AWS Cost Optimization: How To Save On Managing Services
Amazon's diverse portfolio of web services is attracting more and more businesses and individuals. Navigating the complex ecosystem of services is not easy, and a poorly thought-out combination of services can drive overhead costs sky-high. How to avoid unnecessary extra costs?
Amazon Web Services (AWS) provides its public cloud platform to millions of users worldwide. Key benefits of using AWS include DevOps-ready tools and ease of deployment.
Thanks to the comprehensiveness of the AWS portfolio, users can have cloud infrastructure, backup, data recovery, Kubernetes and other services from a single provider, mostly automated and in many cases “as a service” – without concern for the platform itself.
Migrate And Manage AWS Services
The choice of tools, migration and service management in AWS can be a challenge even for the most experienced, and potential mistakes often have a negative impact on the budget. At the same time, the tools can be configured so that you do not pay unnecessarily extra and their management can be outsourced.
But such comfort is not free. Therefore, those interested in AWS services are concerned about how much the new solution will cost them as a result. But, as with most public cloud providers, AWS pricing is unpredictable.
Calculating the price of a solution at Amazon has become a separate discipline. AWS even offers its users a range of tools to monitor spend and analyse the most costly areas of infrastructure.
In addition to costs and future invoices, those interested in the public cloud then consider the issue of migration. Is it worth it to relocate all applications? To keep them as they are or to upgrade the platform they run on? Is the application still useful to us? Wouldn’t it be better to replace it with a third-party solution?
With a growing number of unanswered questions and a selection of appropriate tools from an extensive portfolio, the vision of a comfortable operation begins to seem unattainable. Designing, migrating and operating in AWS doesn’t have to be a nightmare – just reach for a managed solution.
Third-party AWS Management: How It Differs From AWS Managed Services
Amazon itself also offers a service for managing its services. It includes maintenance of hardware, virtualization software and partly other software services. However, it does not cover the configuration of network elements, firewalls, regular updates of tools and the operating system.
An alternative, often more affordable, way is to get support and administration for AWS services from a third party that knows Amazon’s extensive portfolio well and can design a hybrid solution that provides users with cost optimization on data processing, for example.
Compared to the official managed variant of AWS, the major benefit from the partner side are:
- customized infrastructure design, also in hybrid variant
- 24/7 live support
- 24/7 service monitoring
- security patches, encryption, access authentication
- network security, network traffic filtering
- regular system updates
- deployment and configuration of new tools, their maintenance
- performance and cost optimization
- close cooperation, consultation
In this option, the customer only provides application-level management. However, as in many other cases, the concept of managed AWS services can vary across providers, so it’s importnant to check what the AWS service provider will be responsible for in the first place.
Practical Examples Of AWS Administration
Many companies and organizations are still hesitant to migrate to AWS, sometimes rightfully so. The public cloud certainly isn’t for everyone, but fears about it can be unnecessary in many cases, and many businesses can benefit significantly from this “step into the unknown”. So in what cases will AWS services make infrastructure operations more efficient?
1. Reaching Foreign Customers
When expanding into foreign markets, it’s worth considering a different location for your server infrastructure. Reaching customers in the United States will work better from a local datacenter than from Central Europe. A local IT operation will ensure lower latency and therefore a nimble website, which will increase the probability that customers will successfully send their first order.
“For such a case, you can run a virtual instance of EC2 in Ohio, for example, and insure the speed of loading web, images and backups by connecting object S3 storage. There is no need to give up local data processing. The infrastructure in Ohio can be connected via VPN to the current solution in the Czech Republic,” says Petr Bělousov, senior administrator and specialist for managed AWS at MasterDC.
For infrastructures operated outside the European Union, organizations may face legislative problems and, for example, due to GDPR, they are forced to store data within the EU. The ideal solution is to proceed to a hybrid cloud and operate part of the infrastructure in AWS and part in the Czech Republic.
Dictionary Of Selected AWS Services
EC2 – Amazon Elastic Compute Cloud is a part of Amazon’s virtualization platform where you can rent some computing power. These are virtual servers.
S3 storage – object storage from Amazon, which can be managed via a web interface. It is used for backup, disaster recovery, data archiving and operational data casting.
S3 Glacier – file storage that is primarily used for backup and archiving data that does not need to be accessed regularly. The loading time of data stored in S3 Glacier is in the range of 3-5 hours.
Aurora – relational database service compatible with MySQL and PostgreSQL.
Dynamo DB – a proprietary NoSQL database that is provided to users in as a service mode. Thanks to synchronous replication across multiple datacenters, it ensures high data durability and availability.
SNS – using the so-called simple notification service, notifications are delivered in bulk, especially to mobile device users
CloudFront – A CDN (content delivery network) provides a distributed network of proxy servers that store content of voluminous media, such as video, and enable its rapid delivery and downloading.
Route53 – DNS (domain name system), is responsible for routing users to various AWS services and beyond, monitoring the status of applications and their endpoints. The service supports end-to-end DNS resolution over IPv4 and IPv6.
ALB – Application Load Balancer, developed by Amazon, works at the request level (layer 7) and directs traffic to individual virtual servers or IP addresses. It is ideal for load balancing HTTP andHTTPS traffic. It is also optimized for routing requests in modern microservices-based architectures.
Amazon EKS – Amazon-maintained Kubernetes platform for managing container services without the need to install and maintain your own Kubernetes master instance.
2. Cost Optimization For Graphics Data Processing
Application developers and operators working with large amounts of graphics data are reluctant to move their traffic to the public cloud, mostly because processing this data in AWS becomes more expensive. “Data can be processed outside of AWS on servers with clustered graphics cards running in a Czech datacentre,” suggests Belousov.
AWS Direct Connect provides low latency for transmission to Amazon EKS, where the frontend applications run. If set up well, this solution can save up to tens of percent in infrastructure costs.
3. Applications With Microservice Architecture
Mobile applications based on microservices operating in the global market require demanding requirements for stable high availability and low latency. A cost-effective solution for these cases can be designed in AWS.
“The range of services can be really wide – from Aurora databases, Dynamo DB and SNS notification services through CloudFront CDN, Route53 domain name system, ALB application load balancer and S3 storage including Glacier to Amazon EKS for the java application itself,” says senior administrator Belousov.
Such a robust solution can be operated in, for example, five regions to provide global coverage. For example, Frankfurt, Northern Virginia, California, Hong Kong and Mumbai.
Because the entire environment is Infrastructure as Code (IaC), it is provided and configured through code (not manually), making it easy to modify and distribute configurations. Duplication of the environment to a new region can thus be accomplished very quickly, ranging from tens of minutes to days, depending on the amount of customization required and the complexity of the infrastructure.
Migrating To AWS: How, When And What To Move?
Moving a complete infrastructure, especially if it is extensive, can take several months. A more cost-effective solution may therefore be a hybrid mode, where part of the infrastructure runs in a local datacenter and part in the public cloud.
Another option is to consider migration as an opportunity to modernize or revise the application portfolio. Remove legacy tools that could cause compatibility issues during migration, replace others with a SaaS model and save not only on operations but also on internal capacity.
The final form of the infrastructure is always individual and it is worth consulting with someone who has a lot of experience with the design and management of AWS services. This will avoid unnecessary, often recurring, complications.