How to evaluate a NoSQL technician in the big Data age-products and technology

Today, the arrival of "big data" is no doubt, especially in telecommunications, finance and other industries, almost to the "data is the business itself". This trend has made a lot of changes to companies that believe in the power of data.

In the salon of large data technology conference, MongoDB Chinese community founder Cheng, Taobao core system storage System Research expert Yang Zhifeng, Sina Cloud Computing senior technology manager Conglei, 139 lobbyist architect Li Yu, ymall.com technical director Gang Jianhua on how to rationally consider NoSQL, NoSQL benchmark points, NoSQL and cloud computing integration trends, NOSQL and distributed computing, and other issues have been hotly discussed.

Host Cheng: In fact, we from two, we from two different angles to talk about this problem, ultimately you pay the ultimate cost, the top three most concerned about the maintenance of the database, he operation of some of the costs, may be more concerned about the development of the online part of the cost. Look at the two parts of the cost to find a very good balance to do this thing. In this company adopts NoSQL time, will certainly go to find some NoSQL technical personnel, you in the profession process how do you evaluate a nosql technical person, should have what quality? There will be countless people jumping into this circle when we're done. They jump into a brazier, or a fire pit, he needs to have what kind of quality, you think this thing, because the traditional DBA is a standard, basically what content, now nosql different, we all have the difference of view, Let's start with the building.

Gang Jianhua: I think a lot of nosql products for developers compared to two, I think it is relatively simple, as long as the developer can understand such as some of the data modeling model and the original different, the original habitual how to represent the data model. The second is about how to focus on performance optimization, how to find out the problems of their own code. Of course, in order to achieve the ability to optimize, you need to understand the database on the way to query and so on need to have in-depth understanding, and the normal database is no different, including IO, a complete set of architectures are very understanding. Of course, we also know that the relational database, many people even after the completion of the know that they do not know why can also let the program on-line, for MySQL This situation may exist, to ensure that these things will not be too big problem.

In general, as long as the use of, as familiar as SQL, know the rational use of index optimization database, how to know how to put the original SQL mode, forwarding similar to the document mode of development methods, can be based on NOSQL data development.

Moderator Cheng: Basic principles You know how the index, what kind of promotion, to know the SQL technology is good, learn NoSQL also fast?

Gang Jianhua: You can understand that.

Yang Zhifeng: My angle to do nosql people, there is a very famous last year when Google had a Daniel, he had a PPT in the last year to talk about his experience doing Google infrastructure. For example, one of them, you know how long it takes for a network to come back from here to England, a storage architecture, is to do the requirements of the system, you need to know how long a memory access, a disk access, a flash visit, I want to have a concept, in fact, to have a sense of the system, you see a system , you know where the bottleneck may be.

On the other hand NoSQL system is distributed system, from I want to hire, if he has distributed systems, such as graduate students are learning distributed systems, this is a big advantage. For example, you know what consistency is.

Li Yu: I don't feel so advanced, our project is an engineering nature, we recruit when the main use of Java, as long as you can Java, understanding is better. Just say we how to evaluate a person, will have a subsidiary characteristics, not recruit a full-time staff to carry out maintenance. I feel like now, if I had just given that angle, the requirements of our developers, but also to understand how some of the bottom of the index, because we have used a lot of index of the industry, but also on the Java Foundation is better, better understanding, we make use of some graphics, data theory more, For example, I find out what Obama's shortest data is, how to do it, and ask for more.

Conglei: It seems that the general it Internet company has not yet recruited NoSQL position, more projects are to use nosql knowledge. In fact, my point of view is similar to Zhi Feng, first of all, this person to see his needs, in fact, nosql into two layers, one is relatively high, involving a number of distributed theory, the 1th is not to say that this low, not low-level, more involved in local storage of things, very important hardware, I think that from the traditional hardware in the future this storage engine will change, including Google's latest one, there will be beyond the traditional search engine database appears to replace the previous method, from the top level is basically Amazon (micro Bo), Google a few bags, to learn theory, We generally mention that the theory is cap, can be divided into two reassuring, one is extensibility, there is a copy, from this aspect to better understand some.