The present situation of NoSQL

After at least 4 years of intense debate, it is time to make a phased conclusion about the status quo of NoSQL. So many things have happened around NoSQL that it is difficult to make a simple generalization, and it is difficult to judge what it achieves and what it does not expect.

In many areas, NoSQL has not only been successful in the academic field but also in the industry. The university began to realize that NoSQL had to be included in the curriculum. Just repeating the standard database is not enough. Of course, this does not mean that it is wrong to delve into a relational database. On the contrary, NoSQL is a very good and important supplement.

What happened?

The NoSQL field has exploded 50 to 150 new databases in a short span of 4-5 years. nosql-database.org lists 150 such databases, including some that are old but powerful, like object databases. Of course, some interesting mergers are taking place, such as the couchbase generated by Couchdb and membase transactions. But we'll discuss each of the major systems in this article later.

Many people have assumed that there will be a huge consolidation in the NoSQL field. But that did not happen. NoSQL used to be an explosive growth, and it is still the case. Like all areas of computer science--such as programming languages--there are now more and more blank areas that require a large number of databases. This is synchronized with the explosive growth of the Internet, large data, sensors and many future technologies, which leads to more data and different needs to deal with them. For the past four years, we've only seen an important system leave the stage: the German graph database sones. A large number of NoSQL still live happily, either in the open source community, regardless of any financial rewards, or in the business world.

Visible Sex and money?

Another important aspect is visibility and industry adoption. In this respect, we can see that there is a big difference between the traditional industries-to protect investment-and the emerging industries, mainly start-ups. Almost all popular web-based startups such as Pinterest and Instagram are using a hybrid (SQL + NoSQL) architecture, while traditional industries are still grappling with the use of NoSQL. But observations show that more and more such companies are trying to process some of their data streams using NoSQL solutions and analyze them later, including Hadoop, MongoDB, and Cassandra.

This at the same time leads to continued growth in the demand for architects and developers with nosql knowledge. Recent surveys show that the most needed developer skills in the industry are as follows:

HTML5

Mongodb

Ios

Android

Mobile Apps

Puppet

Hadoop

Jquery

Paas

Social Media

of the top 10 technical requirements, there are two NoSQL databases. One is even in front of iOS. If this is not a compliment to it, what is it?!

However, the adoption of NoSQL has become more and more rapid and deeper than initially predicted. In the summer of 2011, Oracle published a famous white paper that mentions that the NoSQL database feels like an ice cream flavor, but you shouldn't cling to it too much because it won't last long. But only a few months later, Oracle has demonstrated their plans to integrate Hadoop into large data devices. Even, they set up their own NoSQL database, which is a modification of the berkeleydb. Since then, all vendors have competed in integrating Hadoop. Microsoft, Sybase, IBM, Greenplum, pervasive, and many companies have already integrated it tightly. There is a pattern everywhere: can't beat it, just hug it.

But another important, but not nosql, sign of the widespread adoption of the NoSQL is that it has become a PAAs standard. With the ease of installation and management of numerous NoSQL databases, databases such as Redis and MongoDB can be seen in many PAAs services, such as Cloud Foundry, OpenShift, DotCloud, jelastic, and so on. As all things migrate to the cloud, NoSQL can put a lot of pressure on traditional relational databases. When faced with a choice of mysql/postgres or Mongodb/redis, for example, it will force people to think twice about their models, their needs, and other important issues that come with them.

Another interesting technical indicator is ThoughtWorks's technical radar, even though you may not fully agree with everything it contains, but it always contains some interesting things. Let's take a look at their October 2012 technical radar, as shown in Figure 1:

Figure 1:thoughtworks Technology radar, October 2012--platform