If you’re like most businesses, you already have at least one workload running in the cloud. However, that doesn’t mean that cloud migration is right for every element of an enterprise’s IT infrastructure. While cloud environments are generally scalable, reliable, and highly available, those won’t be the only considerations driving your decision. At Riversand, we typically see companies choose hybrid cloud solutions as a strategic middle ground and as a stepping stone to understanding and learning about the best ways to use the cloud.
From our work across various industries and types of organizations and businesses, we are in a unique position to hear a lot of good questions and concerns pertaining to cloud migration. We wanted to share some of our key insights from companies still sitting on the cloud-fence — so we invited our VP of Global Marketing, Katie Fabiszak to sit down with Riversand’s Senior Director of Product Management, Nikhil Bhatia and discuss what customers were saying:
KF: What types of hybrid cloud database environments have you encountered?
NB: We have seen a multitude of hybrid cloud environments being used. The trend seems to be many environments moving to the cloud (our own master data management – MDM – application is completely cloud native), but most legacy systems, especially large ones such as ERP and PLM systems, are still on-premise for many companies. In our case, most of the data being managed is business critical and needs to be communicated to our customer’s own eCommerce channels (e.g. pricing, inventory and availability data) in a time sensitive manner. With master data being managed in and sourced from multiple sources for every one of our customers, we deal with hybrid cloud database environments on a daily basis.
KF: What types of applications are still best suited for on-premise database systems?
NB: While all applications can be adapted to be on a cloud-based database system, there are still some that operate best in an on-premise environment. For example, highly integrated, business critical applications such as ERP or finance applications are candidates for on-premise database systems. In addition, while the security and privacy of cloud-based database systems has improved significantly, compliance and regulatory rules may still force companies to host applications containing sensitive data in on-premise systems.
KF: How are enterprises typically dividing up cloud-based applications and data versus on-premise?
NB: Enterprises are typically moving their back-office systems, including those supporting HR legal and IT to the cloud. In addition, cloud-based SaaS solutions such as Salesforce and Hubspot for CRM, Confluence for Document Management, Slack, Jira etc. provide a way for small and big enterprises alike to move independent applications to the cloud. The highly connected, legacy systems such as ERP, PLM, MDM are still typically hosted on-premise. MDM solutions like ours at Riversand provide a completely cloud native SaaS MDM environment so the on-premise systems are seamlessly able to work with the cloud.
KF: Are there enterprises putting entire database systems in the cloud? What is driving such decisions?
NB: We are yet to come across enterprises who have put all of their database systems in the cloud. With its healthy mix of on-demand reliability, high availability, security, and reduced operations costs, hybrid cloud implementations can be attractive. Going hybrid can sometimes give you the best of both worlds.
There are multiple factors to consider for enterprises to decide on the mix of cloud vs. on-premise data:
- Cloud applications are built on the concept of pay-as-you-go hence have no capital expenditure (CapEx) requirement. But the operational expenditures (OpEx) can be daunting especially for large processing and storage needs.
- On premise infrastructure has more CapEx investment, but lower OpEx costs if an enterprise already has IT teams supporting the infrastructure.
- A detailed TCO (total cost of ownership) is required before making such a strategic decision.
- IT departments of enterprises need to maintain the on-premise infrastructure and have to be highly trained and comfortable with it.
- Cloud applications are typically built on the concept of self-service unless one is using infrastructure-as-a-service (IaaS) in the cloud.
- If the majority of the enterprise’s infrastructure is on-premise and a new application is to be integrated, it might be easier to host the new application on-premise as well.
- If a new application is intended for fairly independent usage, a cloud solution without any CapEx cost may be the most viable option.
Data and Processing requirements
- While cloud applications especially SaaS (software-as-a-service) are a compelling consideration, having large amounts of data transfers and processing on the cloud can be cost prohibitive. In these situations, on-premise solutions may make more sense for an enterprise.
- If the data in the application has to be accessed without latency across many geographies, this makes hosting on the cloud with features of geo-replication more desirable.
Security, Privacy & Compliance
- Some applications, due to the compliance and regulatory requirements, may not be put on the cloud.
- With the GDPR compliance framework rolling out, unless the cloud providers can ensure that the data being provided to them is being stored per a geographic location, enterprises will not have the trust to host applications with sensitive data in the cloud. Amazon Web Services (AWS) and Azure have already rolled out region-wide databases which can be selected per region.
KF: What are the main challenges associated with supporting a hybrid data environment? How can these challenges be addressed?
NB: The main challenges with supporting a hybrid data environment are:
First time data migration
- Migrating large amounts of data to the cloud from an on-premise system without any data loss or hampering of ongoing business is a common challenge. This can be mitigated with detailed planning and utilizing compare reports for data which most cloud services will offer. If the data is being moved into a SaaS solution with data manipulation requirements (which is typically the case for our MDM application) this gets even more challenging. Having simplified connector and ETL capabilities is critical for mitigation.
Integrating the hybrid cloud
- Setting up the right integration within the hybrid cloud environment is probably the most challenging piece. The ability to setup the right networking and connectivity in addition to understanding the different natures of stream and batch integrations followed by different applications and having the right ETL capabilities is critical. A new concept of EtLT (extract, small transform, load and big transform) is gaining traction to solve these issues.
- Budgeting for different CapEx and OpEx costs is one of the challenges for enterprises wanting to have a hybrid cloud solution. As long as enterprises and accounting divisions understand those nuances this challenge can be mitigated.
- Geographical and security compliance is something that has to be solved for the cloud hosted applications to be variable for hybrid cloud solutions.
KF: Thank you, Nikhil, for some great insights that will help folks better understand the nuances of hybrid cloud environments!