• During the past five years, Oracle has organized a systematic distribution of new Java releases on a six-month cadence, and it claims to be the project’s leading contributor.
• The Java development division comprises six projects that focus on various performance and feature aspects.

Recently, Oracle Inc. announced that Java 20, the most recent version of the well-liked programming language and development environment, is now available.

Thousands of performance, stability, and security improvements have been made in the language’s most recent version, including seven proposals to improve the Java Development Kit to boost developer productivity and enhance performance, stability, and security. The language has been around for 27 years.

During the past five years, Oracle has organized a systematic distribution of new Java releases on a six-month cadence, and it claims to be the project’s leading contributor. According to Tiobe Software BV, Java is the world’s third most used programming language, while Oracle claims that Java is the best language for organizational development.

Chad Arimura, Vice President of developer relations at Oracle, said, “The innovation pipeline has never been richer. The problem space is changing, and developers have higher demands on their programming languages than ever.”

Divided into Six Projects

The Java development division comprises six projects that focus on various performance and feature aspects. The project delivering new features most frequently is called Amber. According to Arimura, the language is continuously improving to become clearer and less prone to mistakes. It also catalyzes the ongoing transition to more data-oriented programming practices.

Enhancing startup and performance is the focus of Leyden. Loom aims to achieve massive parallelism without the requirement for asynchronous programming. Using Panama, Java native code can communicate with external libraries.

Valhalla, a long-term project, uses value types to increase memory density and improve performance in machine learning and data-intensive applications. With minimal pauses, the Z Garbage Collector aims to collect waste from big heaps.

Staged Previews

A preview system was recently implemented to gather feedback before incorporating new features into the Java standard. Previews last six months, but if there are a lot of comments, more cycles might be added. For instance, pattern matching for a switch is in its fourth preview in this edition. It allows an expression to be checked against various patterns, enabling sophisticated data-oriented queries to be represented succinctly and reliably.

Users can now nest record patterns and type patterns for declarative and composable data processing, which is a significant improvement. For increased usability, robustness, and speed, scoped values, which is still at the incubation stage, allow the use of immutable data within and between threads.

Virtual threads simplify the process of building, maintaining, and watching high-throughput concurrent applications by allowing developers to diagnose more easily, debug, and profile concurrent programs using the JDK’s existing tools and methodologies. Programming with many threads is made more accessible by structured concurrency, which treats each task that runs in a different thread as a separate unit of work.

By effectively calling foreign functions and securely gaining access to memory not controlled by the JVM, the Panama project’s foreign function and memory application program interface enables Java programs to collaborate with the code and data outside Java runtime. A new feature of the vector API boosts performance by describing vector computations that consistently compile to vector instructions on supported processors at runtime.

Oracle provides commercial Java versions for both on-premises and on the cloud, even though Java itself is open source. A single view of Java runtimes and applications used both on-premises and on any cloud is provided by the Java Management Service, a native service on Oracle Cloud Infrastructure. A pay-as-you-go option, the Oracle Java Universal SE Subscription offers support, access to advanced capabilities, and complimentary applications.

• The Oracle Container Engine for Kubernetes has a fully controlled control plane and is entirely compliant with Cloud Native Computing Foundation constructs.
• According to Oracle, when hosting Kubernetes on competing for public clouds, clients can save up to 50% and benefit from additional services not included in Kubernetes clusters.

Recently, Oracle Corp. unveiled new feature updates to its cloud-based Oracle Container Engine for Kubernetes, claiming that it can streamline operations, lower costs, and increase reliability and efficiency in large-scale systems using the Kubernetes orchestrator for software containers.

The improvements are intended for businesses using agile DevOps methods and constructs like microservices to build and run cloud-native apps on Oracle Cloud Infrastructure. Vijay Kumar, Vice President of product marketing for application development services and developer relations at Oracle, said, “Kubernetes is notoriously complex not only to operate but to find the people with deep skill sets. We’re dramatically simplifying the deployment and operations of Kubernetes at scale.”

50% or Less Costs Compared to Competitive Public Clouds

According to Kumar, Oracle supports Kubernetes in various runtime settings, from bare metal to serverless operations. The Oracle Container Engine for Kubernetes has a fully controlled control plane and is entirely compliant with Cloud Native Computing Foundation constructs. According to Oracle, when hosting Kubernetes on competing for public clouds, clients can save up to 50% and benefit from additional services not included in Kubernetes clusters. Kumar added that Oracle provides uniform pricing across all international zones to reduce complexity.

Leo Leung, Vice President of products and strategy at Oracle, said, “A big piece of Kubernetes is compute and on a computer-by-computer basis, we’re less than 50% of the list price of the lowest-cost region of other providers. Then there are additional parts of Kubernetes that require compute to boot up the cluster, and we’re lower cost there as well.”

The improvements come with virtual nodes, which let businesses run Kubernetes-based apps dependably and at scale without dealing with the operational burden of managing, expanding, upgrading, and troubleshooting the underlying Kubernetes node architecture. With usage-based pricing, Virtual Nodes additionally offer pod-level elasticity.

Leung added, “Customers that are deep into Kubernetes may want to have control over worker nodes to get fine-grained control over the infrastructure, such as running all pods inside bare metal. For the majority of customers, though, we believe serverless is the right answer. They don’t want knobs and dials. They want a service that’s going to scale.”

Encompassing Lifecycle Management

With complete lifecycle management that includes deployment, upgrades, configuration changes, and patching, the improvements provide enterprises with greater freedom in installing and configuring their preferred auxiliary operating software or associated applications. Add-ons include access to optional software operators, including the Kubernetes dashboard, the Oracle database, and Oracle WebLogic, and necessary software deployed on the cluster, such as CoreDNS and kube-proxy.

Controls for identity and access management at the pod level are now accessible. The number of worker nodes by default for newly-provisioned clusters has been raised to 2,000. Support for inexpensive spot instances has been enabled—service-level agreements with financial backing for the worker nodes and API server for Kubernetes.

Having the capacity to grow to thousands more nodes, Kumar said, “you can have a fairly large application running on a Kubernetes cluster without having all the networking between clusters.”

• The recent Arcion release also makes it easier for businesses to manage Data Definition Language (DDL) schema modifications and data transformation.

California-based Arcion (previously Blitzz), which provides a fully managed platform to copy transactional data to cloud-based data platforms in real-time, is all set to make data extraction from Oracle databases faster using a new native log reader.

The feature, a part of Arcion’s most recent release, will help enterprises read logs straight from their Oracle instance during replication. This will eliminate the need to use Logminer or other less reliable or less efficient sources. The company claims that this, along with its distributed and parallel architectural design, ensures unlimited scalability and 10 times faster data extraction to target systems like Databricks, Snowflake, MySQL, PostgreSQL, SingleStore, and Yugabyte.

“Arcion is the only end-to-end multithreaded CDC [change data capture] solution that auto-scales vertically and horizontally. Any process Arcion runs on source and target is parallelized using patent-pending techniques to achieve maximum throughput. There isn’t a single step within the pipeline that is single-threaded. It gives Arcion users ultra-low latency CDC replication and can always keep up with the forever increasing data volume on the source. If an enterprise wants to migrate or replicate terabyte-scale data that requires high throughput, Arcion is the answer,” Gary Hagmueller, the CEO of the company.

While newer data integration tools like Airbyte, Debezium, StreamSet, and Kafka Connectors do not provide this feature, older CDC tools (Qlik Attunity, Fivetran-acquired HVR and Debezium) offer this solution. However, as Hagmueller said, all of these traditional solutions demand significant setup and management effort, which is not the case with Arcion.

Facilitating the copying of data

The recent Arcion release also makes it easier for businesses to manage Data Definition Language (DDL) schema modifications and data transformation in addition to providing a native reader for Oracle customers.

As a part of the latter, the platform’s capacity to support schema evolution has been expanded to automatically capture DDL modifications from a source database and copy it to the destination data platform. Data engineers can save on time as they no longer have to manually maintain schema alignment between source and target databases, thanks to this capability. Earlier, they had to halt the replication process and restart it from scratch by taking a snapshot of the source system if the DDL or schema on the source database changed. As a result, this led to downtime, a waste of costly computing resources, and a possibility of human error and data loss.

“Oracle Golden Gate is one CDC solution that supports automatic schema evolution (DDL). But Arcion is the only CDC platform that supports out-of-the-box DDL with modern analytic warehouses like Snowflake or Databricks. Oracle Golden Gate does not provide very robust support for Snowflake and Databricks, so anyone adopting such systems will find that solution inadequate. Alternatively, the data team has to be ready to invest in manual resources to handle the schema evolution with other alternative CDC solutions,” the CEO noted.

Arcion is also launching a zero-code capability that provides adaptable, high-performance streaming column transformations on the fly to help businesses manage data transformations better. As a result, there is no longer a need to use technical resources to develop a staging table (like Kafka) and write customized code to modify data on the destination. SLA delays were also a result of the practice.

Availability of the Oracle log reader

The other two features are now generally available as part of the fully-hosted version of Arcion, while the Oracle log reader is now only available in beta and will be made more widely available later this month.

With this update, Arcion also includes Imply (established by the people who created Apache Druid) as a new target, Google BigQuery, and Azure-Managed SQL Server as new sources. The technology facilitates the replication of data across more than 20 enterprise databases and data warehouses. The business also obtained USD13 million in series A capital a few months earlier, valued at USD65 million.

“The data replication and protection software market showed much greater-than-expected resilience in 2020 despite the pandemic,” Phil Goodwin, research director at IDC’s infrastructure systems, platforms, and technologies group, said. “We expect this market to return to its normal growth pattern, with a 2.7% CAGR through 2025. The public cloud services portion of the market is the bright spot, with an expected 11.6% CAGR during that time.”