Data is the lifeblood of every modern organization. Regardless of where a company may be in its own digital transformation journey, to operate successfully, today’s businesses need real-time access to insights about their revenues and costs, products, customers, partners, employees and more. That’s why enterprise databases are often considered the organizational crown jewels, and just a single minute of database downtime can cause widespread application outage and, eventually, translate into thousands of dollars of lost opportunity.
In traditional IT architectures, technology teams spent an inordinate amount of time nurturing and maintaining databases for optimal reliability and performance. For example, resources were deployed to manage indexes, apply software changes/patches, allocate system storage, run data loads and backups, maintain security and more. And in cases where there is a high volume or velocity of data, performance is often dependent on the network between the database and the applications it supports.
To achieve greater agility, businesses are increasingly moving mission critical workloads to the cloud, with hybrid multicloud being the most common approach. According to the RightScale State of Cloud Report 2019 from Flexera, 84% of enterprises have a multicloud strategy and 58% of enterprises have a hybrid strategy.i In this cloud-first world, how can organizations ensure that they have best possible database and application performance when the architecture tiers are split across a single cloud provider boundary?
Avoiding the slow lane in multicloud architectures
In their journey to cloud, many enterprises either end up or at least consider deploying their applications and databases into different clouds. This can result in a number of challenges such as inconsistent performance, increased complexity and security concerns, particularly if the clouds are connected over the public internet. High-volume workloads moving large data sets will not perform well in these scenarios. Latency and speed are also of utmost concern when it comes to data-intensive processes where large datasets need to be synchronized between large databases (data warehousing), or where applications need instantaneous access to databases (online transaction processing).
Equinix and Oracle are helping organizations solve these challenges by providing them with direct and secure interconnection between Oracle Autonomous Databasein Oracle Cloud and applications in other cloud platforms. With the recently announced Network Edge on Platform Equinix®, this cloud to cloud routing challenge is quick and easy to solve. Network Edge provides virtualized network functions that are pre-integrated with Equinix Cloud Exchange Fabric™ (ECX Fabric™), a software-defined interconnection solution that connects multiple cloud platforms over high-speed, low-latency virtualized connections on Platform Equinix.
Handing over the keys: Oracle’s self-driving database
Offered as a cloud service, Oracle Autonomous Database uses machine learning to deliver a database that is “self-driving, self-securing and self-repairing.”As noted above, there are two use cases where Oracle Autonomous Database is particularly advantageous because it is optimized to address their workloads: data warehousing and online transaction processing.ii,iii
A data warehouse (DW) is a central repository of integrated data from one or more disparate sources that is used for reporting and data analysis. It is characterized by a relatively low volume of transactions, but queries tend to be very complex, long running and over a large dataset. The key priorities for a DW are high data throughput and data quality. The Oracle Autonomous Data Warehouse service supports fast, complex queries and is optimized for DW scenarios.
Online transaction processing (OLTP) is characterized by a large number of short, online transactions where there may be a large number of concurrent users. The system is expected to respond immediately to user requests, so the main characteristics are high throughput and low latency for snappy response times. Examples include online orders, retail sales and financial transaction systems. OLTP transactions can often span a network and include more than one company. The key priorities for OLTP applications are availability, fast query processing, high throughput and transaction atomicity and consistency. The Oracle Autonomous Transaction Processing supports a high volume of simple transactions and is optimized for OLTP scenarios.
Both Oracle Autonomous Data Warehouse and Oracle Autonomous Transaction Processing provide self-tuning capabilities. Using machine learning algorithms, the database automatically tunes internal database structures and optimizes SQL execution as data volumes change or as new access structures are created over time.
Source: Oracle
Automated performance is only as fast as your network
Both DW and OLTP use cases require either the lowest possible latency or high throughput or both. Running reports or analyses in a data warehouse is very data-intensive. Complex queries may need to access large volumes of data to generate aggregated results. A slow network in this case could mean that your report runs for days instead of hours. It’s even more critical in the case of online transaction processing where a lag of even a millisecond could mean an abandoned cart or lost customer.
In a previous blog, I discussed the role that physics plays in the design of hybrid IT architectures and why speed, volume and gravity matter. A key step that enterprises often overlook in the rush to modernize their IT infrastructures is the need to optimize their network. Cloud services like Oracle Autonomous Database offer organizations an unprecedented opportunity to harness advanced technologies like artificial intelligence, the internet of things and virtual reality at a fraction of the cost that custom applications would take. However, best of breed multicloud services are only as good as the network that connects them together. This network needs to be as agile and dynamic as the cloud service, and support the volume and velocity of the data and application workloads running over it.
Getting your network to the speed of cloud
With Network Edge on Platform Equinix, businesses can modernize their networks as quickly and easily as they move to cloud services. Network Edge enables organizations to select, configure and connect virtual network services from industry-leading vendors like Cisco, Juniper and Palo Alto Networks to access cloud platforms in minutes—without additional space, power, hardware or equipment. This makes it simple to securely deploy and interconnect virtual network services such as routers, firewalls, private network terminators and WAN load-balancers with less complexity and cost. Once a virtual connection is provisioned, accessing an Autonomous Database tier on Oracle Cloud to an application tier on another cloud is instantaneous.
A Network Edge virtual device works in tandem with Oracle Cloud Infrastructure FastConnect and ECX Fabric to easily and quickly establish virtual network connections to Oracle Autonomous Database in Oracle Cloud (see diagram below). ECX Fabric is available in 35+ metro locations and interconnected across Equinix International Business Exchange™ (IBX®) data centers around the world, enabling secure access to cloud platforms that may not be local to a business’ data center locations.Oracle Cloud Infrastructure FastConnect with ECX Fabric enables businesses to connect a database or application tier on any cloud to an Autonomous Database tier in Oracle Cloud on demand.
In the video below, I walk you through how to connect to Oracle Cloud using Oracle Cloud Infrastructure FastConnect, ECX Fabric and virtual network devices created on Network Edge.
Learn more about Oracle Cloud on Platform Equinix and signup to test Network Edge.
[i] Flexera, Cloud Computing Trends: RightScale 2019 State of the Cloud Survey, Feb 2019.
[ii] Oracle, Oracle Autonomous Database White Paper, Sept 2018.
[iii] Oracle, Frequently Asked Questions for Oracle Autonomous Database, last updated May 2019.
Subscribe to the Equinix Blog