Five ways to save on AWS costs

As an AWS account manager, saving customers money whilst enhancing AWS usage is at the heart of everything you do. It makes sense – if the customer can see the benefit in financial form, they will keep coming back and will further utilise what is available from Amazon Web Services

Cost optimisation for SAP customers migrating to AWS

Your customers are making the most of AWS for other systems, but how can you help them save further by moving SAP onto the cloud? These five ways covers everything from the compute side to budget set-up and cultural change but all will go towards ensuring your customers are utilising AWS to the fullest extent whilst reducing their overheads.

Introduce cost control as standard

To begin with, the focus of any migration project will be to ensure a smooth transfer from on-premise to AWS. Once that’s done and the customer’s systems and their use of it has settled into business as usual, the tactics need to change to cost control, and therefore saving. We suggest a monthly full financial cost control process to be provided as the customer matures their usage of AWS. This only takes a few hours each month, and the savings can be substantial – one customer saved $200,000 over a year. This is a huge saving that can be generated in a number of ways.

1Reserved instances returns

reserved instances

On the compute side there are two ways in which customers can save. Firstly, capitalising on the returns available when using reserved instances rather than on-demand. There are flexible options over one to three-year periods which can provide large financial discounts of between 40-75% on running each compute. Customers should also be made aware that the flexibility stretches to whether they would to prefer to make a down-payment or avoid upfront costs.

Making the most of the analysis available, even as standard using AWS tools, is key to monitoring and making the most of the opportunities available. Our Reserved Instance Analysis report goes a stage further, whilst using the standard Amazon recommendations in conjunction with it, to provide enhancement based on intelligence gathered during the customer relationship. This means that instead of just relying on automated decisions from analysis over a short period of 1-3 months, we can ensure business requirements inform recommendations based on how the customer has historically used and plans to use their system whilst making sure the set-up is correct for what they are running on AWS.

We also suggest customers be bold with their procurement of instances to make the most of the agile buying and selling process. By actively trading Reserved Instances on the Marketplace, especially when seeking those with a shorter lifespan for specific projects, customers’ can still make the most of the (up to) 75% cost reduction. On top of this, by understanding their usage pattern, customers can also make the most of Convertible Reserved Instances for flexible future trade-ins.

2 Exploit the 70% of the week where systems are not used


The other side of the compute usage is how up and downtime is managed. Traditionally systems run constantly, until irksome downtime interrupts working life, meaning that the costs tally in an upward direction. But our research has shown that many businesses actually only use systems like SAP on a 10 by 5 basis – 10 hours by 5 days which is only 30% of the week. By creating a schedule so that the service can be maintained in up- and down-time consistent with the business needs, customers can rake in yet more savings by making the most of AWS’s pay-for-what-you-use billing structure.

Of course, there are times when systems will need to be up outside of the 10×5.
Another of the functions of Lemongrass’s MiCloud is that there is manual on and off switches for instances such as a developer squeezing in some additional changes on a Saturday morning. There’s even an auto-stop function for when the systems idle outside of their set schedule (be that 10×5 or otherwise) – gracefully shutting down the system when that developer rushes out the door later on that Saturday and forgets to turn it off again…or the many other instances where the system lays idle when it doesn’t need to be on.

All of this is as simple as switching off lights to save money, and results in a cost savings of over £2,000 per month, achieved in seconds with no real impact on the user community.

3Set budgets against tags and manage them

set budgets

Tracking spend against budgets is fundamental to cost saving, but this process can be linked to needed changes in business culture so providing ways of facilitate this is crucial. AWS can be spliced using tags which can be related to what the systems are, who owns or controls them, the type of product is running on each resource. By using as many of these tags as possible, and with the aid of MiCloud CGS platform, monthly budgets can then be set against each tag in order to track individual areas costs against each budget. By grouping the tags together, business units as a whole can also be monitored for cost effectivity and control. If the customer finds that certain business areas are running over or under budget, they can drill down into the detail and make effective changes in an agile way.

The agility of this service is also found in how it is deployed. The MiCloud CGS platform is available on iOS or Android as an app which means business decisions can be made quickly and effectively, as well as providing real-time insights. This is particularly important to the customer’s executives and managers – the people that do care about cost – to see, on a day-to-day basis, what is being spent against plans and budgets. By using the app, trends can be easily spotted and expected behaviour can be checked in on to ensure actions are adhered to; for example, stopping systems over the weekend should show a dip in cost, and a cost reduction should show the same. By enabling strong business insight and oversight, customers can save 25% of their month-by-month AWS spend.

4 Storage cost management

storage cost

The flexibility of AWS means that the customer is not stuck with a one-size fits all solution to their data needs. Four ways in which to make sure the customer is makingthe most of this flexibility are:

  • Migrating storage to more cost-effective tiers
    A cool down can be automated by changing the characterisation of EPS volumes out of the hours (for instance the 10×5 schedule) from being held on high-speed platforms, such as GP2, to slower platforms. This change can provide an additional 22% saving, which isn’t just found in when the systems are idle or off line over weekends and evenings, but is spread throughout the week on an ongoing basis.
  • Application level data management
    SAP HANA dynamic tiering and data aging allows the customer’s SAP HANA database to store warm data on a separate dedicated host that houses the dynamic tiering service. The dynamic tiering service provides the ability to create multistore and extended store tables to store and process your warm data. This relieves the customer of older, less frequently accessed data to an integrated disk tier, whilst still being able to access it with excellent performance, and it lowers the cost of the customer’s SAP HANA system significantly.
  • SAP archiving
    This is the gift that keeps on giving. By moving unused data to deep glacial level storage such as S3, the cost is reduced to 0.9%. That is a 99.1% saving on the original storage costs which just shows how much wastage there can be! There is also the opportunity to use smaller instances, less CPU is needed, less storage is needed for back-ups – the possibilities and the savings increase tenfold with structured archiving.
  • Selecting the right storage type
    This means utilising the ability to change the compute sizes of the SAP systems, as well as the underlying storage transparently. We would suggest providing just enough compute to run the systems efficiently which would reduce the inherent costs of larger set-ups. AWS provides flexibility and speed with this process, unlike traditional on-premise, and you can change your mind and move to larger or smaller instances as needed.

Of course, there may be apprehension from customers who are used to the on-premise model that doesn’t have this level of choice, or they may have experienced complications whilst archiving in the past. There can also be concerns about the time needed for business leaders to decide which data can be archived. This is why Lemongrass provides technical archiving, as well as running hosting, to articulate the benefits and help customers through the process.

5Make the most of native tools

make the most

We’ve mentioned native tools that are available above, as well as Lemongrass products and services, but financial governance is close at hand when using AWS and this is a strong selling point for moving SAP from on-premise to the cloud. AWS Trusted Advisor has a range of ways to optimise infrastructures, and a clear and concise dashboard for clients, and indeed us, to monitor performance.

At the heart of AWS Trusted Advisor is the pillar of ‘Cost Optimization’, which can be used to report on and eliminate unused and idle resources or to make commitments to reserved capacity (see points #2 and #1 respectively). The other four pillars of AWS Trusted Advisor – ‘Performance’, ‘Security’, ‘Fault Tolerance’ and ‘Service Limits’ – all ,go towards cost saving across the customer’s set-up.