The hottest Sydney summit openstack application Ro

  • Detail

Sydney Summit: openstack application road

the infrastructure of cloud has begun to take shape. How to provide a better scheduling environment for various applications on openstack

at 9 a.m. local time on November 6, Sydney, Australia, the 16th openstack summit was grandly opened at the Sydney International Convention Center. More than 2300 participants from 52 countries around the world will have a three-day discussion on the technology and business practices related to open infrastructure with openstack as the core. This article is one of the discussions of the second day

in the openstack user survey in April 2017, we absorbed the design concepts and structural characteristics of famous relaxation experimental machine companies such as Germany and Italy. It can be seen that in addition to being regarded as infrastructure services, openstack is also used for architecture testing and continuous integration, databases, network services, big data, etc. In addition, the recently popular edge computing and IOT also use openstack as the development and research platform

the picture is from openstack official website

at the last summit, it started with the official addition of an openstack policy, which told openstack that it must pay attention to application requirements (), and then at the Sydney summit, various discussions and requirements began to ferment in many projects. We continue to discuss this application in depth

architecture change

in order to meet the needs of applications, the design planning of application credentials has been carried out since the last summit, so that applications can use the permission obtained by keystone (verifier) to run openstack services. Applications can operate openstack services through this permission, but only within the scope allowed by this permission. Therefore, permission management will not be destroyed

during Denver PTG, there has been a technical detail discussion (). At the summit, the application credentials feedback technical agenda mainly reports the specifications currently formulated and collects feedback

this function enables applications to use some openstack services without using user information for authentication. You can use the service by directly passing the application credential to the called service. The scope of services that can be used will be set when the application credential is created

for the modifications discussed during PTG, the feedback is positive. Generally speaking, credential will not be used to re create other application credentials. However, for the requirements of scheduling projects (for example, when automatic capacity expansion, heat needs permission to create new capacity expansion resources, like the current use of trusts mechanism)

why does the author put this function in the architecture change block? In fact, the report made at the last summit also mentioned that surprisingly, the permission structure is the most outpost of change in the current community, because only after permission is opened, can other services start planning the design of subsequent applications

another change belongs to the issues brought about by the cloud native design/refining across openstack (Part II) technology agenda

cloud native, although the vocabulary has always existed, the transformation is regarded as the rise of CNCF (cloud native Computing Foundation) and containerization to improve the container architecture, and the stable Cloud Architecture brought by openstack after its maturity. The combination of the two began to be valued. For newly developed applications, directly let applications use services on cloud framework and container architecture (such as automatic capacity expansion, monitoring, recording, repair, etc.)

in the agenda, review the existing design of openstack and discuss how to make openstack architecture have the advantages of cloud architecture. Undoubtedly, for the application program, the subsequent feasible design will not destroy the use of the existing program (openstack cross version compatibility to the automatic clamping and loosening device, sample clamping force: 690kpa is a major advantage)

therefore, the first knife falls on condition monitoring. The focus of the entire technical agenda is how to establish a condition collection that conforms to the Cloud Architecture ()

the design focus is to provide a platform for all projects to upload their own resource status. Establish a status return tool in ddeware by planning. This tool enables all projects to send all resources that they want to drop into the status list to a unified location

later, each service will choose how to distinguish these states. There will be such a design to provide a platform that can look at the overall openstack status, so that the resource status is not scattered, and there is no need for separate services to independently develop resource status monitoring. Compared with efficiency and unity, it will be much higher. This is the focus of this link as the first knife. At present, there is no relevant practice, but we must be optimistic about this function in the future. This is a good development policy for improving the problems of composite materials such as pa6/montmorillonite, pet/montmorillonite, pmma/montmorillonite, pi/montmorillonite, ep/montmorillonite, ps/montmorillonite and so on

Orchestration: the road to automation of applications

automation is the demand of many users, which requires the interaction of the following links: generate resources - 《 monitor resources - 》 signal trigger events - 《 repair - 》 continuous monitoring. As long as it can meet the above ring. Automation can be achieved. The biggest problem is what services should be used in each link, and whether there is a good serial connection between services to achieve the most efficient automation system

auto scaling is not a new technology in openstack environment, and its operation maturity is also quite high. Therefore, most users share at the summit. Automatic repair is one of the small focuses of the summit. In the technical agenda self-healing and optimization SIG, it is the first time to gather relevant technical experts at the summit to discuss the current repair functions and how to optimize them

at present, heat has done in-depth research on automatic repair in the community, and also provides ways for users to refer to


of course, we hope that the automation we can provide to users is more comprehensive. Therefore, the task of this sig group has just begun, with a considerable number of people participating and many volunteers willing to promote it together

at present, it has been decided to establish this group. There will also be proprietary IRC channels and conferences in the future. More importantly, through the group flag, sign and collect more user needs (-forum)

automated processes also have other standard protocols such as ifttt. It is worth mentioning that in the practice of repair, it is suggested to consider the digital display fully automatic impact testing machine generating signal Mistral service through the high-speed load measurement sensor, and apply openstack to provide process management service during the automatic process, so that the operation can act according to the actual situation process

consider using heat as resource management, while Mistral as operation process management during management, which is a good architecture in many practical use cases. In the agenda Mistral - project update, Mistral PTL introduces the current situation and future development of Mistral

in the above automated process and application management, orchestration is a way to simplify the complexity of user-side operations. Often, applications are composed of several services. In most actual operating environments, applications may have to be scheduled on hundreds of nodes. Therefore, it is more necessary to unify the management mechanism, increase the management intensity, speed up the scheduling process and reduce errors. At the summit, the author was lucky to be responsible for arranging several community technology agendas for the project

on the morning of the first day of the summit, there will be an onboarding technical agenda for arranging the project. Provide project introduction, framework, script explanation, debugging methods and contribution methods. Let developers and operators connect more closely with the community. In order to help the application connect, the content uses the automatic repair and expansion script and the software config script as the commentary direction. So that operators writing scripts can develop cloud native environment for applications more quickly

in the project update section, introduce several new script functions (list_concat_unique, contains, make_url) and new resources of the pike version (including: OS:: neutral:: trunk, OS:: neutral:: segment, OS:: zaqar:: subscription, OS:: zaqar:: mistraltrigger, OS:: magnum:: cluster, OS:: magnum:: clustertemplate, OS:: Mistral:: externalresource, OS:: Zun:: container). Other mentioned functions include that the project itself supports python3, supports the update of choreography according to the actual situation, adds a new heat agents sub project, and improves the distributed services

the update of zaqar resources can be observed in order to improve the serial connection of signal triggering events in application management automation

in addition, OS:: Mistral:: externalresource allows you to use Mistral's workflow to define each independent action of resources (such as create, update, delete, etc.). When an application has resources that do not belong to openstack and attempts to use openstack architecture as a scheduling platform, this resource can greatly assist in the simplification and orchestration of related processes. Of course, even without Mistral, you can put customized resources into the heat script by adding Python source code or registering new resources. Therefore, you can choose the method according to the needs of the application

from the agenda of the summit, it is not only these agendas that the author came into contact with, but also other agendas related to applications and application choreography. In terms of the overall openstack agenda, there are not a few. Many practical cases of applications such as network scheduler or container service requirements are no longer shared. Openstack's power and development needs to assist applications seem to be many. The focus of the service is to meet the needs of users. From the first day of the summit, we can see that the community is indeed moving in this direction

as for whether users are also full of expectations and confidence in architectural applications? The following author gives you a reference

on the agenda future science on future openstack, we should let you know that ska has decided to cooperate with CERN and plans to create the world's largest cloud environment for scientific research

of course, this huge stack of applications is about to be built on openstack. Later, I chatted privately with several speakers. They believed that openstack, as an underlying platform, would achieve high availability and reliability for application support at the beginning of the plan

at this summit, openstack is more specialized in the goals of applications. Compared with the early divergent and numerous development projects, the stability of the current situation makes it easier for developers to cooperate with each other. The rise of other communities, on the contrary, allows the openstack community to see more opportunities and can indeed lead applications to success

Copyright © 2011 JIN SHI