Direct exposure to large data management risk: risk before opportunity

Big data for many companies does not mean opportunities or business potential, before they can manage data well, large data only means risk and endless annoyance.

Three important topics in large data: unstructured data beyond structured data, increased volume of structured data, and commercial analysis of structured and unstructured data, the first two issues are the basis and prerequisite for the third issue to be implemented without good data storage, protection, migration and grooming, Trying to analyze the data is a myth-where does the data come from? How do you analyze unstructured data without structure or clue? The first thing the enterprise does is to reduce the "data storm and risk" in the face of large data.

But the problem is not as simple as it appears to be--it's not just another game of storage capacity--it's a full range of risk-surrounding and alarm bells for businesses, in addition to the capacity crisis, enterprises in the large data age is faced with great management risk, which includes the accumulation of heterogeneous storage architecture and complex environment of "rigid architecture", Limited storage managers are unable to cope with the increasingly complex storage environment, the growing data tension budget, of course, we also want to see as the volume of data explosion and the increase of data type complexity, the enterprise compliance also has a growing impact.

It can be said that in the large data on capacity and data analysis system caused by the challenges and pressures behind, is the great risk that large data poses to enterprise storage management-even simple data and storage work can be risky, and if we take a look at what we're going to do with the data, we'll find the problem even more frightening: backup, recovery, Snapshots, migrations, replication, remote replication, duplicate data deletion, capacity recovery, archiving, encryption and decryption-only roughly, no more than 10 operations on the data.

Whether it is from the Enterprise storage strategy and environment, or from the point of view of data and storage operation, the "management risk" brought by large data is not only becoming more and more prominent, but if it cannot be solved properly, it will surely cause the terrible consequence of "big data is the big risk".

Gartner: What are the challenges and risks of big data?

Global technology research and consultancy Gartner points out that CIOs focus on managing "big data" challenges with large volumes of information by excluding other aspects of information management, but this leaves a lot of challenges to be addressed later. When information managers lose control of data access and eligibility, they may focus on quantity. Gartner analysts cautioned that too narrow a focus would force years to undertake large-scale reinvestment to address other aspects of big data.

"Today's information management guidelines and technologies are simply unable to complete the task of handling all dynamic information," said Mark Beyer, Gartner Research vice president. Information managers must fundamentally reconsider their approach by planning all aspects of information management. The business requirements for accessing vast resources of large data provide information managers with an opportunity for the enterprise to change the way they use information. It leaders must educate their business peers to confront challenges while ensuring a degree of control and coordination so that large data machines will become cluttered with large data, which may increase compliance risks, increase costs and create many islands of information. ”

As noted above, analysts believe that volume is just the most direct and common problem of large data challenges and risks, and CIOs have been plagued by the shift in the number of transactional information into decisions-there are now more types of information needs to be analyzed-mainly from social media and mobile (situational awareness). Categories include tabular data (database), layered data, files, e-mail, metering data, video, static images, audio, stock quotes, financial transactions, and more. Speed involves the flow of data, the creation of structured records, and the availability of access and delivery. Speed means how fast data is being generated and how quickly data must be processed to meet demand.

In addition, better data protection and data containment are issues that enterprise CIOs need to focus on today, as data leaks, losses, and hacking activities are growing. Like some car industry CIO told Doit, previous full backup one night dry 4 hours, now full backup 4 night finish 1 times is good, in a large number of automotive advanced design means of stimulation, the automotive industry, the pressure of unstructured data "surge", far more than the original information system load capacity.

But all of these things need to be backed up or left out, because these are very important car design materials. The great risk lies before the CIO: The longer the Backup window is, the more data de-duplication technology is not the only one that is effective, unstructured, and structured data in the storage system, supported by a fragmented, insular approach to management and storage. So the CIO is very interested in tiered technology, data deduplication, solid-state drives, unified storage, and even cloud storage, and "This is also forced."

Solution: How to avoid the risk of large data management

The first priority in avoiding the risk of managing large data, is not the implementation and deployment of technology or products, most importantly, the enterprise strategy and the CIO's conceptual transformation: Big Data is not the first opportunity but the challenge, the first thing to do is not data analysis, use, but to better storage and management of data, This is the first thing the CIO needs to do in the big data age.

Of course, correcting the concept does not mean that we do not have more effective means and programmes to address the risk of managing large data. In terms of quantity, the "terrible" of large data is the first of its "big", that is, the large-scale effect of data, the existing manual, artificial way naturally is not able to deal with, therefore, it is important to have a highly automated solution to deal with.

Rather than automating and digitizing the enterprise's data storage policies, highly automated solutions require the ability to unify, automate management of a variety of storage devices, including stored data types, and automate, reliable policy execution to reduce the workload of people, while automating monitoring and reporting, Alerts can alert you to any irregularities or backup recovery policies that are not properly implemented. Obviously, automated management can reduce procurement costs and TCO to some extent.

We note that many of the products on the market are starting to work on simplifying the management interface, strengthening automation and intelligent policy Management, whether it is IBM's legitimate mainstream storwize V7000 or EMC launched last year's Vnx, management interface and automation is very high, for example, V7000, Its management interface, which was born out of IBM's high-end storage XIV, is simple and effective, even if there are multiple heterogeneous storage in the San, and when V7000 virtualization of the disk array in the entire SAN environment, the vast majority of management work is focused on V7000 unified execution. This simplifies management and lowers the management risk associated with limited professional management.

In addition to automation, in today's midrange storage market, the midrange storage of the top vendors is almost unified storage. IBM's Storwize V7000 naturally not to repeat, is the industry's mid-tier storage to file Data + Block data Unified storage transformation is very fast and very successful example, EMC's Vnx and entry-level Vnxe also provides a unified storage platform, Dell has FS7500 and entry-level Ns3500,netapp based on the EqualLogic platform. It's a NAS platform based midrange storage. Naturally, there is no need to repeat.

We see a growing trend in the unstructured data of large data, midrange storage is easier to consolidate block data and file data into a single unified storage platform than high-end storage, and these systems consolidate the original block data and file data and different storage systems for the enterprise, and the unified storage is also "virtual storage" : V7000 can virtualize all heterogeneous disk arrays in a SAN environment to form a converged, unified storage pool. Data can be stored and transferred transparently across different arrays, so resources under the entire SAN can be fully mobilized to quickly resolve any space or performance problem, greatly reducing the risk of management as a result of the complexity of the storage environment.

Automation, unified storage of block data and file data, and storage-system consolidation brought about by virtualization can effectively reduce the risk of data storage, especially large data storage.

(Responsible editor: The good of the Legacy)