Most businesses operate in dynamic operating environment with rapidly changing customer expectations, and preferences. This is forcing businesses to be creative in delivering value to their customers and partners through applications or application programming interfaces (API) with new features and capabilities daily or even hourly.
Read More: 1Z0-432: Oracle Real Application Clusters 12c Essentials
Let’s consider a typical airline mobile app, it let's users manage reservations, get flight updates and keep track of membership bDocumenefits. However, not in the distant past, generating electronic boarding passes, seat selection and tracking bags in real-time used be an offline business process.
Businesses are increasingly depending on developers / IT teams with an established continuous delivery process in meeting customer expectations. Unfortunate reality is, business requirements for applications can often be vague, and it is not unusual for business leaders to demand the ability to define or change requirements frequently and within a short notice. This can cause lots of frustration between business leaders and development teams. Enter JSON.
JSON to the rescue
According to Wikipedia, JSON (JavaScript Object Notation) is a lightweight data-interchange format for web applications. JSON is programming language independent, it is easy for humans to read and write, easy for machines to parse and generate. It is these properties that make JSON an ideal data-interchange language. It derived from JavaScript, the most popular language among developers for nine years in a row according to 2021 Stack Overflow survey.
In short, JSON data format allows applications tolerate changes to application model and supporting data model without incurring application downtime. Additionally, JSON can support structured, semi-structured or unstructured data sources, which are becoming common in modern applications. The popularity of JSON has resulted in evolution of document databases that are suitable for high-volume web applications.
Evolution of document databases
Armed with JSON and with promise of speed, agility and optimized for productivity, development teams started using document databases for data persistence. Document model closely resemble JSON and can handle massive data sizes, without need for upfront schema modeling. These databases have become an alternative to fixed relational schema with tables and columns that are required in relational databases. Do they?
While a switch from relational to JSON document storage may offer significant advantages for application developers, it can lead to significant challenges for other consumers of this data. 79% of businesses executives say they don’t thrive when they operate in data silos and are increasingly relying on data analytics for faster innovation, improved efficiencies and ROI. Unfortunately these are not very well supported by document databases and precisely the need for database with converged capabilities. Mind the gap!
In pursuit of ideal data persistence
An ideal database should support open standards such as Structured Query Language (SQL) and process transactional workloads (OLTP) alongside analytics workloads (OLAP) without performance degrade. Developers shouldn’t be forced to add custom application logic and orchestrate complex, fragile, and inconsistent data movement between different specialized data stores just to support Graph, XML, Spatial data types and fast analytical queries to report across multiple and different JSON documents.
1. API driven Access:
Developers should have schema flexibility for their applications with ACID transaction guarantees. Better yet, a simple API-driven access and model-specific languages, while still having recourse to powerful SQL capabilities whenever they want.
Developers should be able to simply load JSON documents into the database natively, generate a schema and indices from JSON objects for transactions, and have ability to join with spatial, graph, and relational data for analytics.
2. Futureproof application needs with Multi-workload support
Application usecases evolve as the businesses evolve, different kinds of database workloads require different kinds of software optimizations. For example, smart sensors track millions of checked bags daily that pass through various airports, taking frequent measurements need a database that can ingest a large number of new records extremely quickly.
But training machine-learning models is a very different kind of job. For example, airline app should be able to provide personized lounge recommendations to its customers and suggest popular activities from their affiliates. This involves extremely large numbers of relatively simple calculations which means a lot of CPU cycles, including hundreds of common machine-learning algorithms.
3. Freedom to get job done
Developer productivity improves when a core database engine fully supports different kinds of application workloads without requiring excessive customization or app tuning. Likewise, data productivity improves when that same core engine also fully supports different analytical workloads critical to data scientists and business analysts. Developers need freedom to spend more time coding, more time learning, and contribute more to community by leveraging cloud and complete machine-learning based automation.
Optimized and fine tuned for JSON workloads
Oracle Autonomous JSON Database is part of Oracle Autonomous Database family that runs on Oracle Exadata engineered systems that are optimized for database workloads, and finely tuned for JSON workloads. Autonomous JSON database delivers industry leading performance with single digit microsecond response times that significantly outperforms specialized NoSQL document stores without sacrificing transactional consistency. This makes it an ideal persistent store for microservices based applications that can constantly adapt to business requirements and tolerate variability in demand. As with any Autonomous Database Oracle Autonomous JSON has full support of SQL including complex joins and long running transactions.
2X faster and 35% cheaper than MongoDB Atlas and AWS DocumentDB
Yahoo Cloud Serving Benchmark (YCSB), an industry standard open source tool that provides common set of workloads for evaluating the performance of different types of NoSQL data stores and Database as a Service (DBaaS) offerings available in market today. We used it to evaluate performance against MongoDB Atlas.
YCSB consists of YCSB Client, an extensible workload generator and core workloads, a set of workload scenarios to be executed by the generator. Source code for YCSB is available and can be downloaded from GitHub
We welcome you to run your workload by following these steps to configure and run the YCSB benchmark against Autonomous JSON Database
◉ Step 1: Provision an 8 OCPU instance with I TB storage. Follow these steps to provision the AJD instance.
◉ Step 2: Provision a VM.Standard2.24 compute instance on Oracle Cloud Infrastructure (OCI) in the same cloud region as Autonomous JSON Database for generating test load. For some workloads, you may need to more than one YCSB client to drive traffic. If so, repeat this step to provision more client compute instances
◉ Step 3. Configure and setup YCSB workload by cloning the GitHub repo. Follow instructions to install and run YCSB for multiple workloads
In our internal evaluation, Autonomous JSON Database with 8OCPU and 1 TB storage configuration that is comparable MongoDB Atlas M60 and AWS DocumentDB on R4.4xlarge configuration consistently outperformed for different type of workloads. These workloads are performed on 4 Million and 81 Million documents with each document.
◉ The first workload performs update-heavy activity, invoking 50% reads and 50% updates of the data.
◉ The second workload performs a medium-range scan that involves 75% scans and 25% updates.
◉ The third workload performs a short-range scan that involves 95% scans and 5% updates.
We used the lower end of performance improvements in the following comparison graph, however, your workloads could run much faster.
Simplified data platform for current and future application needs
Developers, IT and Business leaders should all work together towards primary goal of serving customers with flexibility and independence. Developers should have freedom to choose technology that makes them more productive, IT should have a unified data tier that makes reuse of data and overall management simpler and business leaders, flexibility to make data-driven decisions. An Autonomous JSON database is same as an Autonomous Processing database, except that JSON database is limited to storing 20GB of non JSON document collections. If your application needs change overtime, you can do a 1-click in place upgrade to Autonomous Transcation Processing database without any data migration or application rewrite. Ultimately, what this means for enterprises is a simplified data platform to both digitize business processes quickly and offer valued services more confidently for current and future business needs.
Source: oracle.com
Posts
0 comments:
Post a Comment