There were several years before and after OpenStack, but this year the OpenStack Summit, held in Austin, USA, was the first time I took part. Back, has been in the aftertaste, feel to write something, will I at this summit of observation and thinking briefly summed up, whether right or wrong. Overall, my observation of the OpenStack summit is that OpenStack as IaaS is already in the past, and Solutions on OpenStack is now and in the future.
The basic situation of Austin Summit participation degree
Observe 1:openstack is committed to becoming the infrastructure platform for running a variety of cloud Solutions 1.1 OpenStack two stages of development
I have probably drawn a diagram:
This figure shows the main points:
- OpenStack has developed into its 13th release this year, and its development process can be broadly divided into two phases: the first phase of OpenStack as IaaS, and the second phase Solutions on OpenStack.
- In its first stage of development, OpenStack has been given considerable opportunities, capabilities and recognition as an IaaS. Over the past few years, OpenStack has been consolidating its basic capabilities as an IaaS, with several core modules maturing, its growing role as an iaas, especially as the most exclusive open source option for the private cloud, and we all believe it is a very successful development. , and we are increasingly seeing that OpenStack has begun to exert its force on public clouds as several key issues that have previously hindered it from becoming public clouds have been or are being resolved gradually.
- The key to OpenStack is that its second stage of development is whether it will be able to do so if all cloud solutions have chosen it as the carrier of the operation. If competent, then it will be successfully landed in a variety of enterprises, services for a variety of enterprise applications, to help enterprises on the cloud; if not, there may be a new solution to replace it.
Various applications can be broadly divided into two categories:
- The first, OpenStack ready applications, which are currently running on the OpenStack platform, include developing test environments, enterprise support applications, and some peripheral business applications. The OpenStack platform for this part of the application is also the main target market for the major OpenStack startups today. This part of the application is now able to cloud, because they are not very high demand for the cloud, and both suppliers and enterprises have a considerable accumulation of their needs and needs to solve the problems are very familiar.
- In the second category, OpenStack to is ready application, including Big Data, NFV, Internet of things, blockchain, core trading system of financial Enterprise, core website of e-commerce enterprise and so on. This part of the application of OpenStack is still in the test phase, not large-scale rollout, the main reason is that openstack from the demand for these applications there is still a large gap, even if there are sporadic user cases, but most of them are not replicable.
1.2 Austin summit on the various related performance
With that in view, looking back at the Austin Summit, the different people who do OpenStack may have different feelings:
- Those who focus only on the first stage will think that the summit has nothing new to do with the previous meeting, and that speculation has arisen. The reason is also very simple, that is, OpenStack as the first type of IaaS support application is already very mature, there is nothing new to talk about. To talk about, is nothing more than to say how many customers I have, how good each customer, how many nodes per deployment, what a great improvement, how much code to contribute to the community and so on. There is nothing new about these things.
- People who focus only on the second stage think that OpenStack has developed into a bottleneck because the current capacity of OpenStack is not a little bit different from the real needs of these applications. These new applications give the opportunity, but OpenStack has limited capacity and is not yet able to catch the opportunity, so the pessimistic rhetoric appears.
- The two-stage focus will be on the one hand to cheer OpenStack's achievements in the past, while encouraging the bright future. This is an optimistic argument, and it will be the right argument.
Different people with OpenStack also have different feelings:
- Customers and startups who are more concerned with the first category of applications are happy because customers see the maturity of OpenStack in this respect, the mind will gradually eliminate the worries, start-up companies will be better to sign the list.
- More attention to the second type of application of the customer some anxiety, because the gap is not generally large, and the difficulty is not generally large, to change the code and to add a lot of code, even the structure to adjust, progress is relatively slow, so they are very concerned about the community in this progress, and even some worry. In a booth, I saw a manufacturer who did VNF was anxious to ask if a feature of neutron had been implemented.
Fortunately, the OpenStack community is well aware of the current status and future of OpenStack, as can be seen from what they talk about:
- Combining four types of cases, let customers tell their stories, respectively, to demonstrate the maturity of OpenStack for the first class of applications, to enhance confidence in the present, to demonstrate the forward-looking and control of the second type of application, to enhance the confidence of everyone in the future.
- Three main themes: User experience, manageability, performance, stability, and extensibility. The OpenStack community faces these existing gaps and directs us to work together to solve them.
- Emphasizing collaborating with adjacent communities,interoperability and integration, because the second type of applications are not common enterprise applications, often have the open source community to promote its standards, therefore, The OpenStack community needs to work closely with these communities, and the prospect is to build openstack into a integration engine that supports different user scenarios and solutions.
- Emphasizing one platform for VMs, containers and bare metal (both virtual machines, containers, and physical machines, three pillars), the only way to meet the diverse needs of different applications.
- Put forward the certified OpenStack Administrator program, because OpenStack as IaaS has become a de facto standard in the private cloud, allowing its administrators to begin textual research. It is also a source of income for startups and a reward for the vast majority of people who try to be a community.
- Demonstrate and discuss new solutions, especially public cloud, IoT, NFV, etc. at the current stage of development, and estimate the next blockchain.
1.3.2 problems that still exist
- There is a lack of standard reference architectures that are replicable. Most of the user stories that were shared at the summit were the result of extensive revisions and customizations made by their own teams in the community version of the code. Personally think, when Mirantis such a company won the award, the standard participation in the structure will have.
- There is still a big gap in RAS (reliability, availability, performance, security, etc.) from the production environment.
- Further enhancements are needed for container and bare-metal support, and OpenStack has been too focused on virtual machines and neglected two of the other.
- How to better work with other big open source and standard communities. In the next phase, the development of OpenStack is no longer dependent on its own development, but on the development of the whole ecosystem.
Observation 2: The performance of the major summits participating in the enterprise
- HP: It is said that HP's performance in the past few years is not the same, the momentum is much smaller. The possible reason, on the one hand, because of its OpenStack-based public cloud closed so the topic is less, on the other hand is a lot of Core Memeber lost so can speak less people, and then its Helion sell is not good, can tell the case is not much. After listening to one of their lectures on NFV, we talked about the various shortcomings and possible solutions of OpenStack in NFV support, although it was a good summary, but relatively general.
- RackSpace: The momentum is great, but the main thing is talking about the first stage.
- RedHat: More emphasis was placed on the second phase, mainly NFV, including achieving fine-nine VNF reliability in a telco-grade OpenStack cloud,designing f Or nfv:lessons learned from deploying to Verizon,ipv6 capabilities in TELCO environments; compared with the past, more attention to publicity sessions K + OpenShift as a complete Stack to support a variety of solutions.
- Huawei : This momentum is generally about something that is related to NFV, including OpenStack Tricircle/multisite, Network (Service chain,dragonflow), Container ( Magnum,kuryr), NFV and other content. It should be said that Huawei has done very well in the OpenStack as IaaS level, and there are many cases of propaganda, but why is it just about NFV? Over the past few years, Huawei has focused on the IaaS layer, which may be limited in areas other than NFV in its own domain. For Huawei, their energies need to be placed more on PaaS and SaaS.
- Easystack and UnitedStack: First of all, congratulate them on entering the Community Gold membership list, from the sessions they do, their main focus on the first phase of the application, how in the future in the second phase, is the test of how they will develop in the future.
- VMware: A bit loud, but very small. The development of OpenStack in the second phase will determine how it will go beyond VMware, so these new solutions are basically not playing with VMware.
- Mirantis: In both phases, and in the new applications of the community such as IoT and NFV have been laid out, and through and Ucloud became a joint venture company Umcloud to the Chinese market. Optimistic about its development prospects.
- T (telecom operator): There has been no small progress in NFV, with dozens of talks on NFV, about its needs, gaps, plans, prospects, etc. The company also received the "Super Star" award for its NFV project, and their NFV was part of the production environment.
- China Mobile: also in the promotion of NFV and public cloud, listening to their lectures on the public cloud to practice the experience of OpenStack, in the domestic large state-owned enterprises should be the front.
- traditional telecoms companies such as Nokia and Ericsson : These traditional telecoms companies that had nothing to do with OpenStack have finally found their reason to participate in the OpenStack summit at NFV.
- TCP Cloud: They were impressed by the presence of IoT environments deployed on the Keynote session.
OpenStack Austin Summit observation: OpenStack as IaaS is the past, Solutions on OpenStack is the future