The Missing Ingredient in Your Multi-Cloud Strategy

Contributed by

8 min read

Savvy organizations around the world are embarking on multi-cloud transformations. In some cases, this is motivated by avoiding vendor lock-in, but often it's driven by the explosion and variety of data-intensive apps that are part of an organization's digital transformation. Customers want to go with cloud providers that offer the best services for their application-specific needs. For example, Apple made headlines two years ago by diversifying its infrastructure as a service (IaaS) footprint with Google along with Amazon Web Services. And last week, longtime Amazon Web Services customer Netflix announced that it has transitioned certain workloads, mostly focused around artificial intelligence, from AWS to Google.

The Missing Ingredient

If these bellwethers are any indication, the mass movement to a multi-cloud world is just a matter of time. While moving to a single cloud is painful enough for many organizations, it takes even more careful planning and execution to prevent data silos across different clouds. It's about more than just maintaining a pool of resources that you can use whenever you want. The component approach to cloud, where each provider has a vast catalog of independent services, places an enormous burden on customers to figure out how to connect these services into a cohesive application architecture. Furthermore, cloud vendors don't make it easy to move workloads between different clouds or between their cloud and an on-premises data center.

Some companies are finding a way to move their applications around with technologies like containers and Kubernetes. But there are challenges with containers - data being chief among them. Most containerized applications today are stateless: if destroyed and redeployed, data may be lost, but no services are disrupted. However, focusing on stateless applications is problematic because most applications require some type of state in order to know what's been happening and respond to the next thing. Reliability is another issue. Container environments don't provide the reliable storage needed to support stateful applications.

Meanwhile, big data analytics and machine learning are becoming fast-growing areas in cloud computing. Business executives increasingly expect IT to use data to help them understand their customers' preferences and buying habits. But what happens when your applications cannot access all of the relevant data? If you don't have the data you need, you can't learn about how your business functions and make informed business decisions. Data issues are pervasive, regardless of the cloud platform. That's why it's so important to define a holistic data strategy supported by the right data architecture to ensure a successful multi-cloud strategy.

You need to think beyond point solutions in the cloud. For example, if you are using Amazon, and you put your data in S3, it's in S3. If you have it in Elastic Block Storage, it's in Elastic Block Storage. If you have it in Redshift, it's in that data warehouse. You start seeing data being placed in different locations and being modeled in different ways for different use cases. Furthermore, these data stores are completely disconnected, and it's very expensive to stitch them together.

What is needed is a converged data platform that lets you integrate data within and across various clouds, all while permitting workloads to easily move back and forth. Subsequently, you need to take a hard look at how you factor cost, interoperability, business continuity, future proofing, and global awareness into big data analytics in the cloud.

With MapR, you'll have a better way to execute multi-cloud applications as well as handle quick changes in your environment.

  • Future-proofing applications. Avoid rewriting your software every time your application has a new destination for a particular cloud provider, a specific data center, or a new platform inside your cloud provider. Write to MapR once. You are done. MapR also supports the containerization of existing and new applications by providing containers with persistent data access from anywhere. Of course, these capabilities help not only with multi-cloud transformations but also with the move to a single cloud as well.
  • Controlling costs. You can decrease your storage costs by up to 50% through inline compression with MapR, along with our ability to transparently tier data to low-cost cloud object storage. Automatically moving your data from hot to cold storage (and back again when needed) helps to reduce costs while getting the performance you need. Additionally, data is encrypted as it travels over the network and at rest.
  • Avoid Lock-In. When you look at all of the service offerings from all of the cloud providers, most APIs are non-standard (with some exceptions). A common fear when dealing with a cloud provider's API is vendor lock-in. MapR prides ourselves on a completely pen API architecture that ensures interoperability and avoids vendor lock-in, providing you with options in the future.
  • Maximizing business continuity. MapR offers a replication service, which synchronizes database changes and streaming messages between data centers and cloud providers. This saves you from having to write a bunch of extra code, and more importantly, you can load balance across different cloud providers for solid disaster recovery. Fr files, you can also take point-in-time snapshots, which you can mirror to another cloud or on-premises data center. These snapshots can also be used to run new ML models or applications against old, historical data.
  • Global awareness. When you have multiple data centers, whether on-premises or in the clouds, you need to have a single view of all your data and be able to access it whenever you need to, wherever it might be. You have this ability with a single global namespace from MapR.

The MapR Data Platform opens up the door to all enterprise applications. It provides a place to store your data, a place to process it, and the ability to focus on real-time activity. It can run on-premises or in the cloud, simplifying how you move from one to the other. We have a large number of customers with dozens of petabytes of data, and they only have one primary administrator and one backup person to administer their environment. The robustness of the MapR Platform not only lowers administration cost but also protects your business from disruptions, outages, and integration issues in a complex multi-cloud or hybrid cloud environment.

Related Links:

This blog post was published May 03, 2018.

50,000+ of the smartest have already joined!

Stay ahead of the bleeding edge...get the best of Big Data in your inbox.

Get our latest posts in your inbox

Subscribe Now