Hybrid clouds offer a number of advantages. But they are also a complex architecture. It requires constant monitoring, as well as solutions to intuitively and efficiently manage a whole range of operations: network performance, workload management, security, cost control...
Because of the covid-19 epidemic, many companies are considering migrating to the cloud. The spread of teleworking has revealed that companies' capacities are not really optimized.
The main obstacles to maintaining an "acceptable" level of performance: undersized VPNs and VDI servers. On social networking sites, several executives confessed that they were "having a hard time" connecting to their organization's IS. To avoid bottlenecks, CIOs have had to set up access schedules throughout the day, and even into the evening!
As a result, companies are beginning to reconsider their strategy: they are considering taking cloud integration a step further. In most cases, they have been content to rely on SaaS or public IaaS. However, as we emerge from this period of exceptional crisis, the priority will certainly be to implement a cloud-based Disaster Recovery Plan (DRP). This will be the subject of a future article.
A global view of hybrid clouds
The hybrid cloud appears to be one of the most appropriate solutions for limiting the impact of a serious crisis such as the coronavirus. Its main advantages are well known: flexibility and cost control, distribution of workloads according to their level of criticality, scalability...
On the other hand, the hybrid cloud is complex to manage. So there's no need to rush into it just because it's a good way to regain market share after months on the sidelines.
Any organization considering such a rollout should start putting a transition framework in place as soon as possible. As we've made clear several times on this Blog, the most important decision is which data and applications should be hosted on site (due to data sensitivity in particular), and what will be put in the cloud.
But what then of day-to-day management when an IT environment resembles a mille feuilles? Suppliers offer numerous options and tools, but not all of them are easy to master.
Above all, the absence of standardized APIs means that you can't have a global view of all your services. Many solutions display compatibility restrictions on the components they can manage, locking your management platform into a vendor or group of vendors.
Using standardized open source tools can be an effective starting point to guard against compatibility problems. Tools from the Cloud Native Computing Foundation(CNCF), such as Kubernetes and Prometheus, are good examples. As for HashiCorp's open source tools, such as Vault, Vagrant, Packer and Terraform, they support the management of Multicloud deployments and hybrid clouds.
Prioritize with hybrid clouds
But they are far from sufficient. As with any problem, the most difficult thing is to prioritize. In the case of the cloud, prioritization concerns management requirements. But not everything works perfectly... There are, and always will be, gaps. Whatever the provider!
When selecting a hybrid cloud management platform, it's important to know that nothing is certain until a process has been tested in this environment and against its business requirements. Companies therefore need to determine which functions are most important to them, based on their use cases.
But the human factor is also decisive. As well as getting staff on board, you need to ensure that you have the right skills in-house (particularly for configuration and access control). If this is not the case, training, additional qualifications or support from a specialist service provider may be required.
Ultimately, managing hybrid clouds requires a new team spirit. This new environment can solve many problems. But it can also create others if not properly implemented and managed.
It's vital to make sure you understand the scope of a hybrid infrastructure and how it will be used, so that it precisely meets your needs.
