Virtana announced the grant of US Patent No. 12,348,389 B2 for full-stack optimization of AI-centric cloud applications.
The patented method builds a live application model, links performance targets to the required resources (including GPU-accelerated infrastructure), detects emerging issues, and autotunes the stack to maintain SLOs while computing real-time cost per operation, enabling right-sizing decisions that reduce spend and improve ROI for both AI and traditional workloads. This follows the recent announcement of another patent awarded for creating an orchestration system that dynamically manages AI analytics workloads.
The new patent describes a practical control loop for continuously monitoring cloud applications, building a full-stack application model, mapping performance needs to underlying resources, detecting issues as they emerge, and automatically adjusting the stack to meet service-level objectives (SLOs) while computing per-operation cost in real time.
Why this patent matters for customers:
- SLO adherence with fewer firefights: Proactive detection and targeted adjustments reduce variance and long-tail latency.
- Right-sizing without guesswork: Matches standard and AI-workload needs to resources in real time, eliminating over-provisioning and waste.
- Cost-to-serve transparency: Real-time per-operation cost analysis enables accurate showback/chargeback and better architectural choices across traditional and modern AI environments.
- Faster triage and safer changes: A model-driven approach isolates the tuning layer, reducing the risks posed by broad, manual interventions.
- AI, hybrid, and multi-cloud fit: Works across on-premises and cloud environments for consistent performance and cost governance.
This patent extends Virtana's optimization loop to AI workloads by linking model/service performance to the exact infrastructure required (compute, storage, network, GPUs) and then tuning the right layer at the right time. The same closed-loop records cost as resources are consumed, giving platform teams clear cost-to-serve for AI operations alongside traditional services, so teams can:
- Align LLM latency targets (p95/p99) and vector DB throughput to GPU/CPU, memory, and storage choices.
- Tune RAG and inference pipelines without broad, risky changes—adjusting the narrow layer that drives the bottleneck.
- Identify per-operation AI cost (e.g., per prompt, per embedding, per inference) for comparing models, routes, and deployment options on both performance and spend.
"With a full-stack model and SLO-driven control loop, our platform can adjust the right layer at the right time and capture the cost impact as it happens," said Paul Appleby, CEO and President of Virtana. "That combination drives both reliability and financial accountability at scale. For AI, that means predictable model behavior, lower MTTR, and better ROI on GPU fleets and supporting infrastructure."
The invention formalizes a full-stack optimization method that:
- Builds a live application model spanning services, dependencies, and infrastructure.
- Maps performance requirements to resources (compute, storage, network, GPUs, etc.) using that model.
- Detects performance problems early via continuous telemetry and policy thresholds.
- Dynamically adjusts specific layers (application, platform, or infrastructure) to maintain SLOs.
- Calculates real-time aggregate cost for a specified operation as resources are consumed.
"Cloud applications are dynamic systems, and AI makes them even more dynamic. This patent codifies a disciplined way for organizations to be performant and cost-aware by connecting what the app or AI workload needs to what the infrastructure delivers in real time," said Amitkumar Rathi, SVP of Product and Engineering at Virtana. "Customers get steadier SLOs and clearer economics without constant manual tuning."
The patented capability underpins optimization workflows in the Virtana Platform and is available as part of ongoing product updates. This patent complements another patent Virtana was recently awarded for priority-aware scheduling and backpressure mechanism that dynamically reorders and resubmits analytic tasks based on real-time resource availability [Virtana Patents Orchestration System for Dynamically Managing AI Analytics Workloads].
The Latest
I've spent a lot of time in the channel, and one thing I keep coming back to is this: a partner program is only as good as what it looks like in the field. Many programs look great on paper, but when a partner is in front of a customer navigating a complex hybrid environment or trying to make the case for AI-powered observability, the gap between what a vendor promises and what it actually delivers becomes very clear, very fast ...
Enterprises today operate in a real-time environment where uninterrupted access to trusted data has become a baseline expectation for users, applications and automated systems. Traditional DataOps models, built on manual effort and human triage, cannot keep pace with this always active demand. AI agents are emerging as the operational backbone, ensuring consistent data availability, reinforcing trustworthiness and enabling a level of scale that manual processes cannot achieve ...
For decades, trust in the digital workplace rested on familiar signals. We trusted faces on video calls, voices on the phone, and emails that appeared to come from people we knew. These cues felt human and intuitive. They anchored how decisions were made, approvals were granted, and access was authorized. AI-powered deepfakes have quietly broken that model ...
Cloud migration was supposed to be a one-way door. For most enterprises, it turns out it isn't. Cloud data repatriation is a real and growing trend. A new survey ... finds that 89% of organizations plan to expand their on-premises infrastructure footprint over the next two years — and 75% have already moved at least some workloads back from public cloud in the past 24 months. The findings point to a broad rethinking of where data belongs ...
Over the past few years, large language models (LLMs) have revolutionized the software industry. Given their ability to excel at multi-step reasoning, LLMs have helped enterprises streamline workflows and adapt to the unknown. However, employing such models comes with sky-high costs, latency issues, and limited flexibility. In the realm of IT operations, it is generally wiser to employ smaller, domain-specific models instead ...
For years, DevOps teams operated under a simple assumption: collect enough telemetry, and you can find and fix any problem. That assumption is breaking down. Modern enterprises now operate across microservices, hybrid cloud environments, APIs, Kubernetes, and highly automated delivery pipelines. Releases happen continuously, dependencies shift constantly, and failures spread faster than teams can diagnose them ...
New Relic surveyed IT and engineering leaders from the media and entertainment (M&E) sector to understand what's working — and where challenges persist with their observability practices. The findings reveal how M&E organizations are navigating rising platform complexity, audience expectations, and AI-driven change. Below are five takeaways that stand out ...
Let me start with something I've seen play out more times than I can count. A team hits a wall with the cloud. Costs creep up, then spike. Performance starts to feel inconsistent. Someone in finance asks a simple question like "why did this double?" and nobody has a clean answer ... Maybe this isn't the right place for everything. That realization feels like a breakthrough, like you've identified the problem. In reality, you've just identified the starting line ...
In MEAN TIME TO INSIGHT Episode 24, Shamus McGillicuddy, VP of Research, Network Infrastructure and Operations, at EMA discusses network observability tool sprawl ...
In cloud-native systems, scaling is often as simple as moving a slider. For on-premise databases, the stakes are different. Over-provisioning hardware is expensive. Under-provisioning leads to performance bottlenecks that are difficult to fix once the equipment is in the rack ...