Sensis search backend
goes AWS native
Sensis leverages the flexibility and modern services of Amazon Web Services (AWS) to redevelop its core Search API (SAPI) suite of applications in AWS - part of an overall strategy to exit its data centres and optimise operational efficiency using DevOps and CI/CD
Sensis is Australia’s #1 marketing services company. The Sensis purpose is to engage Australian consumers with businesses, and today delivers on that promise through its leading digital consumer businesses (Yellow Pages, White Pages, TrueLocal, Whereis, Skip), search engine marketing and optimisation services, website products, social, data and mapping solutions and through its digital advertising agency, Found. Sensis is also Australia’s largest print directory publisher including the Yellow Pages and White Pages.
In 2014, a majority stake in Sensis was acquired by US private equity firm Platinum Equity – a business known for driving down the cost base of its acquisitions to extract maximum value. As the ink dried on that deal - the Platinum Equity executive team, working with Sensis CEO John Allan, began a re-alignment of the business to ensure adoption of their lean cost ethos. The onus was placed on each Sensis business unit to right-size for cost-efficiency, and IT was not immune. The question was posed; ‘How can Sensis continue to innovate its digital portfolio, while reducing its IT operating costs?’.
As part of this broad push to drive down expenditure, Sensis engaged Mantalus to assist with an organisation wide assessment of its applications to determine migration pathways to Amazon Web Services. One of those platforms was the Sensis Search API (SAPI).
SAPI is the primary index and search service powering the YellowPages, Whitepages and Whereis websites and mobile apps. Yellowpages itself remains one of the most visited websites in Australia, making SAPI a widely integrated and sensitive platform in the Sensis portfolio. SAPI itself, whilst remaining under continual development, comprised of multiple legacy applications - designed and built using outdated architectures and 3rd party products, as well as a significant portion of in-house developed capabilities. Core functional components, such as the SOLR search index had been in-place through multiple refresh cycles without significant re-investment by the business. This left the platform lagging behind modern search technologies.
Mantalus worked with internal Sensis teams to assist in early phases of application categorisation and 6 R's treatment for migration planning. Through that process, SAPI was identified as a platform that would remain key to the overall Sensis strategy, but one that could be modernised to utilise native AWS services and significantly reduce its running costs / infrastructure footprint. This would be achieved by re-architecting the platform to take advantage of the flex-scale and serverless capabilities AWS has to offer, and re-writing the application for resilience and automated recovery across as many of its components as possible.
Mantalus outlined a migration pathway for AWS by converting existing discrete 'services' of the platform into microservices, and implementing modern development and deployment methodologies. The platform's resulting AWS architecture refactored its componentry at all tiers, replacing ageing middleware and operating systems with available AWS services. On-premises servers running Solaris and RHEL were replaced with EC2 running AWS Linux. SOLR search was replaced with ElastiCache and ElasticSearch. Mongo DB was replaced with RDS (Postgres), Mashery was replaced with API Gateway and other services were incorporated into the modern architecture including Simple Queue Service (SQS) - all connected back to on-premises integration points using DirectConnect.
Mantalus successfully assessed, designed, managed and delivered the AWS build of a complete refactor of the Sensis SAPI platform. Using a strong and repeatable methodology and approach, Mantalus applied its 'secret-sauce' to ensure a successful large-scale AWS migration of this core Sensis platform. A large part of that approach was ensuring the Sensis IT team developed a culture of 'buy-in' to the outcome. Mantalus fostered that buy-in by involving the existing Sensis teams in the project, helping to up-skill their incumbent resources. This is another key pillar in the Mantalus methodology. AWS might all be about pay-as-you-go infrastructure, but at the foundation of that is human capability. By investing in their staff and ensuring they are part of the journey, enterprises can galvanise the entire organisation to support their cloud migration goals.
Not only did the migration approach help to future proof the SAPI platform for the long term, it also allowed Sensis to decommission over 350 Virtual Servers in its on-premises environment, and avoid renewing licensing and support agreements with multiple incumbent vendors.
Sensis put full credit for the success of the SAPI migration in the hands of Mantalus. CIO Baden Roberts summarises his thoughts on Mantalus’ capability; “Not only did the Mantalus team drive a compelling argument for why we should migrate, but they over-delivered in a way rarely seen nowadays. That’s the power in selecting the correct AWS Partner for the task at hand”.