AgiliWay, a software development outsourcing company (https://agiliway.com/), in partnership with a software distributor, Softprom (https://softprom.com/), has supplied to ROI4CIO and implemented AWS, which has improved website performance by 8 times. Implementation project timelines: 28.09.2017 – 18.10.2017: optimal server set-up chosen, optimal AWS services selected, complete transition to AWS performed, including the domain name. Issue: Services used by the website prior to AWS implementation did not support operations with “heavy” functionality, such as processing large volumes of data. For this reason when the number of website users reached a certain figure, CPU and RAM indicators of the server used to rise to 100% and the website crashed. Solution: Thanks to AWS server and services the number of concurrent users of the site has increased several times, whereas the workload on the site and database is half as much as before. The website performance has improved by 8 times, however its cost of ownership has grown only twofold. ROI - 800% Issue:

• "The team has come up against the problem of insufficient website performance. When during testing the number of concurrent users reached a certain figure, the response time delay exceeded 10 seconds. As workload increased, the database failed to support it and the service “refused” to work,” - Oleg Pitsyk, IT Systems’ Architect with ROI4CIO, Agiliway.

Solution: Before the decision was taken to implement AWS for the ROI4CIO project, it was tested whether another cloud-based solution can potentially be used. Comparison of the testing results has conclusively proven superiority of AWS. Softprom in cooperation with Agiliway has developed the optimal solution enabling to enhance the site performance. Representatives of Agiliway, acting as developers and architects of ROI4CIO systems, were in charge of application server and database set-up. Softprom, a supplier of AWS platforms, acted as a consultant with regards to deploying AWS functionality. On top of that Softprom employees have configured the services of monitoring, notifications and AWS Lambda. The website domain was hosted in Amazon Route 53, highly available and scalable cloud Domain Name System service (DNS). Anti DDoS AWS Shield wall was deployed to ensure comprehensive protection from all known infrastructure attacks (levels 3 and 4). A virtual EC2 (C4 type) server was used to optimize workloads, which required high computing capacity, to configure automated scaling of resources and inbound loadbalancing. At website peak load times computing resources are automatically scaled, which allows to support virtually any workload. Amazon Relational Database Service was selected to enhance reliability and fail-safety, as well as to reduce administration spend. It performs the functions of hardware provisioning, database setup, patching and backups. ElastiCashe was also used to increase website throughput - it is a web service which simplifies rolling out and scaling functions in a cloud or in-memory store, as well as their management. AWS CloudWatch was used for extensive monitoring of applications and database. Provided relevant rules are in place, an administrator will be immediately notified by text or e-mail in case of any unexpected emergency. AWS SNS (simple notification service) was integrated in CloudWatch for messaging functions. Besides messaging in some circumstances serverless computations of AWS Lambda are triggered to fulfill automation functions. AWS S3 cloud storage automatically backs up the database and application server as scheduled.

• "Our efforts were mainly aimed at enhancing performance. However, a lot of attention was paid to fail-safety and backups. Quantitatively, we can probably also highlight Input/Output Operations Per Second (IOPS) of storage drive. We have used a high performance drive - Provisioned IOPS. Up to 40,000 IOPS can be provisioned for each DB instance", - Vlad Gavrilenko, CIO of Softprom.

• "The server was migrated very quickly, in two days literally. After that we were configuring the services, which has taken about a week," -  Vlad Gavrilenko, CIO Softprom.

Outcome: By implementing AWS the team has made it possible to increase the resource performance by eight times and at the same time to automate labour-intensive administration tasks. And it should be noted that cost of ownership has risen only twofold. After the query code was patched and all services have been implemented the site maintains stable and uninterrupted performance even at high load levels.

The Challenge
After maintaining on-premises hardware and custom publishing software for nearly two decades, The Seattle Times sought to migrate its website publishing to a contemporary content management platform. To avoid the costs of acquiring and configuring new hardware infrastructure and the required staff to maintain it, the company initially chose a fully managed hosting vendor. But after several months, The Times' software engineering team found it had sacrificed flexibility and agility in exchange for less maintenance responsibility. As the hosted platform struggled with managing traffic under a vastly fluctuating load, The Seattle Times team was hamstrung in its ability to scale up to meet customer demand. Tom Bain, the software engineering manager overseeing the migration effort, says, "We had a fairly standard architecture in mind when we set out to do the migration, and we encouraged our vendor to adapt to our needs, but they struggled with the idea of altering their own business model to satisfy our very unique hosting needs."
Why Amazon Web Services To address these core scalability concerns, The Seattle Times engineering team considered several alternative hosting options, including self-hosting on premises, more flexible managed hosting options, and various cloud providers. The team concluded that the available cloud options provided the needed flexibility, appropriate architecture, and desired cost savings. The company ultimately chose Amazon Web Services (AWS), in part because of the maturity of the product offering and, most significantly, the auto-scaling capabilities built into the service. The Seattle Times' new software is built on the LAMP stack, and the added benefits of native, Linux-based cloud hosting made the most sense when choosing a new vendor. The Seattle Times developed a proof-of-concept and implementation plan, which was reviewed by a team from AWS Support. “They looked over our architecture and said, ‘Here are some things that we recommend you do, some best practices, and some lessons learned,’ ” says Rob Grutko, director of technology for The Seattle Times. “They were very helpful in making sure we were production ready.” After implementing the desired system architecture and vetting the chosen components and configuration with AWS, The Times deployed its new system in just six hours. The website moved to the AWS platform between 11 p.m. and 3 a.m. and final testing was completed by 5 a.m. — in time for the next news day.
How Seattle Times Uses AWS Seattletimes.com is now hosted in an Amazon Virtual Private Cloud (Amazon VPC), a logically isolated section of the AWS cloud. It uses Amazon Elastic Compute Cloud (Amazon EC2) for resizable compute capacity and Amazon Elastic Block Store (Amazon EBS) for persistent block-level storage volumes. Amazon Relational Database Service (Amazon RDS) serves as a scalable cloud-based database, Amazon Simple Storage Service (Amazon S3) provides a fully redundant infrastructure for storing and retrieving data, and Amazon Route 53 offers a highly available and scalable Domain Name System (DNS) web service. The Times is using Amazon CloudFront in front of several Amazon S3 buckets to distribute a huge collection of photo imagery. The combination of Amazon CloudFront and Amazon S3 is used to embed photos into news stories distributed to The Times readers with low latency and high transfer speeds. Additionally, Amazon ElastiCache serves as an in-memory “cache in the cloud” in The Times’ new configuration. The Times is also using AWS Lambda to resize images for viewing on different devices such as desktop computers, tablets, and smartphones.
The Benefits With AWS, The Seattle Times can now automatically scale up very rapidly to accommodate spikes in website traffic when big stories break, and scale down during slower traffic periods to reduce costs. “Auto-scaling is really the clincher to this,” Grutko says. “With AWS, we can now serve our online readers with speed and efficiency, scaling to meet demand and delivering a better reader experience.’’ Moreover, news images can now be rapidly resized for different viewing environments, allowing breaking-news stories to reach readers faster. “AWS Lambda provides us with extremely fast image resizing,” Grutko says. “Before, if we needed an image resized in 10 different sizes, it would happen serially. With AWS Lambda, all 10 images get created at the same time, so it’s quite a bit faster and it involves no server maintenance.” Rather than relying on a hosting service to fix inevitable systems issues, The Times now has complete control over its back-end environment, enabling it to troubleshoot problems as soon as they occur. “When an issue happens, we can go under the hood and troubleshoot to get around nearly any problem,” says Grutko. “It’s our environment, and we control it.” When the company encounters a problem that it can’t solve, it relies on AWS Support. “Our on-boarding experience was quite good with the AWS support team,” says Miles Van Pelt, senior development engineer at The Seattle Times. “It really felt like they went out of their way to answer our questions and research topics that we couldn't readily find in their extensive documentation.” By choosing AWS, The Seattle Times is now better positioned to deliver in its pursuit of being a leading-edge digital news media company. “By moving to AWS, we’ve regained the agility and flexibility we need to support the company’s journalistic mission without incurring the expense and demands required of a pile of physical hardware,” says Grutko .