MulticoreWare deploys HuBe.ai application on OCI, saves 75% in TCO annually

Authored by Kellsey Ruppel, principal product marketing director at Oracle, and Joe Winston, AI Cloud sales specialist at Oracle.

The authors would like to thank Shiva Narayanan, VP of media and entertainment at MulticoreWare for his contributions.

An image of someone typing on a keyboard with a Robot hand coming out of the computer screen.
Figure 1: MulticoreWare offers AI applications and services.

 

The field of artificial intelligence (AI) has made some very important strides over the past year. From breakthroughs in natural language processing (NLP) and computer vision to improved adoption of AI by businesses, we’ve seen several milestones. It's no doubt that AI can continue its explosive growth, and companies like MulticoreWare are offering solutions that help customers reach their full AI potential.  

MulticoreWare is a global technology company headquartered in San Jose, CA, offering software products, AI solutions, and engineering services. They have offices in India, China, and the US and work with clients around the world and across industries, transforming the way technology is used and experienced every day.

With years of research and development activities behind them, MulticoreWare has developed and deployed state of the art AI applications and services that can be customized to various market verticals.

Goals for cloud migration

Human interactions and collaborations have gone virtual, 70-80% of which are video based. You don’t have the liberty of “seeing is believing,” so you don’t have any observational insights. Research says that 93% of all communication is nonverbal, including body language, tonality, and attention span.

In this virtual world, what if you had a way to get visual behavioral insights into a candidate? It could help you make informed decisions. If you had a way to gauge the interest of a prospect you are selling to, your return on investment would be better if you could pitch right-pivot. If you had a way of knowing the likes and dislikes of your audience, wouldn’t you target them better?

To address these areas, MulticoreWare has developed HuBe.ai an API and software-as-a-service (SaaS) platform powered by state-of-the-art deep learning technologies that helps identify behavioral cues. Being the first human behavior intelligence (HBI) platform on the market, their goal is to accelerate the application development for companies that develop applications to track human behavior and augment existing solutions with similar functionalities. Some of the multimodal features include body language identification, human activity detection, and focus engine, which includes distraction and drowsiness detection.

“We believe that HuBe.ai will help enterprises understand human behavior and get insights that can assist them in making good humancentric decisions,” said Shivakumar Narayanan, assistant vice president of marketing and products and MulticoreWare.

To achieve that goal, MulticoreWare turned to Oracle Cloud Infrastructure (OCI). HuBe.ai is running on OCI and had the following goals for their cloud migration:

• Improve the performance of the application to provide better productivity and output to their customers.
• Migrate the entire product setup from their existing cloud provider to OCI without impacting the product performance and service.
• Reduce the overall product and maintenance service cost by adapting OCI cloud services.

Suite of Oracle products used

OCI includes all the services needed to migrate, build, and run IT in the cloud, from existing enterprise workloads to new cloud native applications and data platforms. MulticoreWare used the following OCI services and technologies to run their HuBe.ai application:

• Virtual cloud network (VCN): VCNs are customizable and private cloud networks on OCI. The VCN provides complete control over cloud networking environment like a traditional data center network.
• OCI Object Storage: OCI Object Storage enables you to securely store any type of data in its native format. OCI Object Storage is ideal for building modern applications that require scale and flexibility, as it can be used to consolidate multiple data sources for analytics, backup, or archive purposes.
• OCI File Storage: File Storage is a fully managed elastic file system built for the cloud that enables customers to migrate their enterprise workloads to the cloud. Every file system scales automatically to accommodate the growth of up to 8 exabytes.
• OCI Boot Volumes: When you launch a virtual machine (VM) or bare metal instance based on a platform image or custom image, a new boot volume for the instance is created in the same compartment. That boot volume is associated with that instance until you terminate the instance. When you terminate the instance, you can preserve the boot volume and its data. This feature gives you more control and management options for your compute instance boot volumes, and enables instance scaling, and troubleshooting and repair.
• OCI Vault: OCI Vault is a management service that stores and manages master encryption keys and secrets for secure access to resources.
• Network load balancer: The OCI Load Balancing service enables you to distribute web requests across an array of servers and automatically route traffic across availability domains resulting in high availability and fault tolerance for applications or data sources.
• OCI Networking: OCI networking and connectivity products and services enable customers to manage and scale their networks. You can connect securely to a customizable, isolated VCN, and take advantage of inexpensive data egress charges.
• OCI Kubernetes Engine (OKE): OCI Kubernetes Engine (OKE) is an Oracle-managed container orchestration service that can reduce the time and cost to build modern cloud native applications. Unlike many other cloud infrastructure vendors, OCI provides OKE as a free service that runs on higher performance, lower-cost Compute shapes.
• OCI Identity and Access Management (IAM): OCI IAM provides an innovative, fully integrated service that delivers all the core identity and access management capabilities through a multitenant cloud platform.
• OCI Container Registry: OCI Container Registry is an open standards-based, Oracle-managed Docker registry service for securely storing and sharing container images. Engineers can easily push and pull Docker images with the familiar Docker command line interface (CLI) and API. To support container life cycles, Registry works with OKE, IAM, Visual Builder Studio, and third-party developer and DevOps tools.

MulticoreWare took the following steps to implement the HuBe.ai application:

1. Creation of OKE cluster was required for hosting applications in a virtual cloud environment.
2. Extensive usage of cluster autoscaler was required for worker node pool autoscaling based on the application workloads.
3. Configuration of kubernetes manifests files involving Container Registry images required for application deployment.
4. Integration of application base models in Kubernetes deployments using Object Storage.
5. Implementation of NGINX ingress controller for exposing applications service to the customer.
6. Integrating network file system (NFS) client provisioner with the file storage system was required for file management of the application.

Migration path

Provided at an economical cost, OCI’s exclusive features laid the foundation for the migration of HuBe.ai. They categorized the migration of the product in the following different stages:

• Setting up of cloud infrastructure using Terraform module configurations, including networking, IAM, OKE, OKE node pools, and cluster autoscaler.
• Development of helm charts for application setup, including Kong, Restapi, Keydb, Janus, Starch, user authentication, and base models.
• Configuring and setting up of NGINX Ingress Controller using the classic load balancer.

An image of MulticoreWare’s reference architecture diagram.
Figure 2: The architecture of MulticoreWare’s deployment.

 

Before migrating from their previous platform to OCI, MulticoreWare was charged for various inter- and intralevel data transfer, inbound and outbound). With OCI, they’ve been using this service with sufficient service limits. Before, MulticoreWare’s deployment of Kubernetes applications incurred extra costs for pulling the images from the container registry. OCI’s Container Registry has helped in reducing the cost for this service.

Before migrating the entire HuBe.ai application to OCI, MulticoreWare was using Compute instances with vCPU’s configuration provided by their previous cloud provider. OCI OCPUs helped them in the selection of the appropriate shapes with the wanted tech specifications, including memory and networking bandwidth.

The configuration of preemptible instances in the OKE node pool makes the offering available better and cheaper. Adaption of the preemptible instances with the appropriate selection of Compute shapes using OCPUs makes the product grow and expand extensively.

MulticoreWare migrated the project through a proof of concept (POC), which took approximately 12 weeks for complete deployment of the application. It helped them understand the core features and functionalities of the OCI services.

The major key efficiencies in migrating to OCI included significant savings in the operational cost, improved performance, and Oracle Managed Services helped MulticoreWare reduce the overhead times and ensure that applications are always supported by the most up-to-date infrastructure.

Results

Since running their Hu.Be.ai application on OCI, the MulticoreWare team is working more effectively and driving faster results. It also helps them make smarter decisions about customer journeys and personalizing the same. MulticoreWare was able to create another test environment for their team to innovate, test, and implement, which not only streamlined the process, but also subsequently improved productivity and saved time.

OCI provides smarter integrations and greater access to business-critical functionality, which helped MulticoreWare to automate the application quickly, improve application performance, and accelerate getting the application to market without any downtime. OCI Security Guard automatically detects any misconfiguration errors, which alert administrators and remediates the issues with no extra cost, which was a limitation before the migration to OCI.

OCI also simplifies the access policy by enforcing least privilege and zero-trust principles and makes it easy to understand identity and access management (IAM) policies and denying all access by default. OCI security zones help secure the sensitive data storage and database resources by making security mandatory and always on.

With OCI, MulticoreWare has realized cost savings of around 60% when compared to the prior cloud provider since their migration. They’re anticipating around 75% of cost savings with more added functionalities to the instance configurations, such as preemptible instances configuration for OKE node pools.

Next steps

In the future, MulticoreWare is exploring and planning to deploy OCI Open Search service required for database storage, and OCI Search for Observability and Management and Application Performance Monitoring. They’re looking to deploy OCI Functions to host their static applications, which helps them maintain and showcase the application without any downtime.

The OCI continuous integration and deployment (CI/CD) platform can help MulticoreWare to easily build, test, and deploy software and applications on OCI. Finally, the implementation of Cloud Guard can add another level of security to the infrastructure.

For more information on MulticoreWare’s HuBe.ai application and Oracle Cloud Infrastructure, see the following resources:

• HuBe.ai overview
• Explore other technical case studies from OCI
• Try Oracle Cloud for yourself