Large Data Virtualization 0 Beginnings (i) opening

The virtualization of large data: the trend of enterprise IT development

The virtualization of large data is a trend in the current large data and Hadoop community. At the Hadoop summit conference in June 2013, Gartner said that in order for large data technologies such as Hadoop to truly fall into business, we need to start with specific business requirements, drive the development of large data-related technologies and products, and gradually bid farewell to technology-driven innovation. Gartner also points out that virtualization is an important trend in this new phase. As the proportion of global enterprise IT virtualization has exceeded two-thirds, the data centers defined by virtualized software have become increasingly popular and important to the enterprise, and it has become a real challenge for large data to affect and integrate into the infrastructure of existing enterprise data centers under such a wave. This blog will focus on this topic, from the operational dimension, technology, and solutions from the perspective of the specific content.

What exactly is the virtualization of large data?

To answer this question, first of all, why does enterprise it need virtualization? I think there are several reasons for this:

1. Virtualization can significantly increase server utilization and achieve better utilization by consolidating server resources.

2. x86 server as the representative of the virtualization itself cost relative to minicomputer and hardware and software integrated equipment, more economical, and performance is not inferior, horizontal expansion is a huge advantage.

3. Virtualization takes on important groundwork in cloud computing, whether public or private cloud. Without virtualization, cloud computing flexibility and multiple tenants are often hard to implement.

4. Virtualization has been able to support enterprise-critical applications (such as ERP, mail servers, business production databases, etc.), proving that there is no longer a two choice between virtualization and performance stability. In addition, numerous success stories and technical white papers can help more customers strengthen their confidence. The sign of virtualization towards full-blown maturity has been established.

Obviously the process of enterprise virtualization is not going to stop, and leading vendors, including VMware, are expanding Virtualization 2.0. Not only is the server (computing resources) virtualized, including storage and networks such as the past relatively difficult to be directly used by the virtualization of the island has been the forefront of innovation, such as "Software definition data Center", "Storage Virtualization", "Network virtualization" and other hotspots, have been the emergence of specific products and solutions.

The virtualization of large data is the task of running or migrating large data workloads into virtualized environments. In addition to naturally inheriting the general benefits of virtualization mentioned above, there are several special benefits worth mentioning:

1. Since large data infrastructures are initially difficult to determine how much computing and data nodes are needed, these nodes require one by one to heap on physical servers. Without expert team support, it will be time-consuming and laborious, and the future expansion is very inconvenient, the utilization is extremely low, the management efficiency problem is quite outstanding. Virtualization not only enables rapid deployment of clusters, but also the flexibility to manage them, while significantly increasing utilization.

2. Large data mixes use shared storage and local storage to improve performance. Virtualization can fully meet these requirements and give us the flexibility to extend and design policies.

3. Virtualization can be large data from the bottom up to form a multi-tenant and data Analysis Services, a good isolation of the computing environment, to promote large data that is the basis for service.

4. Virtualization also facilitates the integration and integration of other data applications on a unified virtualization platform, significantly reducing the complexity and operational costs of IT infrastructure.

I think the above not only explains what is the virtualization of large data, but also explains the existence value of this market. So what else do we need? Knowledge and skills. The biggest problem facing the enterprise is not the real demand, but the lack of professional talent to find and deal with it; 57% of companies think they need people with specific skills and knowledge, while management and security are also a big challenge, accounting for 37%. These figures attest to the necessity and value of virtualization. (Data from Microsoft Inc. 's report on Hadoop Summit 2013)

As for the new topic of virtualization of large data, I think there may not be many people in the market who can understand how to implement it and what technologies and products to use. Because of not understanding, it is easy to cause arbitrary judgment, think large data and virtualization is contradictory, even think that the combination of the two is "not reliable". I will elaborate on how to implement the big data on virtualization through a series of upcoming posts, and let readers know the relationship and help them solve their concerns. Therefore, the next series of blog articles are technical "dry goods", will guide interested in further understanding and try this field of readers or enterprises to further explore and hands-on try to use.

Author Introduction:

Zhang June late

VMware Large Data Solution project Manager

Currently responsible for the management and marketing of VMware large data solutions. Vfabricdata Director product Manager for VMware database management products, with in-depth understanding and hands-on experience in enterprise products, technology solutions and Markets for virtualization, cloud computing, relational databases and large data. Prior to this, Microsoft was working on product management and research and development in distributed systems.

See more highlights of this column: http://www.bianceng.cnhttp://www.bianceng.cn/Servers/virtualization/