Cloud First | Cloud First Policies | Informatica US

Cloud First

Cloud First Policy

Congratulations! You work for a large organization that has announced a cloud-first policy: The new default assumption is that any new system will be stood up in the cloud.

The upside reasons include:

  • Flexibility: Stand up new systems in hours not quarters.
  • Scalability: Leverage cloud elastic scalability to manage peaks in demand.
  • Cost: move to a use-based payment model that is CAPEX rather than OPEX.

These are just some of the reasons we’re seeing a cloud-first shift this year among some very large Informatica customers.

But here is the inconvenient truth: For the vast majority of large organizations, moving from all-on-premise to all-cloud is a multi-year transition. We’re working with large companies, including a petrochemical giant, and mid-sized companies like Harvard Business Publishing. They all say the same thing: This it a journey that will take years.

Why does it take so long?

Most cloud systems going live today are net-new systems rather than replacements for older systems, particularly in the analytics space, where the vast majority of new systems are augmenting rather than replacing older systems.

Over time, organizations will build migration and replacement strategies, typically as part of a major modernization initiative. We certainly see companies doing that.

But, even for the most aggressive organizations, this is a long process that will take careful planning. Some of the issues include:

  • Standing up the new system.
  • Ensuring the security of the new system and data movements to and from the system.
  • Ensuring that the new system is loaded with data from all relevant sources.
  • Ensuring that the data fed into the system is of high quality, is timely.
  • Ensuring that the data is governed: data quality and business context are managed on an ongoing basis, the data has business context and meaning attached, there is data lineage to prove to management and auditors where the data came from.
  • Ensuring that there is two-way data synchronization between the on-premise systems and the new cloud systems.

All of this will take a carefully considered enterprise data architecture. The really interesting part is that this data management architecture will have to evolve as the “data center of gravity” moves to the cloud. This concept is based on the idea that it’s easier to do the analytics where the data resides. As you move more data to the cloud, you have to consider that your analytics should probably follow. To keep data movement at a minimum, architect with this center of gravity in mind.

The risks along the way

The number one risk that we hear about is security. The next most common issue is protecting privacy and sensitive data.

But don’t forget regulatory compliance. If you’re subject to compliance, you need be able to audit your data and prove to auditors how it was moved, where and how it was transformed, and who has touched it. This is difficult in an all-on-premise world. It gets much harder in a hybrid environment. Strong metadata management and data lineage will be critical for anywhere important data is located. A major financial institution just received multimillion-dollar fines because they did not have visibility into employee activity that proved to be fraudulent.

The other major danger is in creating data silos, islands of unconnected data. This makes data opaque to auditors and slows down your future analytics activities, where the generation of actionable insights is dependent on accessing data from many sources across the organization in a timely manner.

I’ve written many other posts on this subject.

What you can do to minimize the risks of hybrid data management

It will all come down to careful planning.

  • Carefully define the desired future state of your data management architecture.
  • Define the milestones along the way. As mentioned, your data management architecture will have to evolve as your data center of gravity moves to the cloud.
  • Design for data security and privacy up front. It’s much harder to retrofit security later, usually after a negative event.
  • Design for data auditability and governability up front.
  • Design your data architecture to speed up your data delivery rather than slow it down. The shift to cloud and hybrid will provide the opportunity to rethink your approach and to leverage the speed, flexibility and scalability benefits of cloud. You will also have the opportunity to leverage data self-service and automation to speed up access to trusted data for your organization while reducing the load on IT.




This all sounds relatively straightforward, but the truth is that it will take careful planning and architecture to pull it off successfully.  In the end, the movement to cloud will be a journey for everybody. For some very helpful research on this subject, check out the TDWI Report: BI, Analytics, and the Cloud. Strategies for Business Agility.