Major defense contractor relies on AIStor for portable data lakehouse and S3 hybrid cloud available under any conditions

A major US-based defense contractor needed to create an internal solution that allowed for the secure sharing of critical information with other allies and sensitive actors – a portable data lakehouse that could be scaled up and down on demand and deployed anywhere in the world, across a wide variety of deployment options and in challenging network environments. 

Environment: Mission-critical applications and diverse requirements 

Our customer required a highly-available, cloud agnostic, open standards-based object store which would meet these challenging and diverse requirements:

• Provide a fully S3 compatible, REST-based API
• Operate in DDIL (disrupted, disconnected, intermittent and low-bandwidth) and air-gapped environments
• Support a wide variety of cloud-agnostic deployment options, including Raspberry Pi
• Seamlessly replicate to a larger and more complex object store repository
• Utilize a standard technology stack across deployment architectures to minimize attack surface and risk.

Challenge: Building a highly secure, ultra-portable data lakehouse

To meet all of  these stringent requirements, our customer considered only three options – HDFS, CEPH and AIStor. HDFS was eliminated immediately since a shared file system was not compatible with the customer’s stringent network requirements, while CEPH had complexity and scaling challenges along with similar issues operating in the customer’s challenging network environment.

Solution: MinIO AI Store S3 Hybrid Cloud

AIStor fulfilled the above requirements, aligning directly and immediately with the established core tenets for our customers’ application: 

• Act as a design partner and trusted advisor in architecting the solution
• Lower maintenance and deployment costs based on open standards and open source software (typically 60-85% smaller infrastructure footprint). 
• 100% S3 compatibility
• Rapid deployment and recoverability
• Interoperability with other data layers; mesh and federate rather than replace or compete with data silos. 
• Support future expansion and performance requirements 

Meeting these requirements makes the customer’s programs and products more interoperable, which in turn benefits the organization and their customers. 

Customer workflow with AIStor within the theatre of operation: 

• Image acquisition: Images are acquired and stored on a mobile device
• AIStor Batch Replication: Artifacts are sent to local AIStor for field-level analytics
• AIStor promptObject: Artifact assessment is performed in the field and warfighter has low-latency access to information 
• AIStor Bucket Replication: All assets and metrics are sent to secure cloud for recording of events and evaluation (batch replication) over a 4G/5G link

Outcomes: Optionality at any scale at the edge

AIStor delivered optionality around scale, with the ability to start small and grow rapidly based on their requirements for sharing and collecting data at the edge. And with AIStor’s open standard approach, there is no need to interrupt or change core functions when growth is required. 

Rather than having multiple nodes and systems deployed, our customer now has the flexibility to deploy entirely at the edge and all the way back to a data center or private cloud environment without interference or extraneous developer time. AIStor provides full optionality across every access point, edge to cloud, while allowing the customer to reduce its overall footprint by nearly half. 

AIStor also helps our customer minimize risk and maximize applicability and reusability, reducing friction across disparate environments – a non-negotiable requirement when lives are literally at stake.