August 29, 2020 Refresh2Cloud - Building test systems in the cloud
Non-running business processes have never been good. This is not a completely new insight. Technical progress and ever more optimised business processes, however, ensure shrinking risk and security buffers in business processes in ever larger networks of effects and dependencies. The failure of central systems can therefore have a concrete and devastating impact on business operations and figures within minutes.
Companies naturally take precautions for emergencies. Virtualisation, clusters, sophisticated backup/restore concepts and all the associated organisational measures are the usual solutions. These are also all correct. But there is often some requirement they do not answer. Either they are very complex to deal with, or one of the dimensions from RPO, RTO or RCO is not sufficiently covered. Or they are limited to purely local scenarios.
In order to overcome these local restrictions, there is the classic option of working with remote systems and data centres. However, many companies still shy away from the associated costs.
The general reasons for the remote operation of failure systems are now sufficiently known from risk management: Major fires or floods, widespread network or power failures, or strikes and even sabotage. They all ensure that corporate managers, CRITIS officers, auditors and other bodies are increasingly coming to the conclusion in their risk assessments and risk mitigation considerations that availability and disaster recovery concepts at local or campus level are rarely sufficient.
At the latest, it becomes critical when the campus or an entire region is not available in terms of IT. It is quite clear that the local units cannot work in a meaningful way or also face completely different challenges. But when other company locations are forced to stop work because they cannot access central systems, negligence is now often found.
If companies operate distributed data centres anyway, this may play a rather subordinate role. In these cases, we support DR scenarios with the Libelle BusinessShadow portfolio that actually work in an emergency. Switching over individual databases, complex application environments or entire data centres at the push of a button. Best practices from 25 years with hundreds of enterprise customers.
However, not every company can or wants to set up and manage a second data centre at a different location. This is because both infrastructurally (buildings, security systems, networks, etc.) and system-technically (server hardware, operating systems, software, etc.) these systems must of course also be fully maintained. This in turn means internal and/or external costs and efforts for environments that are mostly not in active use, but must be fully available for emergencies. In addition, their size is often geared to production operation. Or in other words: standing, dead capital on a significant scale, which even causes expenses.
In these cases, Libelle, together with various cloud providers, offers the possibility to solve such situations comparatively simply and efficiently - with DR as a Service - DRaaS.
You operate your production environments locally in your company's own data centre as before. Of course, you can continue to use all existing availability technologies and retain existing procedures. Or you can re-evaluate the existing procedures at any time.
In addition, together with Libelle and one of the cloud providers, you build DR systems that are shrunk to minimal requirements during normal operation. These simply have to be performant enough to accommodate the data flow from the Libelle BusinessShadow. Means: Comparatively very low costs for this emergency system in normal operation.
If an emergency occurs, the performance parameters are increased, similar to virtualised environments. More CPU, more RAM, more network throughput. Here you decide individually which performance should be available in emergency mode. For this emergency operation, the corresponding costs are incurred, but they can be calculated very well in advance and in most cases are still far below the costs of a self-operated system.
The keyword here is right-sizing. For each operating mode, you choose the right system size and only pay for it.
There are many reasons to think about cloud systems. Some of them can also be found in other application scenarios.
However, there are some very interesting approaches especially for the DR concept: Companies do not have to invest in the corresponding infrastructure, plus the right-sizing described above. But other aspects should also be taken into account when making a decision:
Admittedly, such a DRaaS scenario is not worthwhile for every company. We are happy to support you with our best practices in the concrete analysis, calculation and evaluation and, if necessary, also implementation of corresponding scenarios. We work with the big hypescalers such as AWS and Microsoft Azure as well as with regionally oriented providers. The question of your specific requirements always arises in the selection process, which we will be happy to evaluate together with you.
We also see that some of our customers are using DRaaS as an entry scenario into regular cloud operations. The advantage of this, besides the reasons mentioned above, is also the build-up of experience. In addition, there are many other possible triggers: the existing DR concepts and infrastructures, which have been "good enough" for a long time, are reaching their limits in terms of requirements/performance; hardware is expiring; existing service provider contracts are expiring; data centres are to be scaled back; ...
If the topic of the cloud is being considered anyway, but a "big bang" conversion is to be avoided, then outsourcing the operation of the DR systems is, in our view, an absolutely suitable first step to maintain IT security, to fall back on best practices, to gain own experience and also to realise commercial effects.