At Linke we are aware that the current worldwide health and economic situation is shifting companies priorities, not only on their project management and workforce daily activity but also in terms of infrastructure, solutions and services' investment.
To overcome the challenges involved in these difficult times, companies are evaluating the implementation of cloud-based technologies or, in case of those that already benefitting from cloud services, are taking essential cost saving measures using all the power provided by cloud computing models.
Saving on system and hardware investments (maintenance and management) is one of the priorities to ensure business continuity. From products to carry out optimum remote work to the reduce costs assessing services, the AWS public cloud solution is one of the best options.
In this particular article we would like to share five key measures to help you reduce AWS cloud infrastructure costs during this situation. The below recommendations are a summary extracted from our experience with customers, lead by our AWS Reseller Manager at Linke, Agustina Percio Roque.
Saving in the AWS Cloud
In order to meet customers' IT requirements needed to face the current panorama. Cloud systems and investment is rapidly growing and is positioned as one of the most flexible solutions with easily accessible and adaptable resources, always with the guarantee of high availability of business data at any time.
Scalability, security and availability are just some of the main benefits offered by Amazon Web Services. Let's see how to take advantage of these functionalities and some recommended actions that will help companies reduce the costs of IT departments in terms of cloud infrastructure expenditure:
1. Start a Rightsizing process
We recommend reviewing the size of each server, commonly called resizing/rightsizing process, in order to readjust them into families and sizes of Amazon EC2 instances offered by AWS. The goal is to improve cost efficiency but without neglecting the technical requirements of our servers.
During this process, we can use the available tools within the AWS platform and/or specific software providers such as CloudCheckr, in order to analyze that our servers are correctly sized according to their usage (in terms of memory and CPU, for example) and thus give guidelines on how the expense could be reduced.
This is an exercise that is usually carried out with clients, but considering the current situation, the flow of needs must be verified during these weeks and take advantage of the flexibility of the Cloud, adapting to the current circumstances of workloads.
2. Reviewing Amazon EC2 Servers and AMIs
In addition to the rightsizing exercise, it will be necessary to review all those stopped or idle servers, to proceed to their removal if they are not to be used in the near future or to stop them during certain hours if their activation can be reduced depending on the use.
A key tool for this process will be Amazon CloudWatch, an AWS service that will automatically detect inactive instances and proceed to deactivate them if we indicate so.
In relation to these servers, an important concept within AWS are the AMIs (Amazon Machine Images), since they contain the key information for launching EC2 instances.
As mentioned above, in case of having a stopped server (not used at this time) but that would be necessary in the future, it is proposed to create an AMI of it to launch it when necessary. In this way, during that time we replace the cost of EBS storage with the cost of maintaining an AMI.
3. Reviewing EBS volumes and cleaning up Snapshots
As with instances, it is possible to reduce costs based on storage efficiency. We recommend that you review all unnecessary EBS (Elastic Block Storage) volumes and dispose them off.
These volumes are easy to create, use, encrypt, and protect. And why are they so important? Because they are highly efficient for the most demanding workloads, including critical applications such as SAP, Oracle and Microsoft products. The cost impact will depend on the type of volume and storage capacity you can reduce.
How can we identify unnecessary volumes? Those that are not being used are shown in blue as "available" within the AWS platform, this will indicate when a volume is not being used by any instance and we can evaluate its disposal.
Another option to save costs, is to create a Snapshot, that is to say a backup of that volume at a certain point in time. This change implies a cost saving since we would stop paying the monthly provisioning of those volumes (example: General Purpose SSD, $0.10 per GB-month of provisioned storage) for a constant GB cost per month of the EBS Snapshot (0.05 per GB-month of stored data) .
Existing Snapshots can also be cleaned, since it could be that the current retentions of the same ones are more frequent than necessary. Therefore, if there is no need to store information more than a certain time, because our system does not require it, we can proceed to remove the snapshots that exceed this time threshold.
Finally, if our backups are based on an AMI policy (article point 2) we should identify all those Snapshots that are not part of an AMI. Once this identification has been made, we should proceed to remove them.
4. Instance Reservation
Once points 1 and 2 (rightsizing and cleaning of servers) have been carried out, the possibility of making reservations of Amazon EC2 Instances or EC2 Saving Plans must be evaluated, always considering what the use loads will be during the coming months.
At the time of instance bookings, we have some parameters to take into account that affect the possible savings to be generated in the costs of compute usage, they would be as follows:
- Region / location of the servers.
- Type of instances: families and sizes.
- Commitment period: 1 or 3 years.
- Payment method: total or partial payment at the beginning of the commitment, or directly only by establishing a commitment (without any payment in advance).
It is important to note that the more commitment established, the lower the cost per hour and vice versa, which is why we recommend detecting which servers must remain in time (taking into account the infrastructure needs of each client) and which can be deactivated or paused as explained in the first points of the article. And the same happens with the "payment method" variable, the higher the advance payment, the higher the discount percentage we will get.
5. Costs savings on Amazon S3
Finally, as we have mentioned, an important aspect to review is resource retention. At this point we will talk specifically about the object storage service in Amazon S3 (Amazon Simple Storage Service).
AWS provides us with different S3 storage options depending on the frequency of access required to the data (objects). With this in mind, it is proposed that all information requiring less frequent access be transferred from Amazon S3 standard to Amazon Glacier. Since the latter service allows long-term backup and archive storage with a recovery option from 1 minute to 12 hours; with a cost reduction per GB of up to 80% less than in standard Amazon S3 storage.
With these various tips, every company can benefit from the elasticity offered by AWS resources and make sure that all those resources that are not needed at the moment can be stopped in order to reduce costs. The advantage is that no critical data will be lost and instances can be reactivated when desired.