A Free Trial That Lets You Build Big!
Start building with 50+ products and up to 12 months usage for Elastic Compute Service
In fact, for the different needs of enterprise users in the application management solution deployment and delivery model, Blue Ocean provides two product and service delivery modes: The traditional software licensing model (Blueware) and the Cloud based SaaS model (ONEAPM). Blueware mainly for traditional large enterprises, while ONEAPM mainly for Internet enterprise developers.
With ONEAPM, developers can increase development speed, release applications faster, and pinpoint bottlenecks in the usage experience. According to the Blue Ocean News introduction, ONEAPM's function is very suitable to support the agile development model. They want to provide a service, let the application of management is no longer a problem, for developers, you only need to do a good job, the service you want to provide to customers think well, do not need to care about the performance of the problem, this article will be ONEAPM related issues to interview the founder and chairman of the Blue Ocean News He Xiaoyang.
He Xiaoyang, founder and chairman of Blue Ocean News
CSDN: First of all, why do you want to do APM, what is the significance?
He Xiaoyang: My personal microblog may often see such information, there are users spit @ service provider, said your performance experience is not good, I paid the money on the net I did not receive tickets, this kind of thing is very common. Another example is to say that an app used to play an error, pop-up line of code, think this thing is very strange, this line of code what meaning? He was puzzled by the average user. So this thing is obviously not the developer of this application he wanted to see, but he didn't know such a thing happened, so for them, the first will lead to a decline in user experience, and second, consumers directly choose another product. Now especially in China 2C such a field of competition is very fierce, whether in any field, such as online education or the recent fire of the tourism industry, competition is very fierce, if the experience is not good, the end user may choose others, which will directly lead to the decline in revenue.
We want to provide a service, we want to make application can no longer be a problem, for developers, you do your business well, the service you want to provide to customers think well, your business innovation think well, do not need to care about performance issues, we provide this applied to manage total Solution, We are helping all our clients to solve these problems in a unified way. We oneapm this SaaS platform, facing the small and medium enterprises, as well as the entrepreneurial team, which in the United States called Sgt to middle business, in the domestic feel that may be smaller, or even personal developers. As long as you are an entrepreneur, have your own ideals, dreams, you can write a line of code, write an application, we can ensure that your app configured is no problem.
So what are we doing now? We want to give all developers the ability to stop worrying about performance, no matter what code he writes, and use ONEAPM to quickly guarantee that the customer experience will not be compromised by the application's performance problems. It's the first thing we do, and the thing we've done, is how much he's capable of, what he can do, and not because of his lack of it capability that makes the business capacity fall.
What is the main difference between CSDN:APM and traditional operation and maintenance monitoring? such as IBM provides CA, HP's, also is to monitor the infrastructure.
He Xiaoyang: In fact, the basic monitoring facilities of the products we have, but in fact, the basic monitoring of the problem where? This should be said to be related to the APM industry development process, we take the American, the earliest time in the United States before 1998 or 1995 to 2000 such a point of time, This time we know the age of the Internet 1.0, is the site, basically like Yahoo is that time rise, is a website, without any interaction and no business logic in, such as Yahoo to provide you with the news, you only go up to see the copy, you want to have some interaction with it is impossible. This time the speed depends on your network, so at that time some of the application of measuring speed, and today they say they are doing APM. From 2000 to 2005 the point of Time, the computer from the site into a real business, at that time, such as ebay manufacturers can rise, users in the Web site to create the actual business, at that time the application run in where? is the frame. We know that sun is very valuable at that time, selling servers, minicomputer, it sells servers and minicomputer, this performance is naturally need to monitor, infrastructure monitoring is probably the American in 2005 to do a thing, that time in the sun's small machine running database software, professional software, run WebLogic Some software such as server. And then, from 2005 onwards, creator became the core of application, with the advent of virtualized cloud computing, with the landing of SOA and the application of devops, this time monitoring infrastructure is no longer used, because many failures occur within the application, So the third generation of application-oriented monitoring is the application-centric, you have to pay attention to the user experience, emphasis is on end-to-end application performance management.
CSDN: But like CA, IBM are traditional software, you just mentioned APM including BSM they are doing, you now do APM, where is the advantage?
He Xiaoyang: Yes, BSM now no manufacturers can do it, APM they are doing, in fact, the APM technology itself is changing, the previous generation of APM even if we say can perceive the user experience, can do End-to-end, 2008 APM is now very different. 2008 APM is the third generation of APM software, in fact, to do end-to-end application management, at that time what is the IT infrastructure? First of all, the data is vertically expanded within the framework, not too much virtualization, not too many clouds, are deployed on the minicomputer, so this application is up. This kind of application is often placed on Java and. NET, because the rise of these two causes the container is also very expensive, we know that Bea sold 8 billion dollars, this price now want to be sky-high, incredibly and splunk, incredibly expensive than sun. So in a very large and expensive small machine, running a lot of expensive such as WebLogic application Server, this time APM, whether IBM, HP HP SS to do things, are such a solution, is for such a scenario to do. But the problem is, now the application is far from this, now applications are distributed, are virtualized architecture, such as IBM's Itcam, a general set of license costs in the United States are also more than 1 million dollars, because it is very heavy, Its UI, and so on, is geared toward heavy enterprise-class applications and is geared towards a minicomputer environment, so there is no way to face SMB customers. For individual developers, they have no way to use such a software, the first to see, and then can not afford to buy, even if the purchase will find that this thing may not conform to the way he uses, because the traditional APM is supporting the Java language, but now the program is written with what? Many programs are written in Fast Ruby, or on the phone, with Android or iOS, and you don't have a concept like IBM that doesn't support this application, so you can't use it. So the next generation of APM is based on a distributed cloud framework that provides a relatively simple, easy-to-use APM solution, which is probably the feeling.
CSDN: Can you share some of the specific ideas, technologies, and infrastructure that ONEAPM is building on your core products?
He Xiaoyang: The main use of AOP (aspect-oriented programming) programming ideas, the application of all the monitoring logic abstraction, so that users focus on writing their own business logic, in the system startup, through the corresponding technical means to the application of the logic code of monitoring and then woven into the user's application, so as to add application monitoring functions. In Java, for example, in the Tomcat container, the user downloads the ONEAPM probe and adds the-javaagent:/path in the parameters of the Tomcat Java Virtual machine system. Oneapm.jar, and then start Tomat, during Tomcat's JVM startup, the JVM uses the BCI (Byte Code instrument) technology to weave the corresponding application monitoring logic into the application Java class file while loading the corresponding application class. To complete the monitoring of the application.
ONEAPM Technology Stack:
Subsequent technologies: Strom, Kafka, Flume, Elasticsearch, Hadoop and Spark
ONEAPM Architecture (click to enlarge)
CSDN: Can you introduce the fact that there are a lot of startups in foreign countries that are already on the rise, and have got a good development, and there are a few at home now? Can introduce the threshold in the field of APM in the end how high, you in the entrepreneurial time have some difficulties encountered?
He Xiaoyang: First of all, you do the backend monitoring first concept, since you want to do with the application, then your implementation level is lower than the application level, to a bit more low-level. We are in fact and foreign wily technology, AppDynamics are the same, we are all implemented in the language level, such as the Java language, you have to write a probe. If the user is. NET language, your Java probes are completely useless, and you have to write a new one. NET probes. How many languages are there now?. NET, JAVA, PHP, Ruby, Python, Nodejs, go, Lisp, and so on need each language to write a probe, that this probe and not say how difficult, first put each language together into a team, a language a team, Then you need a lot of teams on the back end, plus iOS, Android, for 10 research and development teams, you think about 10 research teams for an entrepreneurial company, and there are people in every language, so China itself is a very difficult situation for developers, and it's hard for you to recruit a technical partner in a language, and you need to call all 7 languages. , it is really very difficult thing, this is the first threshold.
The second threshold, after the probe is finished, starts to collect the application data, but the amount of the data, you can imagine how much this amount will be. We're actually doing SaaS. Our backend servers receive data from our users, each of his actual browsers and mobile phones, and our concurrency is essentially the sum of our user concurrency. Like Sina, everybody, Jiuxian net is our customer, if say your concurrent amount of the sum of their concurrent quantity, can imagine the back-end server to bear how much burden. Let's take a look at the new relic. The event received is 2 trillion data per day, you need the ability to compute and store very very strong, that is, as early as the year before last, the new relic is far more concurrent traffic than Twitter, now should be many times Twitter, How much money do you want as a platform for Twitter, and how much manpower you need to calculate how hard it is to deal with back-end services as a far more than Twitter, so the probes will get stuck in a lot of teams, and you have to do the back end after you do the probes.
and 3rd. Even if the probe is made, he can withstand the test in extreme circumstances, we know that China's IT environment is very complex, the highest in the world, the application of the world's most vulnerable, because most of our programs are fresh graduates write, and even the students have not graduated from writing. Writing an application-level probe in such an environment can easily put someone else's system down or take up a lot of resources, and you need to verify that your probe is OK in many environments, which is especially difficult for many teams.
CSDN: What is the difference between the same founder, Wily and New relic? And, for example, are SaaS environments, just when new relic and appdynamics enter China, what's the difference?
He Xiaoyang: First, the difference between New relic and appdynamics, essentially they are APM products, but their positioning is not the same, New relic its own positioning is to provide developers of the application services, developers must be individuals. AppDynamics refers to large, medium-sized Internet companies and traditional companies. There is also a company in the United States called Compuware, its customers are like the world's top 2000 customers, in fact, our customer base is Compuware plus new Relic, my SaaS service to all developers, my enterprise-class Sorftware service belongs to the largest enterprises in China, We don't do large internet companies in the middle. Product positioning is different, your interface, your UI, your experimental details will be very different, this light and heavy is a difference. As for the fact that they are in China, we have been in the U.S. SaaS service, like Zendesk has been used, the United States a call SaaS billing Zuora we also use. Zendesk access is very slow, our SaaS Help Center is done with Zendesk, users visit our site quickly, but when he makes a ticket, or goes to a document, it becomes very slow.
CSDN: Do you think that the two different groups of developers and enterprise customers will be different in the promotion?
He Xiaoyang: For developers, we want to let people know more about this thing, we think they need this, but they may not know this, so for developers, we just let them know that there is the existence of this thing, you use it, we will not charge a lot of money, I just want these people to learn to use new tools to improve efficiency in the face of more new tools, and that's what the next generation of developers are capable of.
As far as enterprise customers need more services, because Chinese business customers typically are not necessarily aware of what they need, he may not know APM, do not know the infrastructure monitoring, but he knows what effect to achieve, so we give the enterprise is a solution, I said solution this does not necessarily say that there must be someone else's software, we are hoping that we give customers is the effect, the future of our enterprises to do similar like Palantir,palantir the company in the United States, its customers are like the United States FBI, NASA, oil companies like this, His business model is very simple, often their staff to talk to the CEO, you are now most concerned about which five issues listed, these people put their own software, people into your company for 35 months, put all your data, and now tell you the answer to these five questions, and finally give you a final conclusion, say how much money, Maybe a case of 10 million dollars, the company is such a model, now the revenue has been nearly 1 billion dollars.
China innovation "Pioneer" Enterprise Series report serial number company name set up time Ceo/cto official micro-blog Company products/directions 1.2012 Chen Ben
Web site fit 2. Friends 2010
@ friends microblogging C, C + +, Java Product Development &NBSP
3. Aggregation Data
@ Aggregation Data mobile data Service 4. Anchora 2009 Lu Weimin
Mopaas and inpaas
5. Fast enough 2012 years
Chiang Kai-miao @ fast enough technology
6. Evans Hai Hui
2012 Wu Kai
@ Evans OpenStack Public Cloud &NBSP;&NBSP
7. Sohu Cloud 2011 Chu Yingbo
8. Lenovo Cloud Storage 2009 Luo Jinjin
Cloud storage 9. Nanjing Intelligence 2012 She Janxia
Large Data real-time analysis 10. Shanghai San He race 2012
Cloud management, cloud storage &NBSP
11. National Cloud Technology
@ National Cloud Technology Cloud operating system &NBSP
12. SSO365 2012 Jian &NBSP;&NBSP
Cloud security, Cloud identity authentication
13. Cloudil Cloud Solution 2001
Yes @ Century Ding Li
14. Multiple backup
@ mu Lang cloud backup
15. Shanghai Wang Software 2011 &NBSP
Cloud based build station software supermarket
16. Cloud Wisdom 2009 Yinjin @ Monitoring treasure cloud monitoring, based on large data APM 17. Shenzhen Zeyun 2012 He Gianbin
High Performance Storage System 18. Shenzhen Wisdom Crown 2004 Lu Huili
Biological identification of hand veins, virtualization 19. Beijing Vauan Technology 2009 Cao XUEWU Tech Mobile video technology provider 20. Star Ring Information Technology 2013 Vauan @ Star Ring Tech data analysis Platform 21. Hangzhou Count Cloud 2011 Xuanxiaohua @ Hangzhou Cloud Data Mining
22. Red Elephant Cloud
2012 Long @RedHadoop
Large data platform based on Hadoop 23. Apicloud 2013 Shanda @APICloud Cloud API and end API
2012 Wang Tao @SequoiaDB
Large data, cloud computing, NOSQL&NBSP,
Cloud host, Cloud hosting, Virtual Datacenter 26 Isite 2008 Yang Bingfu @ Islab Virtualization and cloud computing
Data Center, virtualization 27. Pro-plus Communications cloud 2011 Ze @ Pro-plus communication cloud &NBSP
Communication Cloud 28. ONEAPM 2008 He Xiaoyang @ Blue Ocean ONEAPM APM based on SaaS platform
Note: September 28, 2014 update, continuous update ...
Note: The Cloud Pioneer Series is built by the CSDN Cloud computing channel, main reports (free) enterprises with unique competitive advantages in cloud computing and large data at home and abroad, with the purpose of communication technology, promote the development of cloud computing technology in China, only you have unique technology, products and services in cloud computing or large data, You can contribute, welcome to submit email@example.com.
Start building with 50+ products and up to 12 months usage for Elastic Compute Service