Currently, the IT industry is the midst of a major transition as it moves from the last generation - the internet generation - to the new generation of cloud and big data, Andy Mendelsohn, Oracle’s EVP of Database Server Technologies.
With Oracle now offering options at all levels of the cloud stack, Mendelsohn recently talked with DBTA about database products that Oracle is bringing to market to support customers’ cloud initiatives.
Over the past 6 months, he said, Oracle has continued to introduce new and updated products and services to help customers take advantage of the next wave of technology innovation fueled by big data and the cloud. In particular, Mendelsohn said, five key trends stand out.
More Pluggable Databases in 12c R2
The first 12c release with multitenant architecture supported up to 252 “pluggable databases” that can be moved between containers, and in a big advantage, customers can manage large numbers of pluggable databases at the container level - including backup, recovery, and upgrades, Mendelsohn noted. In 12c R2 which was announced at Oracle OpenWorld in October 2015, said Mendelsohn, Oracle continued to work on multitenancy with new support for up 4,000 pluggable databases per container. Four thousand database instances is a high number for an on-premise environment, even for a very large Oracle customer, but in the cloud, “4,000 is just getting started,” he said. “In a cloud, you want to have hundreds of thousands, millions of databases, so there you want larger numbers of these pluggable databases per container.”
The multitenancy database structure is also “a great fit” for Oracle’s SaaS applications, said Mendelsohn. Oracle currently has the Taleo applications already up on 12c using the multitenant architecture and the company is in the process of moving all of its Fusion SaaS applications to the 12c multitenant platform as well, said Mendelsohn. “None of the other database vendors have anything like our multitenant architecture. We think this is going to give us a big cost advantage in the public cloud where we can run and manage large numbers of databases at very low cost.”
Engineered systems in the cloud
According to Mendelsohn, unlike other public cloud providers that offer infrastructure as a service that is not optimized for particular workloads, with the introduction of Oracle Exadata and Big Data Appliance in the cloud, Oracle can support customers’ mission-critical workloads. The result is that Oracle now has infrastructure as a service for “commodity vanilla software workloads,” but also provides an optimized infrastructure as a service that uses Oracle engineered systems to support Oracle customers running any kind of workload “including their most mission-critical workloads, their most highly available transaction workloads, and their biggest data warehouse workloads.”
In-Memory Database Technology
In Oracle Database 12c R1, Oracle first came out with the in-memory column store technology and in 12c R2, Mendelsohn said, “we are continuing to enhance that technology.”
According to Mendelsohn, the reason that column store technologies have not been popular historically is that while they are fast for analytics, they are slower for doing transactional processing. Oracle provides a “dual format methodology” that enables high performance transactions but also high performance analytics on transactional data, said Mendelsohn, pointing to benchmark results announced in September 2015.
Combination of SQL with big data
A big trend in the Hadoop community is a renewed appreciation for the value of SQL going natively against HDFS data, Mendelsohn said. In this space, Oracle has introduced a product called Oracle Big Data SQL which provides users with the ability to use Oracle SQL which customers already know, to run SQL against in a Hadoop HDFS, Oracle relational databases, or Oracle NoSQL key-value store database, “and so with one Oracle SQL query, you can scan through all your data across all these data sources,” said Mendelsohn. “This is a very exciting technology in the big data space because the amount of data is so big now that people can’t afford to do ETL and move the data around from one repository to another. They want to just leave it in place. The ability to run this Oracle Big Data SQL query against data spanning multiple data stores is a really important innovation.”
Time of Database Transition
With the introduction several years ago of Oracle Database 12c, “we made major investments in the architecture of the product to meet the needs of the cloud” re-architecting the product to have a multitenant architecture, and with Oracle Database 12c R2 which is now in beta, “we are continuing to evolve the product,” he noted. Further underscoring Oracle commitment to the cloud, Mendelsohn said that when Oracle Database 12c R2 becomes generally available, it will be released as a cloud-first offering, followed by an on-premise distribution.
“Oracle has been around a long time. This is not the first big transition we have gone through,” said Mendelsohn. “In the database group, we are very used to trying to understand what the transition is all about, and evolving our product to meet the needs of the next generation. If you can’t evolve your product when these big generational changes are happening, you are consigned to be a relic of the past and become a legacy product.”