First draft for release D roadmap

Change-Id: I7c3785ed37e8287a2869eb97206aabc1fb87cc6a
Signed-off-by: Yujun Zhang <zhang.yujunz@zte.com.cn>

1 files changed, 100 insertions, 0 deletions

diff --git a/docs/roadmap.rst b/docs/roadmap.rst
new file mode 100644
index 00000000..42caec92
--- /dev/null
+++ b/docs/roadmap.rst
@@ -0,0 +1,100 @@
+.. two dots create a comment. please leave this logo at the top of each of your rst files.
+.. image:: ../etc/opnfv-logo.png
+  :height: 40
+  :width: 200
+  :alt: OPNFV
+  :align: left
+.. these two pipes are to seperate the logo from the first title
+|
+|
+
+Roadmap for Release D
+=====================
+
+The development of QTIP has been paused after Brahmaputra release due the
+shortage on resource. We will skip Colorado release and target for Release D.
+
+The project will stick to the original scope as a benchmark platform and
+continue to develop on existing framework.
+
+QTIP will focus on
+
+- integrating more benchmark tools
+- supporting new technology applied in OPNFV
+- improve the result dashboard for better visualization
+
+Besides the technical parts, QTIP is also aiming to attract more contributors by
+
+- providing more comprehensive documents
+- refactoring source code for better maintenanability
+- improving development management for better collaboration
+
+Framework
+---------
+
+Error Handling
+^^^^^^^^^^^^^^
+
+The QTIP will be used against different environment. It is not possible to run
+without any error all at once.
+
+We will not be able to get rid of errors, but we may handle them gracefully.
+
+Comprehensive error messages will help to locate the issue quickly and help user
+to resolve them.
+
+Declarative Playbook
+^^^^^^^^^^^^^^^^^^^^
+
+QTIP uses ansible for setting up the environment. It is nice and powerful tool
+we will keep for Release D.
+
+However, existing playbooks is full of hardcoded shell scripts which sometimes
+will fail in specific OS distribution.
+
+Although most system administrators will be familiar with shell script, it is
+not easy to tell the purpose of a long command line at a glance.
+
+Ansible's solution for these issues is to provide modules as an abstract layer
+to handle the devergence, and it will also be more compact and easier to
+understand. This is something we should leverage.
+
+Scenario Configuration
+^^^^^^^^^^^^^^^^^^^^^^
+
+Currently the scenario configuration is hard coded and not able to be run under
+different environment. The variables should be separated from the configuration
+template.
+
+Features
+--------
+
+Benchmarks
+^^^^^^^^^^
+
+1. vswitch perf
+2. Cyclictest
+3. Stress
+4. Lmbench
+5. Sar
+
+Technology
+^^^^^^^^^^
+
+Some new technology is introduced into OPNFV and it would be good if we can
+support them at the first time.
+
+- SR-IOV: a key technology to improve network performance in VM and the VM can
+  achieve nearly physical NIC's performance.
+- DPDK: a key technology to improve the NIC's performance through poll mode
+  which dismisses physical interrupt as less as possible. The byproduct of DPDK
+  is nearly 100% CPU usage. It can also be used in VM.
+
+Development Management
+----------------------
+
+We will make improvement on development management
+
+1. Continuous Integration
+2. Documentation
+3. Issue Tracking