Learn how businesses like yours can begin to optimize for today and plan for tomorrow with Cloud-Ready IT Infrastructure

Oracle Exadata and Oracle 18c: Pushing the Efficient Frontier of Infrastructure

Guest Author

Innovation is the driver behind every new iteration of Oracle Database and Oracle Exadata. And the driver behind the innovation is giving businesses more performance, availability, scalability, and cost-effectiveness. In the words of Exadata Product Manager Gurmeet Goindi, Oracle Database 18c running on the Exadata platform continues to push the efficient frontier of IT infrastructure.  

Not one to make a statement without backing it up, Gurmeet provides plenty of evidence for why Oracle Database 18c and this engineered system are such a powerful combination.

Why the combination of Oracle Database 18c and Exadata?

Goindi: Exadata is our flagship platform to run the Oracle Database. After more than a decade, it’s still, by far, the most efficient, the best performing, and the highest available platform to run Oracle Database. While Exadata has massive and successful on-premises deployments in many leading enterprises, it is also the cornerstone of Oracle Cloud. Exadata runs our software-as-a-service (SaaS) applications, and provides database-as-a-service (PaaS). 

How has the increased focus on DevOps factored into enhancements to 18c running on Exadata? 

Goindi: Indeed, developers are increasingly dictating how the database should run, and we have to respond to that. One of the ways we’ve done that is to fully embrace a fast-provision, multitenant architecture—enabling up to 4,000 pluggable databases on the same system. 

Second, we added a feature that allows us to refresh or maintain a second copy of these pluggable databases and switch them over instantaneously—a kind of pluggable database failover feature. Say you have hundreds of developers working off the same machine, and the data on one of those pluggable databases becomes corrupted. You need a quick way to switch over that one particular database and not disturb anybody else. This architecture enables the customer to do that—a very useful feature when running in a DevOps environment.

Also, in this shared environment, you need granular data access control. We upgraded the security feature so that the pluggable databases support native encryption keys to isolate different developers or apps.

We also support Docker containers on Exadata, which allows developers to spin up new containers, stick their app or a new version of database in, and get it going. 

What types of memory innovations have you introduced in Exadata that enhances 18c performance?

Goindi: As a company, we believe the future of analytics is in-memory. In-memory columnar store delivers real-time analytics because nearly every business needs to run a real-time organization today. 

The first phase, which came out in our last Exadata release, used the same server architecture for storage that we have in the database tier, which makes the in-memory technology go very fast on the computer and at the storage tier as well. With this latest release, we store the data in flash in the same format as an in-memory columnar store, and we use the same SIMD technology. The hottest portion of the dataset remains in the computer memory, and the rest spills over into the flash tier; the entire dataset now benefits from the columnar format.

You still have datasets that need the core in-memory performance. With Database 18c on Exadata, we introduced a feature we call Automatic In-Memory that basically eliminates tuning and, instead, automatically evicts less-accessed data to flash and keeps the active dataset in DRAM. These two innovative features, the in-memory-flash formatting and the Automatic In-Memory, provide the performance of DRAM and the capacity of flash.

One important thing I want to note is that Exadata is the only platform on which you can run Oracle In-Memory on a standby database. This allows you to, for example, offload reports to a standby database and use the active database for ingesting new records. We even have a function that allows you to lazily upgrade your standby database to keep it in sync with your active database so that you don’t have to wait until the end of the day to upload billions of records to your data warehouse.

Would you talk about all the latest automation built into 18c and Exadata, and how it affects management and security?

Goindi: With new features or functions we add, we strive to simplify management by automating processes. First of all, the system comes pre-configured. Some features may have to be enabled, but they work in an automatic pre-configured fashion. Customers can tune if they want to, but the default is they don’t have to.

In the latest 18c release, we’ve added a seamless, automatic software update feature. What we’ve done is to create what is really an export system in which you specify the servers that you have to update. You specify a time period for the updates and the software version to update to. You might specify, for example, "Update my Exadata to the latest version on Saturday between 10 p.m. and midnight." This tool will wake up at 10 p.m. It will fetch the bits from wherever you have specified, and it will start updating the servers, one at a time in a rolling fashion until midnight. When midnight comes, it will stop, wherever it is. Then, next Saturday, it will resume where it left off.

If you're running on a non-Exadata system, and something like the Spectre and Meltdown vulnerabilities occur, to protect your system, you have to wait for every vendor to put its security update out. They all have to be lined up on the same day, and then all these teams have to work together to implement the patches because if even one component is not updated, you are not secure.  

In Exadata, one software update patches the updates and secures the whole stack top to bottom, and with only one vendor.

How does the latest functionality facilitate cloud-scale deployment?

Goindi: Since we run the same Exadata platform in the cloud, the business investment in Exadata is protected, and your transition from on-premises to the cloud is smooth, no matter how much you want to move to the cloud or how quickly (or slowly). 

At the same time, we make our applications run faster in a more available fashion. Built-in functionality allows native database charting across many geographies. It also provides management efficiencies of cloud scale, enabling management of all your systems as one.

Isn’t Exadata perceived as being the high-end, expensive platform that’s out of reach for many businesses?

Goindi: Exadata aims to address all customer requirements. Running Oracle Database on Exadata actually provides tremendous cost efficiency while delivering the best combination of performance, availability, and scalability. Because the customer gains all this productivity and reduces staff resources for day-to-day management, the cost-effectiveness keeps increasing as more and more innovation is incorporated. That’s why we call it the efficient frontier of IT infrastructure.

About the Author

Gurmeet Goindi is the Master Product Manager for Oracle Exadata at Oracle. 

Join the discussion

Comments ( 1 )
  • Amit Shivpuri Wednesday, September 26, 2018
    Hi Gurmeet,

    The insight on 'Automatic Software Update' is very interesting and quite useful. Can you point towards more technical details on it like what all is covered in this update, does it update database views also, what are the impact and level of human intervention/monitoring is required.

Please enter your name.Please provide a valid email address.Please enter a comment.CAPTCHA challenge response provided was incorrect. Please try again.