Large data needs to be combined with three it trends

If you are an IT company in the Asia-Pacific region, you may consider deploying large data solutions in your business. Over the past 12 months, big data has been the most discussed topic for IT companies in Asia. Forecasts from IDC, an analyst firm, suggest that the big data market in the Asia-Pacific region is expected to grow from $258.5 million trillion in 2011 to $1.76 billion trillion in 2016. However, analysts predict that while the region's interest in big data is booming, the excessive hype and lack of sufficient knowledge of big data could lead to "irrational exuberance and unreasonable expectations".

Large data is not only the research object of the enterprise. It is certain that most businesses can benefit from it. For example, in the UK, Tesco, a supermarket chain, has developed promotional plans and adjusted commodity prices by analyzing the vast collection of customer behavior information data collected.

IDC observed that markets in the Asia-Pacific region have unique characteristics, such as distributed manufacturing centers centered on a large population of mega-cities, and that data-sharing and fluid regulations create significant new opportunities. In Asia, however, many large business centers have only built large data repositories under the data initiative. While this is indeed a very important first step, it does not mean that the enterprise really realizes the understanding of the concept of large data benefits. By purchasing all the required servers to process and analyze data, even if the process may take only a few hours, one weeks, or one months, such expensive costs can lead to resources being underutilized at other times.

To be more economical and strategic, CIOs in Asia must combine the power of the three major trends in it: Big Data, virtualization and cloud services. Virtualization and cloud computing are driving forces for large data use, making it possible to create large pools of highly automated computing to handle large volumes of data. The combination of the three trends will create the foundation for a flexible, scalable, intelligent, large data application.

Thus, economically, an enterprise can "lease" almost unlimited storage capacity from the cloud provider in a form of infrastructure, a service, and only pay for the capacity they use. Since then, the problem of storage and processing power has been solved.

However, migrating large datasets from the Cloud data center will show a weak link in the "Big Data System", a network that is not used to process large amounts of data. With the dizzying rise of new services and data, the problem of network capacity has been highlighted, which cannot be solved by simply installing or leasing large data storage space.

Instead, using a virtual data Center architecture is a better service method for handling large data efficiently, and the physical wall in a single datacenter effectively breaks the connection to multiple data centers as a logical entity. In other words, this creates a "no walled data center" that uses a high-performance "cloud backbone network to seamlessly connect to a pool of shared resources between a larger enterprise and carrier data centers." ”

By virtualizing and centralizing all of the data center and network assets, organizations can allow flexible configuration and migration workloads based on changing needs. This is the only setting that can be very economical to meet the infrastructure challenges of large data.

Forrester said that the vast majority of Asia-Pacific data centers are mainly large data repositories, followed by providing users with more advanced data mining and visualization tools. But they are not "data centers without Walls" and cannot allow large data processing in the most efficient way to meet business goals.

In order for large data to be appropriate for your business so that its computing power can easily participate in the processing of your enterprise's dataset, you need a high-performance connection: from your data center to the data center of your cloud service provider. However, static connection services that are always connected are expensive to maintain. A "performance-required" connection service allows you to turn on bandwidth when you need it and turn it off when you need it.

(Responsible editor: Fumingli)