By Mike Faden
Data management is at the heart of fundamental trends now transforming how businesses use technology. Companies are exploiting vastly expanding volumes of data from a growing number of sources. Rapid analysis of that data has become key to maintaining a competitive edge. And cloud technologies for gathering, storing, analyzing, and using that data are enabling dramatically increased agility as well as entirely new business models.
With Oracle Database 12c, Oracle has engineered its industry-leading enterprise database to support these transformational trends while protecting businesses’ investments. Multitenancy, a foundation of Oracle Cloud, is a powerful feature of Oracle Database 12c. The Oracle Database 12c multitenant option also facilitates greater agility on premises and zero downtime migration to Oracle Cloud, supporting instant provisioning while slashing administrative effort.
“Every ten years, the computing world changes completely,” says Andy Mendelsohn, executive vice president for database server technologies at Oracle. “The huge generational change going on right now is that we’re transitioning from internet computing in customer data centers to the cloud.”
And with Oracle Database 12c, he says, Oracle continues with innovations that give customers the flexibility to choose their own path in that direction and pursue it at their own pace.
Colliding Particles, Exploding Data at CERN
Among the organizations taking advantage of Oracle Database 12c capabilities is CERN, the European Organization for Nuclear Research. Located near Geneva, Switzerland, the particle physics lab seeks to explore fundamental questions about the structure of the universe (becoming famous along the way for inventing the World Wide Web, in 1989). To study particles and how they interact, CERN uses the world’s largest scientific instrument, the Large Hadron Collider (LHC). The LHC accelerates particles to near the speed of light, creating collisions that help scientists gain a better understanding of how our universe works. Among the LHC’s key achievements was providing, in 2012, experimental evidence of the Higgs boson—the previously unverified final piece of what is known as the Standard Model of particle physics.
Headquarters: Geneva, Switzerland
Employees: More than 2,250
While creating collisions between particles, the LHC also creates extraordinary amounts of data. Even after filtering out more than 99 percent of the raw data produced by experiments, CERN still expects to generate about 50 PB in 2017, says Eric Grancher, group leader of CERN’s database services group. The data is stored and analyzed on the Worldwide LHC Computing Grid (WLCG), a cloud-based global distributed computing infrastructure, coordinated by CERN, involving more than 170 computing centers worldwide. The WLCG runs more than 2 million jobs daily.
More than 100 Oracle production databases are used to manage some of the vast amounts of control data generated by the LHC’s systems. “We are talking about a sustained level of 150,000 changes per second,” says Eva Dafonte Perez, deputy head of CERN’s database services group. CERN’s “quench protection” database alone generates 150 TB of data per month, she says, and the LHC logging database captures 1.5 million signals from a variety of sources.
In addition to providing hardware and software to CERN for 35 years, Oracle has now been a partner in CERN openlab for well over a decade. CERN openlab is a collaboration between the laboratory and leading information and communications technology (ICT) companies that works to accelerate the development of the cutting-edge ICT solutions needed to make groundbreaking research possible. A number of Oracle solutions have been tested through CERN openlab, with CERN’s input helping to guide the development of new database features.
The organization runs a combination of Oracle Database 12c Release 1 (12.1) and Oracle Database 11g and is planning to take advantage of Oracle Database 12c Release 2 (12.2). One feature that CERN uses is Oracle Database In-Memory, which accelerates key database operations and uses a unique dual column and row datastore to support analytics and production workloads on a single database.
“The big breakthrough with the in-memory technology in Oracle Database 12c is that you can do high-performance analytics against your live transactional data,” Mendelsohn says.
This can really change the way people think about computing architectures. The in-memory enhancements are one of the features we’re going to be deploying in Release 12.2 as soon as possible.”–Eric Grancher, Group Leader, Database Services, CERN
For CERN, one key advantage of the in-memory capability is transactional integrity for specific scale-out applications, Grancher says. Rather than maintain data caches at the application level, the entire database can be stored in memory, ensuring a single source of consistent information and guaranteeing transactional integrity. Grancher says one project has already changed its plans to take advantage of the new capability. “This can really change the way people think about computing architectures,” he says. “The in-memory enhancements are one of the features we’re going to be deploying in Release 12.2 as soon as possible.”
CERN is also looking at the possibility of using in-memory capabilities together with Oracle Active Data Guard to flow data workloads from primary LHC databases to secondary stores for analysis. With Oracle Database 12c, says Mendelsohn, you can now create an Oracle Active Data Guard standby for your transaction system and create only the in-memory column store technology that supports those high-performance analytics on the standby database. “You can send the transactional users to the primary database,” he says, “and you can send your business analytic users to the Oracle Active Data Guard standby, where the data might only be a few seconds behind the data in the production system. Now, you have the best of both worlds, where you can completely isolate the transactional users from the analytic users, but still deliver incredible high-performance analytics against data that’s almost real-time.”
Big data analytics will support an expanded range of use cases across the laboratory, says Manuel Martin Marquez, a big data specialist in CERN’s database services group. “With big-data technologies, we’re able to perform real-time analytics; we are able to perform analytics on bigger datasets with better performance,” he says.
CERN expects to start deploying Oracle Database 12c Release 2 when the LHC enters a scheduled shutdown phase next year. The LHC operates in carefully planned cycles, typically consisting of a three-year period producing collisions followed by a roughly two-year shutdown for upgrades and maintenance. The next upgrade will allow the LHC to run with much higher performance, resulting in more particle collisions and further opportunities for scientific analysis and discovery. The higher performance will generate even more data at faster rates, especially because CERN expects to retain a greater proportion of the raw experimental data for analysis. By 2027, the accelerator could be generating around 600 PB per year, says Grancher—requiring new, more-efficient ways to analyze data.
As the data management requirements expand, CERN, in collaboration with other research institutions, is looking to the cloud to help handle the load. CERN already manages its main data center (and an offsite extension in Hungary) through a private cloud environment based on OpenStack, with approximately 300,000 processor cores.
As part of this, CERN is exploring the integration of Oracle Cloud Infrastructure (formerly Oracle Bare Metal Cloud) into its current environment, says Katarzyna Dziedziniewicz-Wojcik, a senior database administrator at CERN, and is in the process of testing a variety of Oracle Cloud capabilities. In one very large test, CERN ran a scientific workload on 10,000 cores in Oracle Cloud Infrastructure. CERN also plans to run applications on Oracle Cloud Infrastructure Compute, to assess performance, the cloud resources required, and the effort involved.
Additional tests for which CERN is preparing include cloud-based disaster recovery using Oracle Active Data Guard. “We have quite a complex environment, with multiple networks and high-performance network links. We want to see how complex it would be to set up such a solution in the cloud and also whether a cloud solution would fulfill our performance needs,” Dziedziniewicz-Wojcik says. CERN is also looking at using containers to switch workloads to the cloud for disaster recovery. The organization has already shown that it’s possible to move applications from an internal Kubernetes cluster to a cluster in Oracle Cloud Infrastructure, for example.
Viscosity: Consolidate and Simplify for Customers
Like CERN, many other organizations are finding that key features introduced in Oracle Database 12c are valuable both in the cloud and on premises, says Charles Kim, who has extensive experience with implementations of the database both as president of the niche database and cloud consulting firm Viscosity North America in Dallas, Texas, and through his role as president of the Independent Oracle Users Group (IOUG) Cloud Computing Special Interest Group.
Headquarters: Dallas, Texas
Revenues: US$7 million
For example, one Viscosity healthcare client, a software-as-a-service (SaaS) company, is using multitenancy to drastically simplify its environment with server consolidation, he says. Previously, the company was supporting its users on many different servers, based largely on older hardware. “Managing all of that became a nightmare,” Kim says.
With Oracle Database 12c, Viscosity has been able to help consolidate those databases using multitenancy by implementing them as multiple pluggable databases within database containers. Because Release 12.2 supports consolidation of pluggable databases with different character sets, the company has been able to achieve greater efficiency by consolidating databases supporting customers in different regions.
“They were able to perform extreme consolidation, which addressed scalability and stability issues—and alleviated their headaches,” says Kim. Release 12.2 has also facilitated testing by enhancing support for flashback in pluggable databases.
Businesses are also taking advantage of some of the less high-profile—but extremely practical—new features in Oracle Database 12c, Kim says. For example, Release 12.2 introduced the ability to move entire tables online, while automatically updating the indexes. That means easier online maintenance with less or zero downtime. Another key example is the index usage tracking capability in Release 12.2, which provides additional information about how individual database indexes are used over time “As an administrator, it’s very important to know, because if a specific index is used only once as part of a batch job that runs every three months, why even keep it there? Eliminate the overhead,” says Kim.
The longer object names allowed in Release 12.2 are proving valuable to customers as well. Before the new release, all object names were limited to 30 bytes; now, the limit has been increased to 128 bytes, making it easier to provide descriptive names. Kim says that’s valuable in a variety of situations, such as when moving data into Oracle Database from other datasources that support long object names.
“We want to be at the frontier of Release 12.2 adoption to help customers migrate to Oracle Cloud with zero downtime,” says Kim, whose company has authored more than 18 books focused on Oracle technology.
AS ONE Uses JSON for Real-Time Collaboration
Oracle Database 12c includes important new features for developers, including the ability to store JSON as text directly within the database. Japanese company AS ONE is taking advantage of that JSON support as it uses Release 12.2–based Oracle Database Exadata Express Cloud Service to support a key requirement: providing its 4,300 dealerships with real-time information for more than 300,000 inventory datasets.
Headquarters: Osaka, Japan
Revenues: US$559 million
AS ONE is a general trading company offering a broad range of equipment and scientific instruments for users in the research, industrial, and medical fields. It offers products through catalogs and the company’s ecommerce website, and it provides them through a delivery system that connects users with thousands of dealerships and manufacturers. With a product range that has rapidly swelled from around 100,000 items to 1.7 million, the company’s systems need to provide both high performance and high availability. To support that expansion, AS ONE created its core business systems using a combination of Oracle Database, Oracle WebLogic Suite, and Oracle Enterprise Manager Cloud Control 12c.
AS ONE started receiving requests from dealerships for inventory information to be linked to the company’s systems in real time, instead of updated once a day. To support that need, the company began looking for a solution that would provide real-time links while minimizing the burden on the company’s core infrastructure.
That’s when Shinichi Hakoda, manager of the IT Promotion department at AS ONE, learned about Oracle Database Exadata Express Cloud Service and the fact that Oracle Database 12c includes RESTful interfaces that enable access to real-time information via APIs. This allows data access via JSON and other means without the need for an API server, he says. “From the development perspective, it’s possible to develop REST services with only SQL knowledge and screen configurations,” which allows a high level of developer productivity, he says.
“Especially the new-generation web developers are making heavy use of REST to call into the database,” says Mendelsohn. “Oracle REST Data Services makes it easy to query and operate the database via RESTful interfaces.”
Hakoda believes that becoming an early adopter of Oracle Exadata Express Cloud Service is giving the company an edge in being able to deliver advanced API-based services to its customers. “We can be a step ahead of our competitors,” he says. The company’s internal users are already using the information provided via the APIs, and external customers are starting to test it.
Hakoda says that choosing a cloud-based service for the new system offers several benefits. “From the standpoint of operational costs,” he says, “the latest environment is always offered, and security is supported by Oracle’s high level of technical strength, which we feel is a major advantage.”
Additionally, JSON support in Oracle Database 12c will enable AS ONE to rapidly add further real-time information services. “These services make it easy to expand from our current provision of inventory information to things like price information APIs, product information APIs, and delivery period response APIs, which we also feel are extremely strong advantages,” says Hakoda.
AS ONE is also moving toward the use of chatbots as it continues to seek out new technologies to increase efficiency and help customers. “We feel that chatbots have extremely high potential,” Hakoda says, partly because they provide a simple interface that can make systems easier to use and reduce training costs. The company has already linked the inventory information API on Oracle Exadata Express Cloud Service with chatbots and is making bot responses to inventory inquiries from within the company. Next step? “We would also like to deploy it externally,” says Hakoda, “and boost customer satisfaction levels.”
“Oracle has led the industry by providing the fastest, most reliable, and highly secure database for organizations of all sizes,” says Mendelsohn. “With the current transition to in-memory computing and cloud, we will continue to innovate to ensure we have the best database technology for the future.”
LEARN more about Oracle Database cloud services.
LEARN more about Oracle Database 12c.
Illustration by iStock.com/kentoh; Photography by Darrin Vanselow/The Verbatim Agency, Bob Adler/The Verbatim Agency, Christopher Jue/The Verbatim Agency