CIOs face the choice of large data architecture

For large data architectures, CIOs face a long-standing dilemma of choice: buy or build? New business problems, scarcity of vendor solutions and the emergence of a large number of new technologies have made it harder to make decisions. Moreover, the relevant nouns are so ambiguous, such as large data.

The feeling is stronger as Tesla announces the launch of a cheap electric car in 2017. But in the IT circle, CIOs are more concerned about how Elon Musk (Tesla's CEO and product architect) has the CIO to deal with this challenge: building its own enterprise resource planning system (ERP) rather than upgrading it based on SAP.

This is not just the work of the ERP, but the creation of a trend. At the MIT Sloan CIO Forum, participants gave Tesla more far-reaching implications, just as the Facobook plan builds its own CRM system. The implication behind such decisions is that the market has not been able to fully meet customer customization needs, and the scales have been skewed from traditional vendors to specific optimization schemes, especially in the context of the gradual infiltration of large data into business strategies. The most critical thing for CIOs should be as Babson Eton (Massachusetts Wellesley) information Technology and management professor Tom Davenport said.

"I don't know if Tesla's system can outperform SAP or Oracle products. What really struck me was the decision to develop everything on its own. However, it is always meaningful to build a strategic system for the differentiation of the business. "Davenport said when hosting large data-related research.

Identify your most critical technology requirements, those that are rarely available through existing vendors, and those that are most closely related to business issues.

When CIOs build large data systems, they may face a long-standing it decision dilemma: Build or buy? Now, the standardised schemes that dominate the market for years have not been able to deal with the bottlenecks in the big data infrastructure. More choices are starting to get into the spotlight, with memory calculations, NOSQL databases, cloud computing, Open-source software, and developers like Facebook and Tesla. However, in any case, the CIO first to eliminate the "big data" the word of the various chaos and confusion, under the tide to see the enterprises face the technical pain point. In the end, CIOs are likely to adopt a more targeted, point-to-point approach, rather than a unification solution covering all aspects.

First, what is the big data?

The word "Big data" has been in existence for more than a year, but its meaning is still confusing. "Big Data is an all-encompassing marketing term that you're interested in, and you can include it." "Curt Monash, founder of Monash, who is based in Massachusetts Acton, said.

Worse, Monash in a 2011 blog post, is that the definition of big data (including Gartner's 3V definition) is often misleading in itself, leading to market turmoil. Monash that the idea of having a technology stack that solves all the new IT problems is extremely naïve.

Darrell Fernandes, chief information officer of the Fidelity Investments Professional services group, said the abuse of the word "big data" might raise sales for vendors, but it also caused a lot of trouble for potential buyers. At the MIT Forum's Big Data workshop, Fernandes that this would hurt the industry, especially when the link between technology investment and business returns is ambiguous. In the 1990 's, you could clearly point to the business value of the CRM technology, and the large data coverage was too broad, the lack of a clear point, and the desire and anxiety of the business and it ends in vain.

It is not only Fernandes who hold such a tone. For example, in its latest report "Reset on" Big Data, Forrester (a consulting firm based in Massachusetts Cambridge) deliberately avoids defining large data as a technology or a technical problem that could lead to a narrow view of the technician and miss the trend. Instead, according to Forrester analyst Brian Hopkins, one of the authors of the report, the big data is "bridging the gap between your ability and existing data, so that information is truly translated into business insights, which will be an ongoing process." ”

In the beginning, you need to know where to start. Forrester's recommendation is that business and IT leaders work together on a practical scenario before building or buying. Sometimes it takes an investment to implement a scenario. However, there are times when a change in corporate culture is required, that is, based on the data to make decisions. Large data means that the business model is data driven. This transformation requires the ability to have a clear understanding of the company's current situation before making any investments.

Single vendors may not be able to handle

When large data technology investments are unavoidable, Hopkins reminds CIOs not to be limited by certain nouns (such as Hadoop, which has recently been used as a synonym for large data), but rather to revert to nature and focus on actual functionality. He recommends focusing on the "high elasticity, cost-acceptable data storage and analysis methods" that Hadoop offers.

The basic certainty is that the technology stack that solves the data problems facing the current enterprise will not come from a single vendor. "You can't get a one-stop solution." "Hopkins said. Buying from a single vendor would be very expensive, Monash says: "The high investment comes not only from money but also from management costs." ”

Hopkins refers to companies such as IBM, Microsoft and Oracle, which have a complete solution stack: "But I do not think that their products, such as the manufacturer's propaganda, the application, potential customers should also understand this." "If you choose Open Source software, Hopkins thinks there will be trouble too." Vendors such as Cloudera, MAPR and Hortonworks have opted for different open-source versions to build their products. Sometimes the new features developed by these vendors coincide with the open source community. So customers may have to make trade-offs at the level of functional modules.

In this respect, Monash has its own principles: "To find their most urgent technical needs, that is, few vendors can meet the functions and those with the current business issues the most closely related to the technology." ”

Of course, consideration of business problems or outputs cannot be overstated. As Fernandes explains, the company is gradually making the concept of big data clear and concrete: "We have been looking at specific assumptions, specific scenarios and specific outputs over the years." Because the focus is very focused, we can continue to achieve progress. ”