Gain Total Efficiency with Cloud
Survive from pandemic by moving to the cloud
When this article is written as the world attempts to recover from the economic devastation caused by lockdowns and social distancing during the pandemic, it is clear that more needs to be done in order to prevent a recession. Although the pandemic is not over yet, countries must begin to focus on rebuilding their economies. Otherwise, they will be met with an even greater crisis in the near future.
The government is preparing a series of incentives to jumpstart the economy, and the banking industry is being asked to be more flexible to facilitate micro and small businesses. This is seen as a necessary measure to help get the economy moving again.
As a national bank, Mandiri takes great efforts to restructuring credit and spending trillions of money. This step is taken wholeheartedly for the common good of Indonesia.
Even though our seniors have never explicitly requested to cut the budgets, we understand the situation and think hard to contribute to the company. We still need to develop and operate our business, and we still have targets to achieve.
As IT personnel in the banking industry, our activities focus on building and maintaining digital banking products. In order to make our operations more cost efficient, we first need to determine what our expenses are.
This is especially true when it comes to building and maintaining digital products, as our everyday spending is mostly funneled into buying support from third-party partners and investing in the infrastructure required to keep these products operational.
We did the sizing and estimating in the early stages of product development, factoring in the total number of predicted users, the processing units, memory, disk space, and network latency they would require, and then purchasing and allocating the infrastructure and supporting components accordingly.
However, this is a real risk, as if marketing does not reach the target, all the preparation will be for naught. We never know what will happen tomorrow, especially in these uncertain times.
The infrastructure is needed for the application and compliance with regulation. Expenditures from partners are expected for compliance with the regulation that states that every software product we use as a component requires support from external parties. This provides a level of comfort, even though it is not optimal or useful. There are rumours that these external parties exist only to be someone to blame when something bad happens.
Although they may not be as technically advanced as some of their competitors, this does not stop them from being confident in their abilities. They know that support for certain technologies is limited, and they also know that opening a new partnership with banking is not so easy, so they take advantage of it. Therefore they will charge a high price for their services. At least per-component we need to spend billions per year on this support.
An API gateway is a common component used in current technologies. It is used to manage customer identity and access, databases, message queues, and distributed caches.
Pricing for support services is usually based on the number of processing units or the amount of disk space used. This is similar to how infrastructure is priced, but multiplied by the number of years the support is needed. This means that the sizing of the support will affect the price of both the infrastructure and the support.
A more massive multiplier for sizing is not only on splitting and balancing the load, but to ensure services are always available to our users whatever happens to the bank, whether affected by an earthquake, fire, or any disaster. To accomplish this, we divide our physical servers into multiple locations and at least double it all the components above.
In order toaccumulate everything and not include development costs, only its components, and infrastructure, it will take at least tens of billions to spend. This is a difficult task to accomplish during the current pandemic.
Cloud Platform Solution
In 2006, Amazon Web Service introduced elastic Cloud Computing, which has gained popularity due to its elimination of the technical nightmares associated with server preparation and provision, especially for Small and Medium Enterprises and Technology Startups.
Netflix changed their luck when they moved from DVD rental and retailer to video streaming provider in 2008. They did this by scaling up their data center and moving all their infrastructure to AWS. This made them a first-class video streaming provider.
Netflix has been able to take advantage of the scalability of the Cloud to accommodate their growth. As they have added new customers, they have been able to provision additional processing power to meet the demands placed on their infrastructure. This has allowed them to provide an optimal experience for their users while also being cost-effective.
In 2016, Spotify decided to migrate its infrastructure to Google Cloud Platform in order to keep up with its competitors. By 2018, it had shut down its data centres and was entirely reliant on cloud-based resources.
If you think about the amount of effort it takes to maintain compute, storage and network capacity for a global company that serves more than 170 million users, that is a sizeable amount of work — Ramon van Alteren, Spotify
Resilient Application
Applications that are born in the cloud have an inherent resiliency. The number of instances will be adjusted to the application load, and there will be an ingress in front of them that is able to balance the load evenly. When one instance crashes, the cloud platform will bring a new one to the swarm, and from the user’s perspective, the service will always be available and perform well.
This resiliency not only applies to the application but also to the components such as the ingress itself as an internal cloud component, API gateway, customer identity and access management, databases, message queues, and distributed caches as additional user components.
If we choose to bring our own stack into the Cloud, we run the risk of having to manage and maintain the availability and performance of the components ourselves. However, if we use components that are native to the provider (such as Google Cloud Platform), we can rely on the provider’s engineers to manage and maintain them, ensuring that they are always available and performing well. This is a much more cost-effective solution, as we share costs with other customers of the provider.
No Up-Front Payment
This way of doing things eliminates the need to make a huge upfront investment in data center infrastructure, and instead allows us to pay only for the resources we actually use. This makes it a much more flexible and cost-effective way to scale our operations as needed.
Most cloud providers will bill us monthly, providing transparent resource usage reports. This allows for a pay-as-you-go approach that is rare for data centre providers.
In the traditional approach to high availability, we would have to set up at least two data centres, DC and DRC, to guarantee application time or availability. However, cloud providers usually serve out-of-the-box application availability for fully managed services. This means that we don’t have to pay for the DC-DRC, doubling the cost.
We don’t need to prepare a budget for support because it is included in our monthly billing. This is a more efficient way of doing things because the cloud providers are responsible for everything that happens on their platform.
It’s not mandatory to size early anymore.
No Hidden Cost
Building infrastructure always comes with a certain degree of uncertainty. We often find that the final price is lower than our original estimate, due to unforeseen factors such as supports, hardware, or even network costs. By being aware of these potential risks, we can better manage our resources and avoid any potential surprises.
We never would have thought that the price for activities as easy as installation would require hundreds of millions of rupiah to start. Even worse, on the network, pulling the cable is not as easy as drawing a line on a picture. When we implement a new network device, it will cost us not only hundreds of millions rupiah but also hundreds of thousands of US dollars.
This situation has never happened in the Cloud because our infrastructure is always set-up through a provider. This means that we always know the costs associated with our infrastructure, and there are no surprises. We simply pay our bill, and we can see all the invoice details.
Conclusion
To deal with this pandemic and its domino effect, we need to think smartly about where we focus our budgets. The Cloud Platform is one of the brightest choices we can make to survive and focus our money on saving others.
