On Analytics, Data Platforms and Smart Applications
Oracle steps up data and analytic services as part of its platform-as-a-service push. New Exadata, Big Data and Analytic cloud services are all in the mix, but the strongest appeal is to all-red-stack Oracle shops.
Are you ready to run in an all-red-stack cloud? Oracle CTO and Executive Chairman Larry Ellison did a masterful job on Monday, during an hours-long Oracle Platform-as-a-Service event, of making that sound like the only sensible choice an enterprise can make.
Ellison stuck to familiar themes and messages that have been very consistent for Oracle over last two years. He and fellow Oracle executives presented at least 24 new services said to complete what they described as the most complete cloud platform available, encompassing infrastructure as a service (IaaS), platform as a service (PaaS) and software as a service (SaaS). As my colleague Holger Mueller points out, Oracle is still catching up in the infrastructure as a service (IaaS) arena, where one of the new services it announced was the Oracle Archive Storage Cloud. But the bulk of yesterday’s announcements were about Oracle’s PaaS.
It’s here at the platform level that Oracle is bringing all of its middleware to bear, supporting services for analytics, application development, content and collaboration, data management, integration and mobility all around a common core of management capabilities.
Given my focus on Data to Decisions research, I’ll focus here on Oracle’s Business Analytics and Data Management announcements. You have to manage data before you can analyze it, so let’s start with the new data-management services, which include the Oracle Database Cloud Exadata Service and the Oracle Big Data Cloud Service. As the names suggest, these services are based on the Engineered Systems of the same name, providing another example of Oracle’s oft-repeated message that what you run in the cloud is identical to what you run on-premises, making it easy to move workloads into the cloud and vice-versa.
Oracle Database Cloud Exadata Service
Oracle Database services have been available for several years, but now in addition to Standard Edition and Enterprise Edition in the cloud, you can now get the Exadata Service level. This is the Extreme Performance Edition on Exadata steroids, with every database feature available, including Exadata and RAC scalability, high-availability, Data Guard, and the In-Memory Database option. Use cases include mission-critical OLTP, data warehousing, in-memory analytics or all of the above, with multiple database instances supported though the same service. You also can use the Exadata Service as a disaster-recovery backup for an on-premises Exadata deployment.
MyPOV: You better have serious database needs, as this service demands a big commitment. Subscriptions are handled on a monthly or annual, not hourly, basis. You can scale up or down elastically, but the meter reading is on a month-to-month basis. There’s also a minimum subscription starting point of a quarter rack, with 28 cores and 42 terabytes of storage capacity. That’s big, though clearly less daunting than buying an Engineered System and running it yourself. If you want a heavy-duty option but aren’t ready for this level of commitment, there’s always Oracle Database Enterprise Edition Extreme Performance, the highest-level DBaaS cloud option short of the Exadata Service.
Oracle Big Data Cloud Service
Here’s another cloud-based way to try out (or avoid buying) an Oracle Engineered System. The Oracle Big Data Cloud Service gives you access to the same software delivered through the appliance, including Cloudera’s flavor of Hadoop, the Oracle NoSQL database, and Oracle R and open source Spark software for big data analytics. There are three ways to get data into this service: a high-speed connection to the Oracle Storage Cloud, an Infiniband connection to the Oracle Database Cloud Exadata Service, or a direct connection (option) to your data center.
Once the data is in the cloud, you can access data through secure shell (SSH), site-to-site VPN or optional high-bandwidth connections to your data center. Clusters are provisioned through the Oracle MyServices Web-based management interface and they’re managed through Cloudera Manager.
MyPOV: As with Exadata in the cloud, the Oracle Big Data Cloud Service entails a monthly subscription, so don’t expect hourly or per-day rates. I scoured the documentation in search of minimum subscription levels, but I didn’t find and restrictions. I would not be surprised to discover that there’s a minimum number of nodes.
Oracle executives talked up the Oracle Big Data SQL Service as a complement to the Oracle Big Data Cloud Service, but the data sheet specifically says it lets you query across Hadoop, NoSQL and the Oracle Database Service Exadata Edition.” That suggests it’s not available with plain old (non-Exadata) Oracle database services. I’m not sure whether this dependency is tied to Exadata Infiniband connectivity or Exadata processing power, but that’s a big commitment to make to gain cross-platform Big Data SQL querying.
Business Analytics Cloud Services
On this front, Oracle has introduced (or at least announced) five important new services:
The Big-Picture POV
The data- and analytics-related services discussed here represent roughly one third of the total Oracle PaaS landscape discussed this week. Nonetheless it’s the important stuff for data-management and big data professionals.
As I said up front, Oracle has been nothing if not consistent about its cloud plans in recent years, and it certainly has put together a comprehensive platform. The message that “what you run on premises also runs in the cloud” is also compelling, particularly if most of what you run is from Oracle.
Ellison and Thomas Kurian also stressed that you can bring any app, not just Oracle apps, into the company’s cloud. But if you run a heterogeneous data center with lots of third-party middleware and perhaps even third-party databases, I suspect Oracle’s IaaS is not nearly as inviting. Oracle likes its all-red stack, and it’s bent on making it the most compelling choice it can offer in the cloud.
We also heard repeatedly about the ability to move from on-premises to the cloud and back again. That’s a comfort to those who are sticking to development, testing and disaster recovery in the cloud. But for companies that moving into the cloud in a big way, it’s more of a one-way street. In that context, the question is, do I want to start fresh rather than migrating everything I have on-premises into the cloud? That changes the equation and brings many more non-Oracle options into consideration.