The present situation and trend of dialysis large data load platform and technology

In the previous article, we have scanned the recognition and attention of large data based on China's different regions and industries, and seen the acceptance and development potential of the x86 architecture as a large data base platform, in combination with the ZDNet launch of China's large data recognition and application market research. This article will be based on the previous article, combined with database application and related technologies to interpret the research report, and summed up several major findings and conclusions.

I. Strong demand for large data analysis and storage

Research data show that, in addition to having recognized that large data has large capacity, with the characteristics of high speed and many data types, including unstructured and semi-structured data, more enterprises will associate with data storage and data analysis and other large data processing technology, it is not difficult to see on the basis of the recognition The enterprise has shown strong desire and interest in the processing of large data, and it is very likely to be an important it planning of the enterprise.

Two. x86 database application Mainstream platform

Although data has multiple sources, there is no doubt that databases are the primary source of enterprise data, and building databases, data warehouses, and business intelligence analysis platforms are also critical to the larger data process. In the data age, database application is undoubtedly one of the core applications of enterprises, from the enterprise's existing database platform, 72.93% of the enterprise current database applications based on x86 platform, only 20.79% of the existing database applications based on Itanium, SPARC, Power and other Unix systems.

From the existing database platform, x86 architecture has been widely recognized and trusted by enterprises, good "congenital advantage" for the x86 to deal with the challenges of large data to play an important role in the foreshadowing, with the database and other key applications in the field of performance, the x86 platform in the field of database applications will be more extensive and in-depth.

At the same time, in the construction of data Warehouse, business intelligence, the survey data and the enterprise existing database application load main platform consistent, 82.99% of the enterprises tend to choose x86 platform to build such data analysis and business intelligence applications, in particular, there are 48.99% Enterprises tend to adopt four-way and above x86 servers, 34% of enterprises tend to adopt dual x86 server, which is also with four or more high-end x86 server positioning match, in recent years more and more four-way and above high-end servers began to enter the database, virtualization and other enterprise key applications, and has been widely recognized by the enterprise.

Three. Hadoop is widely accepted

Unlike existing traditional structured data processing, large data processing will involve massive unstructured and semi-structured data processing, in which the techniques and methods for traditional structured data processing are obviously difficult to cope with, and because of this, more and more large data processing technology is emerging, and the enterprise has been favored, such as the hot distributed computing system Hadoop. It is noteworthy that the birth of Hadoop itself is closely related to the x86 platform, and it is the birth of Hadoop that makes it possible to run a distributed computing framework on a low-cost general-purpose hardware cluster, in which Hadoop itself is designed and optimized for the x86 platform.

Hadoop was originally an open source framework, developed by the Apache Foundation, and its important innovation is to break down tasks and run them in multiple nodes, and when the data is too large to be processed on a single server, the advantages of distributed computing are reflected, in addition, By combining this technology with Linux servers, you can replace large scale arrays with a very high price/performance ratio. It is also seeing the potential of Hadoop in the future to deal with large data, and many organizations and suppliers have started to develop Hadoop technology and have formed a certain size, Intel is the representative.

As a catalyst for the x86 platform, Intel has clearly foreseen the enormous power that the x86 platform could unleash with Hadoop technology in response to large data, and in July 2012 the Hadoop release was officially released, with a combination of soft and hard work to fight large data, The Hadoop release has gained a high level of market recognition, and the Intel Hadoop release has become the most popular Hadoop version of the company in a 44.02%-percent share of the knowledge research for the Hadoop version, which can also be seen in addition to the hardware platform, Intel is also admired for its large data software platforms.

Four. Challenges facing large data platform scenarios

As more and more companies focus on big data, and finally will build large data platform as the enterprise's it planning, a large data platform and solution selection has been inevitable, through research, in the platform and project selection, enterprises most concerned about the first three factors in order: price, security, stability, it is obvious, More companies want to deploy large data platforms to cope with the pressure of data growth, while worrying about the cost of investing in large data platforms; In addition, data security and privacy have always been a huge potential problem in data processing, Concerns about the security of large data processing are also due to the lack of data security standards and specific technologies specifically designed to protect enterprise data privacy, and the fact that stability concerns have been reduced by concerns about the stability of general hardware platforms This also has to do with the performance improvement of the x86 platform in recent years, and the combination of x86 platform and Hadoop also provides a lower cost option for processing large data.

Five. Knowledge of large data technology providers

If the approval degree from the database load platform and related technology can reflect the x86 architecture's technical strength in dealing with large data, the perception of large data-technology providers is more intuitive, with 19.42% of users thinking that Intel is closely related to large data and is more receptive than IBM, Oracle, It vendors such as Microsoft, see Intel as a large data technology provider has been widely recognized.

With the advent of the large data age, data-centric application is undoubtedly the core, the enterprise will pay more attention to data storage, analysis and management, the infrastructure can meet the large Data environment scalability and provide excellent performance is undoubtedly a big challenge. Intel is widely accepted as a large data technology provider, with Intel in recent years in the field of large data, but more importantly as a provider of data center infrastructure, Intel's ability to analyze and store data is beyond doubt, in the new wave of large data, Intel is also working to drive the data center infrastructure toward a large data platform, and, with the optimization and expansion of the Hadoop feature, Intel has built up the best practice experience in Hadoop applications, and is certainly more comfortable with massive data and unstructured processing.

(Responsible editor: The good of the Legacy)