The decision between a Software-as-a-Service (SaaS) or in-house respectively on-premise solution is a crucial aspect of any IT strategy. This is particularly important for large enterprises looking for competitive advantages in a highly competitive market.
At the beginning of the cloud era, many companies thought that this new type of infrastructure could solve all their IT problems. But very soon they realized that both on-premise and cloud have their advantages and disadvantages. In other words, CIOs and CTOs are still often torn between the two when it comes to understanding which model is best suited for their specific use case.
The right choice depends on various case-specific requirements, circumstances and challenges. To make the right decision, it is therefore beneficial to get familiar with the most significant differences between the two models.
Before we dive into these, however, we look at the history of both first.
The evolution from on-premise data center to the cloud
Data centers played a central role in the evolution of modern business as we know it today. In the late 1950s, for example, American Airlines cooperated with IBM, which together developed the very first reservation system for passengers. This was the first time that computers were connected in such a way that the network could be used by people all around the globe to enter data, retrieve information and do business. Possibly inspired by this project, IBM drove innovation and designed the first mainframe computers.
Until the introduction of personal computers (PCs) in the 1980s, these mainframe computers were clearly dominant in terms of information processing . In the early 90s, however, companies started to use microcomputers as cheaper alternatives to mainframes. They piled them up in piles so that they could reproduce computing power similar to that of mainframe computers. They called these data centers. So it's actually not surprising that all those companies who worked with a lot of information gradually switched to this type of on-premise facility.
Between 1999 and 2001, the next wave of on-premise innovation followed. Responsible for it this time, was the introduction of the VMWare Workstation. This enabled IT teams to create new virtual servers and thus, optimize their use of resources. Another critical development took place in 2002, when Amazon Web Services (AWS) launched its cloud-based service business. It's when data hosting evolved into a more complex concept. With these new possibilities, companies could now use computing power and data storage without ever having to build their own IT infrastructure or adapt it during operation. As this business developed rapidly and represented an interesting opportunity, other technology companies such as Microsoft and Google jumped on the bandwagon. Today, companies have a very wide range of options to build their business in the cloud, on-premise or in a hybrid form (hybrid forms are not further discussed in this blog article series).
Decision-making: Cloud or on-premise?
When you come to the point of making an important decision on cloud versus on-premise as part of your IT strategy, it is imperative that you think about your data before you argue about how the software should be delivered. The following questions may give you some food for thought here:
- Do you have specific requirements or business needs?
- How large is the relevant data volume in the scope of your project?
- Is the possibly already existing hardware sufficient for these requirements?
- What are the scaling possibilities in relation to the anticipated ramp-up phase?
- What are the possible data costs that you already have (i.e. hosting fees, IT infrastructure, maintenance & support, and other internal costs)?
- How many users need data at which touch points?
Let us now take a closer look at the concrete characteristics of the two infrastructure models categorized according to some of the most important aspects.
If we compare the acquisition costs of cloud and on-premise solutions, we quickly find that those of cloud services are much lower. In some cases even massively lower. It is precisely because of this cost efficiency that such solutions are very interesting for small and medium-sized companies. It is therefore not surprising that we can observe a trend towards the cloud among SMEs. Anyhow, it is an exciting movement when we consider that SaaS was once exclusive to big corporations. Nonetheless, I think it's fair to say that for some services, the annual cost for the subscription is high and in the long term, in purely nominal terms, can exceed the cost of on-premise solutions. But for many, the great flexibility that the cloud offers is worth that premium. In addition, these companies use the advantages of the cloud to minimize costs for internal IT resources and to outsource all IT support to the cloud provider. An additional benefit of this approach is the reduction of complexity in budgeting, as costs are very predictable, which also makes your CFO happy.
As already indicated, the acquisition costs of on-premise are relatively higher. Moreover, there are annual costs for maintenance & support (usually somewhere between 15 and 20 percent of the purchase price), service costs for setup and configuration, regular advice on updates and the like, costs for internal IT maintenance and of course costs for server hardware and licenses.
Especially in our business, the document processing, on-premise solutions incur additional and higher expenses, because you need a setup and training phase for new document types and you depend on permanent consulting for OCR improvements. In part 2 of this series you will learn more about these things.
Implementation and integration
In terms of implementation, cloud solutions take significantly less time to set up. Normally, customers can start operations immediately after signing up. This means that such services leverage their existing platform that they have already implemented, provisioned and tested, and simply add more customers. Another advantage of their nature is that you can quickly build different software integrations and thus, connect or optimize various services and business processes.
In contrast, implementing on-premise requires the physical presence of different employees of the provider. The whole procedure can then take several days or even weeks, and as a result, it's more expensive. Integration with other services is also possible but tends to be more cost-intensive.
Individual configuration of the solution
As far as configuration is concerned, cloud solutions are generally very limited and users can usually make only minimal adjustments, if at all. If adjustments are possible, however, they can be implemented quickly and easily without the need of any support.
On-premise solutions have a clear advantage here. They offer a very high degree of flexibility and can be adapted almost completely to individual needs. This also creates more context knowledge and control over the way processes are handled and information is stored or displayed.
Maintenance & Support
Since cloud providers take responsibility for the consistent availability of the service, and also cover security and potential recovery measures, you do not need to worry about these issues at all. If something comes up here, it will most likely be limited to the validation review or possible individual configuration.
With on-premise this is a little different. Because maintenance, guarantee of continuous availability, security and recovery must be organized internally. In return, a company that makes use of such a solution has ownership of the data and thus, at least in theory, more control.
While cloud solutions can be scaled to an unlimited extent and can therefore always be adapted to the needs of a company, on-premise requires early planning. For companies that have volatility in their business, such solutions are not really suitable.
With respect to updates, the advantages are clearly on the side of cloud services too. Updates are continuously and only very rarely installed with internal IT involvement. And all of this normally without any direct additional costs.
If, on the other hand, you want to update your on-premise solution, you have to plan this carefully and in longer project work. In addition, you have to reckon with further costs for the update and your IT has the responsibility for the deployment and validation of the update.
Security and compliance
Security is a very exciting topic in this context and may be counterintuitive for many. Because a relatively small or even medium sized on-premise system is only very rarely the target of cyber attacks. Nevertheless, you have to protect it anyway as there is still a potential danger and as a result, your company possibly at stake. This means that you still have to plan and mobilize additional time and resources for these things. On top of that, there is the compliance work, which you also have to do yourself. But since your team has complete control over the IT environment, this should be pretty easy to do.
With cloud solutions, especially with high-end providers, you get top-notch security talent that takes over the supervision of the server and network and always tries to protect you from all kinds of attacks. They also have a corresponding amount of resources available for this purpose. You will be able to profit from top-notch talent, because the larger and more valuable the amount of data managed by these services is, the more interesting is the target for hackers. And without appropriate protective measures, the cloud provider's business would be at serious risk. In terms of regulatory compliance, cloud operators offer a clearly defined service. Many are working with the SSAE 16 (SAS70) audit in this context.
In the next part of this blog article series, I will specifically address the pros and cons of cloud and on-premise solutions in the OCR area, explain our approach to a global document platform, and try to draw a conclusion on the topics covered in both articles.