The main challenge for many data shops will not necessarily be tackling unstructured data, but rather machine-generated semi-structured data. “Unstructured data is relatively rare and is mostly just voice and video, which is usually processed to generate tags or full text transcripts,” said Zweben. “Most of the data explosion of big data comes from semi-structured data, such as web logs. This has meant that data management systems, such as Hadoop, need to be flexible enough to handle structured, semi-structured, and unstructured data, as well as cost-effectively ingest hundreds of thousands of records/sec and stores hundreds of terabytes, or even petabytes. Companies have created data lakes to capture this data and have just started to gain insights using analytics and data mining tools.”
Real time—an essential component of the big data explosion—also is becoming important in Oracle environments. Often, real time and big data go hand in hand. “Business leaders realize that recent technology advancements for big data and real-time analytics have disrupted, or provide the potential to disrupt, the way successful businesses operate,” said Gary Chan, strategic alliances director for Nimble Storage. “This is a good thing, as the potential of big data and real-time analytics merged with operational mission-critical transactions, through technologies such as in-memory databases like Oracle 12c, enables customers to respond rapidly to changing customer preferences and market changes.”
Changes are on the horizon for partners, vendors, and enterprises that comprise the Oracle ecosystem.
In-memory technologies—such as those now supported by Oracle—are significantly improving the speed of business decision making, especially against big datasets. “Oracle’s Database In-Memory option uses a hybrid approach to in-memory functionality,” said Caruso. “It combines a traditional row-based model stored in buffer cache with a new column-based format, and is now delivered as part of the 12c Database and licensed as an additional feature. The in-memory option can be used not only for analytics purposes but also for mixed use OLTP systems. Many of our customers use this technology in conjunction with a purpose-built hardware appliance such as Oracle’s Exalytics.”
Oracle’s in-memory positions the vendor and its products squarely in the middle of the big data and real-time analytics surge. OAUG’s English sees “the largest noticeable impact to the public through big data use in customer experience. With quick analytical capability, companies can easily target ads, product links, and other information based on historical data, resulting in a higher return of repetitive purchasing from the consumer. This is a positive experience for the consumer which means happier customers, repeat business, and with social media, extremely rapid expansion of business exposure. It’s a win-win-win.”
This is helping Oracle and its partners keep up with fast-changing demands. “Previously, customers would wait for relevant reports,” said Sudip Kar, vice president of delivery for Cambridge Technology Enterprises. With the advent of real-time data analysis, they now need analysis at a much faster rate. Recent investments in these technologies have resulted in huge dividends in terms of productivity, market responsiveness, development of better products, and insight into consumer sentiment and behavior, Kar noted.
In-memory may still have its limitations, some observers believe. “In-memory technologies have the potential to dramatically speed up the querying and processing of data,” said Zweben. “However, most in-memory technologies today require all of the data to be in-memory, which can be 10x to 20x more expensive than disk-based solutions. This limits them to niche applications that can justify the additional cost. But stay tuned—hybrid architectures are emerging that can intelligently merge in-memory—whether cache, main memory, PCI-based memory, or SSDs—with disk-based technologies.”