MapR on Metal Cloud is 10 times faster than other Hadoop distributions, especially at large scale.
- Record-breaking data processing power
- Infinitely-scalable file system
- High availability
- One-click setup on high-performance bare metal environment
- Configurable distribution of services across all nodes in the cluster
MapR on bare metal cloud delivers consistent low latency, data protection and the highest uptime of any Hadoop distribution.
Bare Metal Performance
Running on physical servers instead of VMs makes MapR perform several times better than other Hadoop distributions in the cloud.
MapR on Bare Metal Cloud has NFS multi-node HA and provides out-of-the-box solutions for backup.
Multi-tenancy means that you can logically divide your MapR cluster in a manner that best suits your application's needs.
MapR employs strong security thanks to Kerberos-based authentication and encryption. MapR uses Access Control Lists and Expressions that can be easily configured and enabled.
MapR on bare metal has both proprietary file and database access, and standard file and database access, as well as a wide range of pluggable services.
The All-In-One Solution for Big Data
MapR supports a wide selection of Hadoop applications*. No matter what your use case is or how many resources it needs, you can rely on a single and easy to use solution from beginning to end.
MapR on Metal Cloud Advantage
High scalability, instant access and consistency issues are the traditional Hadoop concerns. MapR on metal cloud solves them by making use of the proprietary MapR-FS instead of HDFS and also by employing a no-NameNode architecture. Thanks to these two innovations, MapR on metal cloud ensures high availability across nodes and services for both file-based and NoSQL applications.
Highly Scalable File System
NFS - mounting over NFS gets rid of the limitations of HDFS's append-only model and addresses major limitations for applications such as HBase.
Compression - the MapR filesystem supports automatic compression and an unlimited number of files.
Data Control - MapR uses hardware more efficiently by controlling data placement in the cluster.
Dynamic Resource Allocation
Distributed HA - MapR on bare metal cloud employs the "no single point of failure" design principle by using a distributed HA model.
Snapshots - thanks to a redirect-on-write storage system model you can snapshot a petabyte in just a few seconds.
Mirroring - differential cluster-to-cluster mirroring is useful for disaster recovery, remote backup, and production vs. research cluster setups.
From Big Data to Big Opportunities
Real-Time Stream Processing
MapR provides a simplified publish-subscribe model for real-time stream computation using Storm or Spark Streaming. Data feeds can be written directly to the MapR platform, which allows for less overhead and a shorter app chain.
Predictive Analytics, Full Search & Discovery
Gather trend data, search Hadoop data directly or index standard files with no conversion or transformation needed. Content and results are highly available, automatically compressed, and can be protected using snapshots and mirroring.
Large-Scale Distributed Datasets
Low latency interactive query capability, hierarchical data structures and schema discovery make MapR on the Bigstep metal cloud perfect for working with large-scale distributed datasets. It supports NoSQL, Hadoop and traditional RDBMS.
Security & Risk Management
Analyze real-time data from network or other security devices, process application log data preemptively or use pattern and anomaly recognition capabilities to improve security and reduce unnecessary risk.
Managed Services on Metal Cloud
Bigstep’s Managed Services is aimed at alleviating the need for specialized big data personnel needed to manage an infrastructure. It is a very advanced support service that can handle interventions to a cluster that it manages on behalf of the customer and that also handles troubleshooting and escalation to the vendors of the underlying technologies for situations that demand it. The Big Data Managed Services connects customers to a team of certified experts in many of the technologies we support, especially in the the Hadoop-based technologies.
Infrastructure Level Troubleshooting
Bigstep proactively resolves all incidents related to hardware, network infrastructure, operating system or software used by Bigstep to support services provided to the customer.
Consistent End-to-End User Experience
Bigstep handles the end-to-end solution setup and configuration according to the changing needs of the customer and is working closely with vendor’s support organizations to troubleshoot advanced software issues.
Application Level Troubleshooting
Bigstep resolves all incidents related to the automatic provisioning mechanism used to deliver all "Application as a Service" included in our offer.
We are specialists in Spark-based stacks and we can build and maintain private data lakes, real-time analytics, data processing stacks, and data science laboratories.