Pinakastra - Powering Sovereign Cloud Infrastructure Globally
Pinakastra Computing provides an integrated private cloud & virtualization platform Pinaka ZTi to simplify IT infrastructure management, improve efficiency, and accelerate digital transformation.
The incumbent SDDC contract is broken. Enterprises need a sovereign alternative — now.
At Pinakastra Computing, we tackle the critical challenges faced by traditional private cloud and hyperconverged infrastructure providers like VMware and Nutanix, such as complex deployment, high costs, limited scalability, security issues, slower innovation, and vendor lock-in
Cost shock
Per-core pricing has tripled across the industry. A 200-core cluster that cost ~$50K/year now costs ~$150K+. Enterprises paying >$500K/year on incumbent platforms now project $735K+ at renewal.
Lock-in by design
Bundled SKUs, mandatory add-ons, no perpetual escape. Migration off becomes harder every quarter. Average estimated time: 62 minutes per VM to migrate.
Innovation stalled
Roadmap uncertainty. Talented teams leaving. 64% of organizations will need more than one full-time engineer to achieve their new VM strategy.
Industry-Wide Pricing Surge
Per-core pricing has tripled across the industry. A 200-core cluster that cost ~$50K/year now costs ~$150K+. Enterprises paying >$500K/year on incumbent platforms now project $735K+ at renewal.
One platform. Four planes. Zero proprietary lock-in.
Pinaka ZTi is an enterprise-grade, validated, and secure Software-Defined Data Center (SDDC) platform that enables organizations to build and manage on-premises private cloud infrastructure for hosting virtual machine (VM) and container-based workloads with integrated compute, storage, networking, and virtualization orchestration.It allows organizations to build flexible, highly automated private clouds.
Unified SDDC Architecture
Compute
VMs and containers at native speed. KVM/QEMU hypervisor with libvirt orchestration via OpenStack Nova. GPU passthrough, vGPU partitioning and RDMA for AI/ML & HPC.
Storage
Software-defined block, object & file storage powered by Ceph. BlueStore + RBD backend. Self-healing, scales from terabytes to petabytes.
Networking
Full SDN with VXLAN/VLAN overlays, distributed virtual routing, load balancing and tenant isolation via Neutron + OVN. No 3rd-party NSX licenses required.
GPU
Flexible GPU virtualization for AI/ML and HPC workloads. Full PCIe passthrough, NVIDIA vGPU with SR-IOV, Multi-Instance GPU partitioning, and time-slicing for maximum utilization.
Six reasons enterprises pick Pinaka ZTi for sovereign cloud.
⚡ THE COMPETITIVE ADVANTAGEFrom racked hardware to a tenant-ready, HA OpenStack cloud typically under 4 hours, repeatedly, without bespoke automation. Compare with 6–12 weeks of consulting-led VMware VCF, OpenStack-from-scratch, or Nutanix bring-up engagements.
Zero-Touch Installation
The entire solution fits on a 5GB USB ISO. Rack the hardware, click deploy — full-stack OpenStack + Ceph + HA orchestration in under 4 hours.
Perpetual Licensing
Pay once per CPU socket. Own it forever. Absolutely no per-CPU-core tax, no renewal punishment, no surprise hikes. Predictable, drastically lower costs.
Enterprise-Grade Hardening
CIS-aligned defaults, TLS everywhere, encrypted control plane, mandatory MFA on admin. Continuous CVE pipeline; CVSS ≥ 9.0 hotfixes within 2–4 hours.
Stability That Scales
Not vanilla OpenStack. Integrated Distribution Validation + hardware certification. Effortlessly scales from a single node to clusters of 100,000 cores.
Online & Air-Gapped
Works in fully internet-connected or completely offline air-gapped environments. Offline license-key activation. PSU, defence and BFSI clouds run identically.
Optimized for AI/ML & HPC
Native GPU passthrough, vGPU partitioning, MIG slicing, RDMA networking, NVMe-optimized Ceph BlueStore. High-performance for inference, training and HPC.
Built for the workloads that matter most.
Enterprise-grade, integrated software stack for hosting VM and container workloads on-premises. Built on OpenStack + Ceph + open-source SDN wrapped in single-click deployment, hardened security, and 24×7 platinum support.
Government & PSU
Central & State departments, defence, telecoms. Air-gapped, sovereign, perpetual licensing — non-negotiable.
- Sovereign cloud, foreign-license free
- Air-gapped install for classified workloads
- GovTech / PSU procurement-ready
BFSI & Exchanges
Banks, NBFCs, brokerages, mutual funds. Hypervisor-modernisation RFPs running this quarter.
- Audit-ready CIS hardening
- RBI/SEBI compliance posture
- 99.99% production SLA
Academia & Research
IITs, NITs, IISc, research labs. HPC + AI/ML workloads. Budget-constrained, open-source friendly.
- Free training included (5 days)
- GPU passthrough for ML training
- Multi-tenant project quotas
Mid-Enterprise IT
Manufacturing, healthcare, ITeS. Modernising VM estate and exiting subscription-locked stacks.
- Free V2V migration from any KVM-/vSphere-class hypervisor
- Predictable per-socket pricing
- Channel partner ecosystem
AI / ML & HPC
Sovereign GPU clouds for foundation-model training, inference and scientific HPC.
- NVIDIA GPU passthrough & MIG
- RDMA & DPDK network paths
- Up to 100,000-core scale-out
Edge & Branch Cloud
Remote sites, retail branches, distributed dev/test labs, telco edge POPs.
- 3-node All-In-One footprint
- Same API as central cloud
- 10/25 GbE commodity hardware
When Digital Infrastructure Demands Sovereignty and Control
Pinaka ZTi powers most critical workloads through secure standalone clusters for intelligence and defense operations, and scalable multi-tenant clouds for large-scale public governance and citizen services.
Pinaka ZTi Production Footprint
Pinaka ZTi - Turnkey Cloud Appliance for AI/ML Workloads
Compute acceleration with GPU MIG and GPU Time-Slicing in PinakaZTi, delivering optimized GPU resource allocation, improved utilization, and greater flexibility for your virtualized environments.
Open-source ML platform
OpenStack-based environment for AI/ML. Flexible deployment of TensorFlow, PyTorch, JAX, Hugging Face. Foster innovation through open standards.
7-layer AI architecture
Infrastructure → Virtualization → Container Orchestration → ML Frameworks → MLOps Platform → AI Services → Application — fully integrated.
NVIDIA-accelerated
GPU PCIe Passthrough or NVIDIA vGPU drivers. From RTX 3080Ti to A100 to H100/L40S. RDMA + GPUDirect for distributed training.
Cost-controlled private cloud
Move AI off expensive public-cloud GPUs. Predictable per-socket pricing. Sovereign data residency. Built for enterprises, academia, research.
Three SKUs. From lab to foundation-model training.
Cirrus
Entry-level AI/ML lab
StratoCumulus
Mid-scale training & inference
GPU Pod
Foundation-model training · HPC
Squeeze every FLOP out of your NVIDIA fleet.
Pinaka ZTi gives you four GPU-sharing strategies — PCIe Passthrough, NVIDIA vGPU, Multi-Instance GPU (MIG), and Time-Slicing — combinable on the same cluster. Match performance, isolation and tenant density to each workload.
PCIe Passthrough
Entire physical GPU assigned to a single VM. Bare-metal performance, full CUDA feature set. Best for foundation-model training, large fine-tunes and HPC.
Complete — 1 GPU = 1 VM
NVIDIA vGPU (SR-IOV)
A physical GPU split across multiple VMs using SR-IOV virtual functions. Each VM gets dedicated frame buffer; compute is time-sliced or MIG-backed. Secure isolation via IOMMU.
Strong — dedicated VRAM, IOMMU-protected
Multi-Instance GPU
Hardware-level partition of an Ampere/Hopper GPU into up to 7 fully isolated instances. Each MIG slice has its own SMs, memory, L2 cache and bandwidth. Predictable QoS for noisy-neighbour-free serving.
Hardware-level — dedicated SMs, memory, cache
GPU Time-Slicing
Software-based round-robin scheduling. Many users share the same GPU sequentially. No memory or fault isolation, but maximum density. Works on older GPUs without MIG support and combines with MIG.
Soft — shared memory, sequential scheduling
Pro tip: combine MIG with Time-Slicing for hybrid density. MIG provides baseline hardware isolation; Time-Slicing multiplies user density inside each MIG instance — perfect for Kubernetes AI inference pipelines.
From bare metal to production cloud in under 4 hours.
Pick the footprint that matches your scale, redundancy and budget. The deployment methodology stays identical across all three - Zero-Touch.
Three blueprints. One platform.
3-Node All-In-One
Branch offices, dev/test, edge labs, remote sites
- 3× nodes (controller + compute + Ceph collapsed)
- 32–64 cores per node · 256–512 GB RAM
- Hyper-converged Ceph (3 replicas)
- 10/25 GbE networking
- ~30 production VMs supported
Disaggregated Cluster
Mid-market BFSI, regional banks, mid-PSU, enterprise IT
- 3× controllers (HA) + 6–10 compute nodes
- Dedicated 5–7 node Ceph cluster
- 64–96 cores per compute · 512 GB RAM+
- 25/100 GbE leaf-spine fabric
- 300–600 production VMs · 99.99% SLA
Multi-AZ / Multi-Region
PSU/Govt clouds, exchanges, telcos, AI/ML clusters
- 3+ availability zones, optional multi-region
- 20–500 compute nodes per AZ
- Dedicated Ceph + object storage tier
- 100/200 GbE spine · BGP-EVPN underlay
- GPU pools · 5,000+ VMs · 99.999% SLA
Six automated steps. ~3h 10m end-to-end.
PXE-boot · hardware inventory · BMC enrolment
BIOS settings · firmware · network · disk
Base OS · kernel · CIS hardening · agents
OpenStack roles · Ceph topology · SDN fabric
Smoke tests · HA failover · quota policies
Tenant project setup · admin training · go-live
Free V2V migration. 25 VMs per pack.
Inventory your existing hypervisor workloads.
Map workloads to Pinaka ZTi equivalents.
Deploy Pinaka ZTi via Zero-Touch.
Free V2V migration (up to 25 VMs/pack).
Unified Skyline dashboard for management.
Engineering long-term stability on an OpenStack foundation.
BFSI, PSU and government auditors don't accept generic OpenStack. Pinaka ZTi ships with hardening, advisory pipelines and a deliberate lifecycle architecture baked in.
Major Release Stream
Re-base on OpenStack SLURP releases · New hardware enablement · Architecture & features.
Minor Release Stream
Cumulative errata · No breaking changes · API stable.
Security Stream
Quarterly patches · Emergency hotfixes for CVSS ≥ 9.0 within 2–4 hours.
Innovation contained. Stability enforced. Security constant.
- Innovation is contained to Major releases — you control upgrade pace.
- Stability is enforced in Minor releases — no surprise API breaks.
- Security is constant — CVEs tracked daily across Debian + NVD + OSSN.
- Upgrade path: N + Skip (SLURP) managed — no painful N+1 linear hops.
- Current version v2.0 (Antelope Base) · Production / Enterprise Stable.
CIS-aligned hardening
Benchmark-driven baselines on every node. TLS everywhere, MFA on admin.
Continuous CVE pipeline
Debian + NVD + OSSN feeds polled daily. CVSS ≥ 9.0 → 2–4h hotfix SLA.
Auditable & observable
Centralised audit logs, encrypted control plane, OpenStack Security Notes tracking.
Air-gapped capable
Offline ISO + offline license activation. No phone-home, no telemetry leaks.
A transparent capability comparison.
Eighteen capability rows that procurement teams actually care about — licensing, networking, V2V, GPU sharing, lifecycle. Pinaka ZTi delivers every one perpetually.
| Capability | VMware | Nutanix | Proxmox | HPE | Pinaka ZTi |
|---|---|---|---|---|---|
| Hypervisor | vSphere | AHV / vSphere | KVM | KVM/vSphere | KVM |
| Cloud / VM Mgmt | vCenter | Prism | Web GUI | VM Ess. Mgr | Skyline Portal |
| Networking / SDN | NSX | Flow | Limited | VLAN/VXLAN | Neutron + OVN (full SDN) |
| Storage Backend | vSAN | NDFS | Ceph / ZFS | SAN / NAS | Ceph · Gluster · Swift |
| Perpetual licensing | |||||
| Single-click full-stack deploy | Zero-Touch | ||||
| 100% Air-gapped install | |||||
| Validated OpenStack distro | |||||
| No core-count tax | per-socket | ||||
| Live Migration | vMotion | Yes | Yes | Yes | Live Migration |
| Object Storage | Objects | Swift + Ceph | |||
| DBaaS | Era | Trove | |||
| Kubernetes (CaaS) | Tanzu | Karbon | LXC only | External | Magnum |
| LBaaS | NSX LB | VIP | External | Octavia | |
| V2V Migration | Converter | Move | Limited | Included free | |
| GPU MIG / vGPU | vGPU | Limited | Passthrough | Limited | PCIe · vGPU · MIG · Time-slice |
| Licensing Model | Per Core Sub | Subscription | Open-source | Subscription | Perpetual + SnS |
| Pricing | Very High | High | Low | Low | Low |
One perpetual license. Modular add-ons. Transparent SKUs.
Three blueprints. One platform. Pick the footprint that matches your scale, redundancy and budget. The deployment methodology stays identical across all three - Zero-Touch.
Private ZTi - License
License (Pinaka ZTi Cloud Software). Minimum 6 sockets license must be purchased. Licence cost of any 3rd party tools/softwares not included.
Professional Services (per day)
Cloud Design & Architecture · Add-On Services on T&M (GPU integration, 3rd Party tools etc) · Training the technical team. Minimum: 10 days deployment+testing, 5 days training.
VMware to Pinaka ZTi Migration
VMware to Pinaka ZTi V2V migrations supported · KVM to Pinaka ZTi V2V migrations supported · Single migration package covers up to 25 VMs. Additional packages available.
- • Minimum 6 sockets license must be purchased (PIN-160921).
- • Professional Services minimum: 10 days deployment + testing, 5 days training.
- • V2V single migration pack covers up to 25 VMs — buy additional packs for larger estates.
- • 1st-year support included with license. Year 2+ requires a valid AMC contract.
- • Add-on services available on T&M basis: GPU integration, 3rd-party tool integration.
- • Hardware support is not included in ZTi Premier Support; OEM warranty handles hardware.
24×7×365 support. Real engineers, real answers.
We know customers have varied preferences in engaging support resources. Register tickets, email us, or call our team — they'll be glad to address your issues and inquiries.
Pick the tier that matches your operational maturity.
- Installing & configuringPinakastra components installation
- Managing users & tenantsUser accounts and tenants
- Managing network & storageNetwork and storage resources
- Managing compute resourcesHypervisors and VMs
- Monitoring & troubleshootingPerformance monitoring
- Upgrading & patchingSecurity patches
- Installing & configuringPinakastra components installation
- Managing users & tenantsUser accounts and tenants
- Managing network & storageNetwork and storage resources
- Managing compute resourcesHypervisors and VMs
- Monitoring & troubleshootingPerformance monitoring
- Upgrading & patchingSecurity patches
- Installing & configuringPinakastra components (Keystone, Glance, Nova, Neutron, Horizon)
- Managing users & tenantsUser accounts, tenants, roles, permissions
- Managing network & storageConfigure network and storage for instances
- Managing compute resourcesHypervisors and VM instances
- Monitoring & troubleshootingPerformance and issue resolution
- Upgrading & patchingLatest security patches and updates
Raise a ticket on pinakastra.freshdesk.com
or email cloud@pinakastra.com
Login once with your AMC credentials engineers respond per your tier SLA.
Build a sovereign-cloud business on Pinaka ZTi.
A controlled, high-margin partner program. Built on a 20/20 margin model with up to 36% net effective discount capability. Quality over volume discount approvals are not automatic.
Authorised Distributor
Entry level
Foundation training, demo licenses
Preferred Distributor
Mid-tier · co-marketing
MDF, co-marketing funds, priority support
Strategic Distributor
Highest support · NFR · exclusivity
Regional exclusivity, NFR licenses, joint GTM
Get in touch with us.
Have a question or want to learn more? Drop us a message and we will get back to you.
Eight partner tracks. Pick your fit.
Server / appliance manufacturers shipping pre-imaged Pinaka ZTi.
Software vendors layering applications on Pinaka ZTi.
Managed private cloud as-a-service to end customers.
Public / sovereign cloud operators on Pinaka ZTi.
Colo operators offering BMaaS or sovereign cloud.
Integrators running cloud transformation and VMware migration.
Two-tier distribution aggregating volume across resellers.
Customer-owning reseller positioning Pinaka ZTi.
Ready to leave the subscription lock-in tax behind?
From discovery call to MoU in 30 days. Free V2V migration. Perpetual licensing. Real engineers on real customer clouds.
Discovery Call
Walk us through your VMware estate, growth plan and sovereignty constraints. We map a ZTi fit.
Pilot Cluster
We build a 3-node ZTi cluster on your hardware. Run real workloads. Validate the savings.
MoU & Rollout
Sign perpetual licensing. We co-engineer the full migration plan with V2V packs and AMC.