Dataware for data-driven transformation

Supercharge Your Cloud ROI with MapR

Contributed by

14 min read

The MapR Data Platform adds excellent value to your cloud strategy, architecture, and roadmap. In particular, the MapR Data Platform empowers organizations to:

  • Port their data and apps across clouds and other deployments, including on-premises and edge environments;
  • Develop applications atop one integrated set of services, using open APIs;
  • Achieve inter-cloud capabilities through, for example, cloud bursting;
  • Supercharge your cloud ROI by moving existing applications more quickly to the public cloud and reducing total costs.

Each of the above is explained in greater detail in our cloud whitepaper. In this blog, we focus on one aspect – supercharging your cloud ROI with MapR.

Organizations are attracted to using a public cloud provider's platform to reduce or avoid the traditional costs of buying, housing, operating, and managing on-premises server and storage assets. It also promises flexibility to quickly grow or shrink computing resources as business demand fluctuates, without incurring greater fixed capital costs. These value propositions are attractive, whether the organization has a "cloud only" strategy or operates a hybrid environment. And, as we will show, MapR can provide significant additional value to either strategy.

How Does MapR Supercharge Cloud ROI?

Building on the extensive infrastructure and services provided by the major cloud providers, MapR can deliver high value capabilities that increase the business benefits to your organization while at the same time delivering incremental "better together" reductions in total costs.

Migrating Applications to the Cloud

For customers migrating on-premises applications to the cloud, the MapR Data Platform can make redeploying much faster and recoding cheaper. Moving applications to the cloud is generally a very costly proposition, both in labor months and elapsed time needed to rewrite and test the applications on the cloud APIs. Since MapR includes both NFS and POSIX APIs, many or most of the targeted applications can be rapidly migrated to the cloud at a much lower cost. Organizations that are planning to move applications to the cloud should answer the following questions:

  • How many applications do we plan to move to the cloud?
  • What is the total cost and the elapsed time required to rewrite these applications to work and verify they work correctly in the cloud?
  • What would our cost savings be if we could rapidly migrate most or 75% of these applications to the cloud using MapR with NFS or POSIX?
  • What is the business value of both the time savings in redeployment of the applications and the ability to redeploy scarce IT resources to other project priorities?

The answers to these questions alone may easily justify making MapR a central component of your cloud strategy.

Multi-Cloud for Maximum Agility and Leverage

Another huge benefit of using these MapR API interfaces is that they also enable you to quickly and cost-effectively move applications from one public cloud provider to another. The MapR Data Platform is cloud-agnostic and can add "better together" value to any major public cloud deployment, including multi-cloud scenarios.

One clear advantage MapR provides is lowering switching costs between cloud vendors by greatly reducing the need to rewrite applications to work on a cloud environment. While your organization's current strategy may be with a single cloud provider, we believe it's very important to consider the longer-term implications of single cloud lock-in and the high likelihood that the variety, quality, and net pricing of available cloud services will evolve over time across multiple providers. For that reason, the use of MapR in your cloud architecture can lower switching costs and ensure that your organization maintains maximum agility and purchasing power over time.

Increased Development and Data Science Productivity

By using the MapR Data Platform, customers realize substantial improvements in the productivity of their developers, data scientists, and other business users due to the faster, less labor-intensive cycle times associated with building, enhancing, and evolving their growing number of cloud-based use cases. As an example, one of our customers reports a 3X improvement in productivity with expectations for even greater improvements in the future. Other "cloud-only" customers report similar benefits linked to the use of their MapR-enabled cloud architecture. While the value of these improvements can be viewed or quantified as a cost savings, we generally believe that the value in terms of agility, time-to-market, and velocity of new and enhanced applications is substantially greater.

Storage Cost and Performance Optimization

MapR XD is a scalable, high performance distributed file and object store that supports object tiering, compression, and global namespace. These capabilities allow our customers to optimize the use of both low-cost cloud storage and the superior, high performance file and object store essential to meet business requirements and SLAs, such as real-time analytics. MapR does this by supporting both a hot data tier and cold object data storage tiers and, with global namespace, manages and utilizes all data holistically without the need for physical data management or movement.

In addition, MapR compression decreases the size of data storage used in the cloud, which is a primary driver of the billings generated by public cloud providers. With object tiering and

compression, MapR can reduce your total billings for storage services while, again, also providing a high-performance data tier for greater business value.

The illustration below shows the total billed terabytes with and without MapR. It is based on a realistic example of having a 15% MapR hot data tier and a standard 50% rate of MapR compression. While the relative sizes of the hot and object tiers are key variables in determining your potential savings, we generally expect the size of the higher cost, high-performance hot data tier to decline over time as a percent of total volume.

MapR Automated Storage Tiering (MAST) handles all of the data movement between the tiers. It ingests file data into the "hot" (performance) tier and then, depending on customer-defined rules and schedules, the file data is offloaded to the "warm" (capacity) or "cold" (archive) tier. Data in the MapR performance tier is highly available and resilient for faster, reliable access. MapR Erasure Coding for the warm tier offers cost-optimized capacity with data protection, with the ability to handle up to 3 failures.

Data tiering is an intelligent solution for cost-effectively managing ever-increasing data growth. MAST eliminates the need to move data manually while intelligently scaling and translating it for the cloud. Users and applications accessing files do not have to take any special action to take advantage of object tiering. Erasure coding for the capacity tier makes a compelling choice, due to its resilience to failures and storage efficiency. The way that the MapR Data Platform separates the decisions about detailed storage for data from the use of the data by applications is an example of automating policy that can drive down costs without sacrificing performance. With MapR, capacity and performance can be achieved simultaneously and cost effectively!

High Value, Lower Cost Database and Streaming Services

The MapR Data Platform includes both MapR Database and MapR Event Store for Apache Kafka. In our use case analysis, both of these services can be used as alternatives to similar services offered by major public cloud providers to hold down the costs but still deliver superior results.

MapR Database is a high-performance NoSQL database management system built into the MapR Data Platform. It is a highly scalable, multi-model database that brings together operations, analytics, and real-time streaming and database workloads to enable a broader set of next-generation data-intensive applications.

Based on our analysis of pricing data and use case scenarios, we found that using MapR Database can cost up to two-thirds less than the cost of related public cloud offerings, such as DynamoDB and Cosmos DB. Our analysis included the GB per month costs plus the read and write capacity unit charges. MapR can also coexist with these public cloud database services to lower or partially offload the levels of relevant storage volumes that drive billings for services such as Redshift.

MapR Event Store for Apache Kafka is a global publish-subscribe event streaming system for big data. It connects data producers and consumers worldwide in real time with unlimited scale. Based on our analysis of pricing data and use case scenarios, we found that the incremental cost of MapR Event Store for Apache Kafka can be up to 50% less than the cost of related public cloud offerings, such as Kinesis or other event hubs. Our analysis factored the direct service costs (e.g., rate per hour) plus any applicable charges, such as throughput units.

Increasing Value into the Future

As companies gain more experience using the cloud for their business, their data and application footprint typically expands rapidly, and the number of their use cases grows and becomes more complex. New use cases for AI and ML are found, and the amount of data grows exponentially. As one of our more sophisticated customers with data streaming and real-time business requirements observed recently, "We're not using the cloud and MapR just to park data."

Many of our customers are moving well beyond the basics of IaaS and are finding increasing value in MapR's data management and developer capabilities and ability to reign in direct cloud costs. They have seen that MapR's value as a component of their cloud architecture has risen as the size, diversity, and complexity of their cloud environment has grown.

By the Numbers

We analyzed total annual billings across multiple customer use cases and across multiple public cloud providers and have consistently observed cost savings when using the MapR Data Platform with a public cloud offering. The use case types range from relatively small and simple analytics cases to much larger and more complex cases, requiring database and streaming services. Our modeling included various scenario and sensitivity analyses. While there are many customer-specific variables and other variations across public cloud providers, our analysis consistently found that using MapR can:

  • Lower total annual cloud billings by 15% to 31%
  • Lower the cumulative 3-year cloud costs for an average-sized analytics use case by about 23% or an estimated $550K
  • Lower the cumulative 3-year cloud costs for an average-sized complex use case by about 23% or an estimated $2.7M

MapR can provide valuable services, such as database and streaming, at a lower incremental cost than the pricing for the same services offered by the major public cloud providers. These services apply to relatively advanced use cases that can take advantage of the MapR Data Platform. As one customer recently said, "With MapR, everything is in one place. Why would you choose to take on the time and cost of assembling isolated services?"

MapR can also reduce the total storage costs across many use cases, including both relatively simple analytics cases to far more complex use cases requiring, for example, database and streaming. This is primarily due to object tiering and compression, as I will describe below.

In examining a large-sized analytics use case, we found it yielded a $286K or 25% reduction in total annual billings plus a one-time $45K savings in development costs. This example happens to be based on AWS including S3, and the appropriate MapR XD subscription including AWS EBS and node charges. Other key variables included 2,000 usable terabytes of file storage, two AWS data center locations, a MapR hot data tier sized at 10% of total volume, and a 50% MapR compression rate. The cost savings we observed on key AWS billing metrics are also seen on other cloud providers such as Azure. The value of improved development/data scientist productivity, including reduced "build costs" and faster time-to-market, are addressed later in this blog.

Next Steps

Throughout this blog, I've shown many ways that MapR has added value to our customers' cloud strategies, no matter if it is "cloud-only," hybrid, or multi-cloud. I have shown both direct and indirect cost savings and how that value grows as the size and complexity of your cloud footprint grows. But I also realize that each company has unique challenges and strategic directions.

Hopefully, I've proven that you owe it to yourself and your company to see whether MapR can deliver faster, higher-quality results from your cloud strategy while at the same time controlling and even lowering costs. I urge you to contact your MapR representative today, so they can work with you to show you the value you will achieve by making the MapR Data Platform an integral part of your cloud strategy.

To learn more, read our Cloud Whitepaper: Realizing the Full Potential of Your Cloud Investment.


This blog post was published February 28, 2019.
Categories

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