmm/release/bin.

Before testing apps, users can update configuration files as needed.
Copy two configuration files from /dmm/config/app-test to /dmm/release/bin

   #sudo cp /dmm/config/app_test/module_config.json /dmm/release/bin
   #sudo cp /dmm/config/app_test/rd_config.json /dmm/release/bin

After building the DMM, sample test apps will be generated like below, vs_epoll and vc_common

Note: Currently we only support dmm + kernel stack, export NSTACK_LOG_ON=DBG

Examples:

server vbox:

   #./vs_epoll -p 20000 -d 172.16.25.125 -a 10000 -s 172.16.25.126 -l 200 -t 5000000 -i 0 -f 1 -r 20000 -n 1 -w 10 -u 10000 -e 10 -x 1

client vbox:

   #./vc_common -p 20000 -d 172.16.25.126 -a 10000 -s 172.16.25.125 -l 200 -t 5000000 -i 0 -f 1 -r 20000 -n 1 -w 10 -u 10000 -e 10 -x 1


If users want to test their own apps, copy program executable in app-test directory. Run program using LD_PRELOAD from app-test directory

  # sudo LD_PRELOAD=../../release/lib64/libnStackAPI.so ./your_app    

5. Post testing cleanup. After completing testing, issue the following command to clean up the vagrant VM's:

  # vagrant destroy -f

Get Involved

Reference

If you are using Chrome, you can install extension "Markdown Reader" to read the MD file.

[1] Enabling “Protocol Routing”: Revisiting Transport Layer Protocol Design in Internet Communications, By Dr. Kai Zheng November 2017
[2] Empower Diverse Open Transport Layer Protocol in the Cloud Networking, By Qing Chang At FD.io ONS2018 mini summit
[3] Video presentation, By Qing Chang At FD.io Mini-Summit at KubeCon + CloudNativeCon EU 2018
[4] Doc: DMM Developer Manual
[5] Doc: How to build DMM
[6] Doc: How to test APP linked with DMM
[7] Ray kinsella introduce fd.io and DMM in NJ summit
[8] Li Dan introduce Light and DMM in LC3 DMM workshop
[9] Wang Hailong introduce F-stack and DMM in LC3 DMM workshop