The pros and cons of On-Premise to the Cloud

Every organization in which IT plays a crucial role has an IT infrastructure. Think of servers and networks: the Operating System. Some companies have their IT infrastructure on-premise (local). But an increasing number of companies are moving their architecture to the cloud. But what are the benefits? And what are the drawbacks? That is what I describe in this blog.

Organizations choose to maintain the infrastructure on the basis of costs, knowledge, and existing infrastructure for management or self-management. Not all organizations manage their own IT infrastructure. For example, you can also outsource the management, which can be done internally, also known as insourcing, or you can place the management outside the organization (outsourcing). Both forms relieve the organization of worries.

Companies that manage their own IT infrastructure are usually companies whose business process outputs are physical end products (CAD / CAM processes) or organizations that have enough space for servers. These companies often have their own in-house development teams and system administrators / network administrators.

From on-premise to cloud

Companies where the IT infrastructure is on-premise experience limitations in the scalability of the infrastructure. It is also difficult to optimize and renew processes. As a result, they are overtaken by the competition and lose market. (due to overpriced products as a result of ineffective processes).

Also, there are fewer and fewer (information) systems available that can be installed on the on-premise infrastructure. This involves two culture changes:

1. The disappearance of Technical Application Management. The systems are in the cloud and the technical maintenance is also carried out by the supplier. Functional management directs the supplier.
2. Current BI processes and knowledge no longer meet the requirements. SaaS suppliers do not provide access to their database, but make extracts available via REST API.

Especially for companies that deal with Big Data (which includes the three V’s: Volume, Variety and Velocity) it is not possible to scale up sufficiently compared to the amount of data they receive. Problems also arise with the processing of this data. This shows that the current infrastructure and chosen platforms no longer meet the requirements of today and the future.

Scale-out Architecture

The solution to this problem is a scale-out architecture in the cloud. Instead of equipping the hardware more heavily, more (virtual) machines are clustered. In this way, computing power and memory are used in a more targeted manner. This enables us to scale up and down linearly in both performance and volume. In addition, we want to store data we receive from various providers as it is offered (“write before scheme”) without being held back by the limitations of relational database table structures. Then we will apply the big data semantics, so that the data is retrievable.

When implementing this solution, a change must first take place in the IT infrastructure. The result is a hybrid architecture. Existing systems continue to run on-premise and new systems (Greenfield) move to the cloud. Technical management is increasingly concerned with activities in the cloud. In the long term (say five to ten years), technical management is no longer required to manage your own equipment and platforms.

In order to respond to this change, a shift of competences must take place from traditional management to the cloud. To give an idea of how this shift can take place, I have made an overview of the activities that a traditional administrator performs and how they can shift to the cloud.

Server maintenance

Figure 1 shows how the maintenance of physical servers and operating systems is becoming less and less. The speed of the competence shift depends on the chosen services in the cloud. For this, there are three choices:

1. IaaS (Infrastruture as a Service) : Renting physical computing power, disk space, memory and an operating system
2. PaaS (Platform as as Service): Renting platforms including infrastructure (Database, Enterprise Service Bus)
3. SaaS (Software as as Service) Renting software, platform and infrastructure.

Management will be the lowest when SaaS services are selected and the highest with IaaS (see figures 1, 2 and 3). But management will always be there. Just in a different form.

Integration management

The need to connect different software packages to existing users, systems, networks and platforms will grow. In cloud terminology, we call this “Integration Management”.

There is also a shift from one-time costs for purchasing systems to the use of the chosen services in the cloud. Optimizing and balancing costs and usage will increase for the management team. Where the greatest need will arise is the integration of the various separate SaaS products with each other.

Workplace and database management

When selecting a Cloud PaaS Database to replace the on-premise database, the DBA activities regarding “Backup & Restore” are shifted to the cloud provider and cloud engineer. In addition, activities related to patching software throughout the DTAP street are taken over by smart “Building Pipe Lines”. As a result, this work is shifted to the developers.

Advantages of the cloud

For a transition to the cloud, it is first and foremost important that an organization must be able to work in an agile way to a certain extent. The Agile philosophy fits seamlessly with the possibilities offered by the cloud platform. Here are the biggest benefits of the cloud:

Scalability

In the cloud, it is possible to scale up and down “on demand” on both computing power (CPU, memory and writing space) as well as platforms and software. This ensures that the IT departments can work in a more customer-oriented manner and can focus more on customer needs.

Time-to-market

Because you no longer have to wait to scale and set up hardware, you can start realizing customer wishes more quickly. IT departments become flexible and customer-oriented.

Pay according to use

In the cloud you only pay for what you use. You can save considerably when your certain cloud services are not used. It is of course important that this is monitored, otherwise you will lose your advantage and this could even have a negative effect.

Availability

Data in the cloud has high availability as standard. This means that the IT department does not have to worry about storing the data redundantly.

Disadvantages of the cloud

The disadvantage of the cloud is that using data costs money. Transporting data to the cloud is (usually) free of charge. But reading data from the Cloud costs the data owner money. It is also impossible to estimate the use of data. This creates a certain degree of uncertainty.

Another disadvantage is the “vendor lock-in”. In addition to infrastructure, major cloud providers also offer platform services and software in the cloud that you will use more quickly if you have already moved the data to the cloud.

In addition, broader knowledge is required of the coherence of infrastructure, platforms and software services. This role is usually assigned to a solution architect or a cloud engineer.

Consequences

A transition to the cloud has implications for the culture and operations of current IT management. This department must be included in the decision to move to the Cloud. It is important for these people to be included in the process and receive coaching. Otherwise you run the risk that they will not support your vision and the business will suffer.

Future

In order to be able to innovate or optimize processes, it is necessary to have a flexible technical shell. To scale flexibly (with platforms and infrastructure) you cannot do without the cloud. That is why it is not possible for most companies to continue to work on-premise. A cultural change is necessary to ensure that the transition runs smoothly.