Community Blog Why PrestoMall Stops Using Oracle but Chooses ApsaraDB for PolarDB

Why PrestoMall Stops Using Oracle but Chooses ApsaraDB for PolarDB

This blog shows how Southeast Asian companies like PrestoMall are saying goodbye to Oracle and introduces the relational and non-relational databases.

Goodbye Oracle! Exclusive Insights into PrestoMall's Transition to Alibaba Cloud

PrestoMall is a Southeast Asian company specializing in e-commerce that was established in 2014. Until recently, the enterprise had been using Oracle databases to host their business services and applications. All their services were supported by a set of Oracle databases, whereas another set of Oracle databases were used to support their test environments.

In August 2018, three months before their Oracle database software license expired, Prestomall decided to stop using Oracle databases all together, looking for an alternative. So, why did they decide to stop using Oracle anyway? What alternatives were out there for them? And just how did the entire migration process work in the end?

Why Did PrestoMall Stop Using Oracle?

In recent years, with the development of mobile Internet throughout Southeast Asia, PrestoMall experienced a boom in growth. In terms of revenue scale, the company achieved a growth of 256% over the past three fiscal years. With PrestoMall's business successes, the amount of data they need to process started to increase sharply, which made their IT costs also soar. To ease these costs, PrestoMall saw getting away from Oracle as a solution.

One reason for this conclusion was that the licensing fee for Oracle was a sizable cost in their budget. They felt it was too high. Moreover, as the number of PrestoMall's customers and its business data volume increased, this expenditure eventually came to account for a majority of their IT budget, which greatly restricted and hampered business development and expansion at the company. Therefore, three months before the license was due to be renewed, PrestoMall began to actively and aggressively seek out alternatives.

In addition, with the rapid development of their business, the vertical architecture that used Oracle databases also restricted the elasticity of PrestoMall's IT systems—something that stopped them from further growth. The main problem was that traditional databases are not powerful enough to meet the rapid development of Internet Plus, which led PrestoMall to decide they needed to say goodbye to Oracle.

Choosing the Best Option

The main motivation for PrestoMall to part ways with Oracle was reducing its IT spending. Therefore, early in the process of looking for alternatives, IT costs and fees were naturally a major concern. Plus with the current license about to expire, quick migration was also an extremely important consideration.

Initially, open source databases were all preferred. Regardless, PrestoMall came up with three alternative:

  1. Use a commercial database product other than Oracle, such as IBM DB2 or Microsoft SQL Server.
  2. Use an open-source database product, such as MySQL or PostgreSQL.
  3. Continue using their Oracle databases and renewing the license when it was due.

PrestoMall wanted to stop using Oracle and free themselves from being locked down by traditional commercial database vendors. Therefore, the first and third options were far from ideal. Although retaining Oracle or using another commercial database product would be the quickest and easiest option, it was far from the best option. PrestoMall would only consider one of these options if there were no applicable alternatives or there was not enough time to implemented their preferred solution. PrestoMall team was aware that this is a "now or never" situation, as not switching now would make switching in the future all the more difficult.

So, after evaluating their options, the PrestoMall team shifted to using open-source databases. Among the alternative open-source database products out there, PostgreSQL provided more SQL functions than MySQL and was a less expensive option, so PostgreSQL was naturally their top choice.

After the initial considerations, the Prestomall team performed a further detailed assessment of the migration solution by diving into the following concerns:

  1. How compatible would the selected database system be with the existing Oracle databases?
  2. Would the existing local deployment solution be able to be retained or would a cloud-based migration solution need to be implemented?
  3. How intense would the migration process be and how long would it take?
  4. Is there sufficient technical support to adopt the new technical solution?

In the end, they decided to go with Alibaba Cloud, given both the support for open-source PostgreSQL and their good and quick migration services. In particular, they decided on using Alibaba Cloud's database management system solution ApsaraDB for PolarDB, and use Data Transmission Service (DTS), and Alibaba Cloud's Advanced Database and Application Migration (ADAM) service for migration.

At the beginning, PrestoMall wasn't looking for a solution deployed in the cloud. Their original Oracle databases were hosted locally. However, since PostgreSQL was selected as the language of choice, and Alibaba Cloud's PolarDB was their database management system of choice, they decided to migrate their services to Alibaba Cloud.

Let's discuss the products and services that PrestoMall decided on:

  1. Alibaba Cloud ApsaraDB for PolarDB is somewhat of a unicorn in the database management systems of the world. It is a distributed, cloud-native database system that was independently developed by Alibaba Cloud. The service was developed with high demand industries like the gamming industry and e-commerce in mind. PolarDB is fully compatible with MySQL, PostgreSQL, and also highly compatible Oracle syntax.
  2. Alibaba Cloud Data Transmission Service (DTS) is a migration tool that supports data communication among RDBMS, NoSQL, OLAP, and other data sources, incorporating the data migration, subscription, and synchronization process all together.
  3. Alibaba Cloud's Advanced Database and Application Migration (ADAM) service supports the complete and comprehensive migration of Oracle-based systems.

One question remains, however, why exactly was this solution ultimately chosen in the end? Yang Lin, the director of Alibaba Cloud's database and application migration product solutions, said that the client, PrestoMall, told him it was due mainly to three main reasons:

First, cloud migration can easily adapt to future business development. Previously, PrestoMall used locally hosted Oracle database systems. By moving to Alibaba Cloud with its database solution of ApsaraDB for PolarDB, the company could start to take advantage of the elastic scalability that cloud database systems have to offer meaning that they could in turn request resources on demand, responding to sudden changes in business. This, of course, is the best possible solution for an e-commerce enterprise.

Second, the migration cost is optimal. Here cost not only refers to the monetary costs related to using database resources after migration but also includes the migration process in all of its complexity, including code modification and migration time, among other things. Through a thorough investigation, it was found that ApsaraDB for PolarDB was highly compatible with Oracle database systems, and therefore all the associated migration costs could be reduced significantly.

Third, risks are controllable and technical support was guaranteed.

In 2000, Alibaba began to use Oracle databases. But in as early as 2008, Alibaba decided to stop using Oracle databases for other solutions. As you can see, the experience of PrestoMall mirrors that of Alibaba. Over the past decade or so, Alibaba's technical staff have figured out and perfected many of the details involved in migration from Oracle-RAC databases to Aliabab Cloud database solutions like ApsaraDB for PolarDD, both from an on-premises data center and from another cloud environment. Throughout the journey, Alibaba has accumulated lots of experience, from which we managed to developed products like ADAM and DTS. These success stories were highly attractive to PrestoMall during the decision process.

ApsaraDB for PolarDB

At the initial proposal phase, the Alibaba Cloud Database team provided a highly detailed transformation plan based on ADAM. This plan included insights into how to implement automatic mapping, automatic parsing, and automatic conversion at the database level. Also, this plan explained how to modify code at the application layer. At the same time, code modification analysis was performed on the compatibility of different database products.

De Mai, the product director of Alibaba Cloud's ApsaraDB stack on the international site, noted that, after performing the ADAM analysis, we found that, without ADAM, we would need to modify more than 80% of the code when migrating from Oracle to PostgreSQL. However, by using ADAM to migrate data to PostgreSQL, only about 10% of the code needed to be modified, and merely 5% for migrating data to PolarDB. This was exactly another important reason why ADAM and PolarDB were included into the final migration solution.

Relational vs. Non-Relational Databases

Since 2016, Alibaba Cloud has focused its attention on self-developed cloud-based databases, and the extensive range of data management services it can offer to its customers. It's these services and solutions that can provide your business with the data backup, monitoring, migration, and disaster recovery it requires to operate within this data-heavy world.

Database management systems tend to fall within two groups: relational databases (also known as SQL databases), and non-relational databases (otherwise known as NoSQL databases). Alibaba Cloud offers a number of products in both categories and can help find the right database solution for your business.

This article lists a small selection of the solutions Alibaba Cloud offers.

Related Products

ApsaraDB for PolarDB

ApsaraDB for PolarDB is a cloud-native relational database compatible with MySQL, PostgreSQL, and Oracle. ApsaraDB for PolarDB provides the performance and availability of traditional enterprise databases and the flexibility and cost-effectiveness of open-source databases.

ApsaraDB for PolarDB is designed for business-critical database applications that require fast performance, high concurrency, and automatic scaling. You can scale up to millions of queries per second and 100 TB per database cluster with 15 low latency read replicas. ApsaraDB for PolarDB is six times faster than standard MySQL databases, and delivers the security, reliability, and availability of traditional commercial databases at 1/10 the cost. ApsaraDB for PolarDB embodies the proven database technology and best practices honed over the last decade that supported hyper-scale events such as the Alibaba Double 11 Global Shopping Festival.

Related Documentation

Specifications and pricing

This topic describes the specifications and pricing of PolarDB, including billing methods, billable items, and the prices of compute nodes, storage space, data backups, SQL Explorer, and global database networks (GDNs).

User Guide

PolarDB is a next-generation relational database that is developed by Alibaba Cloud. PolarDB is compatible with MySQL, PostgreSQL, and Oracle engines. Each Cluster Edition cluster supports a maximum storage capacity of 100 TB. You can scale out nodes in a single PolarDB cluster to a maximum of 16 nodes. PolarDB applies to the various database scenarios of enterprises.

PolarDB uses an architecture where computing is decoupled from storage. All the compute nodes share one set of data. PolarDB allows you to upgrade or downgrade configurations in minutes, and supports fault recovery in seconds. PolarDB ensures global data consistency, and offers free services of data backup and disaster recovery. PolarDB incorporates the highlights of commercial databases: stability, reliability, high performance, and scalability. It also has the advantages of open source cloud databases: ease of use, openness, and self-iteration. PolarDB for MySQL is fully compatible with native MySQL and ApsaraDB RDS for MySQL. You can migrate data from MySQL to PolarDB for MySQL without modifying code or configurations of your applications.

0 0 0
Share on

Alibaba Clouder

2,624 posts | 723 followers

You may also like