Virtual Machines on Hikube
Hikube Virtual Machines (VMs) offer complete hardware infrastructure virtualization, ensuring the execution of heterogeneous operating systems and business applications in isolated environments that meet enterprise security requirements.
🏗️ Architecture and Operation
Compute and Storage Separation
Hikube uses a decoupled architecture between compute and storage that ensures optimal resilience:
💻 Compute Layer
- The VM runs on physical servers in one of the 3 datacenters
- If a node fails, the VM is automatically restarted on another node
- If a datacenter fails, the VM is automatically restarted on another node in one of the 2 remaining datacenters
- Downtime is limited to restart time (generally < 2 minutes)
💾 Storage Layer (Persistent)
- VM disks are automatically replicated across multiple physical nodes with "replicated" storage
- No data loss even in case of multiple hardware failures
- Disks survive failures and remain attachable to the relocated VM
This separation ensures that your data is always safe, even if the physical server hosting your VM becomes unavailable or if a datacenter becomes unavailable. We guarantee the resources!
Multi-Datacenter Architecture
⚙️ Instance Types
Complete Range for All Needs
Hikube offers three instance series optimized for different usage profiles, ensuring performance adapted to each workload:
S Series - Standard (Ratio 1:2)
Compute-oriented instances with a CPU/memory ratio of 1:2, ideal for CPU-intensive workloads.
| Instance | vCPU | RAM | Typical Use Cases |
|---|---|---|---|
s1.small | 1 | 2 GB | Light services, proxies |
s1.medium | 2 | 4 GB | Workers, batch processing |
s1.large | 4 | 8 GB | Scientific computing |
s1.xlarge | 8 | 16 GB | Rendering, compilation |
s1.3large | 12 | 24 GB | Intensive applications |
s1.2xlarge | 16 | 32 GB | HPC, simulations |
s1.3xlarge | 24 | 48 GB | Distributed computing |
s1.4xlarge | 32 | 64 GB | Massive computing |
s1.8xlarge | 64 | 128 GB | Exascale computing |
U Series - Universal (Ratio 1:4)
Versatile instances offering optimal balance between CPU and memory for most enterprise applications.
| Instance | vCPU | RAM | Typical Use Cases |
|---|---|---|---|
u1.medium | 1 | 4 GB | Dev, tests, micro-services |
u1.large | 2 | 8 GB | Web applications, APIs |
u1.xlarge | 4 | 16 GB | Business applications |
u1.2xlarge | 8 | 32 GB | Intensive workloads |
u1.4xlarge | 16 | 64 GB | Critical applications |
u1.8xlarge | 32 | 128 GB | Enterprise applications |
M Series - Memory (Ratio 1:8)
High memory instances with a CPU/memory ratio of 1:8 for memory-hungry applications.
| Instance | vCPU | RAM | Typical Use Cases |
|---|---|---|---|
m1.large | 2 | 16 GB | Redis caches, Memcached |
m1.xlarge | 4 | 32 GB | In-memory databases |
m1.2xlarge | 8 | 64 GB | Analytics, Big Data |
m1.4xlarge | 16 | 128 GB | SAP HANA, Oracle |
m1.8xlarge | 32 | 256 GB | Data warehouses |
- Intensive computing, CI/CD → S Series (ratio 1:2, CPU optimized)
- Classic web applications → U Series (ratio 1:4, balanced)
- Databases, Analytics → M Series (ratio 1:8, memory optimized)
🔒 Isolation and Security
Multi-Tenant by Design
Each VM benefits from complete isolation thanks to a secure architecture that strictly partitions resources between different tenants. This isolation relies on several complementary protection layers:
- Tenant : Logical separation of resources at the application level, each tenant having its own execution space
- Kernel isolation : Network and process isolation at the Linux kernel level, ensuring no VM can access another's resources
- Storage classes : Automatic encryption and data isolation, with cryptographic separation of volumes by tenant
🌐 Connectivity and Access
Native Access Methods
Access to Hikube virtual machines is done via native mechanisms integrated into the platform, eliminating the need for complex network infrastructure. The serial console provides direct low-level access independent of the network, ideal for debugging and system maintenance. For graphical environments, VNC allows connection to the VM's user interface via secure tunnels. Traditional SSH access remains available either via virtctl ssh which automatically manages connectivity, or directly via the assigned external IP. Application services can be selectively exposed via controlled port lists that intelligently filter traffic without compromising tenant security.
Software-Defined Network
Hikube's network architecture is based on a Software-Defined approach that completely virtualizes the network layer. Each VM automatically receives a private IP in a network segment isolated by tenant, ensuring isolation while allowing internal communication. The system can optionally assign a public IPv4 IP for external exposure, with automatic routing that maintains secure segmentation. The distributed firewall applies granular security policies directly at each VM level, with restrictive default rules that dynamically adapt to application needs.
📦 Migration and Portability
Import of Existing Workloads
The Hikube platform facilitates migration of existing infrastructures thanks to universal import mechanisms that preserve workload integrity. Standardized cloud images (Ubuntu Cloud Images, CentOS Cloud) integrate natively for immediate deployment with cloud-native optimizations. For custom installations, importing ISO images allows recreating custom environments while preserving all specific configurations. VMware snapshots are automatically converted from VMDK to RAW format, ensuring seamless transition from traditional virtualization infrastructures. Compatibility with Proxmox and OpenStack formats (QCOW2) guarantees interoperability with most existing cloud solutions.
Lifecycle Management
The lifecycle management system integrates automated mechanisms that ensure operational continuity of virtual machines. Snapshots instantly capture the complete state of the VM, including memory and storage, to enable precise rollbacks during maintenance or incidents. Automatic backup orchestrates scheduled disk backups with configurable retention, automatically replicated across the three datacenters to guarantee recovery in case of disaster. Live migration moves VMs between physical nodes without service interruption, facilitating hardware maintenance and load optimization without impact on critical applications.
🚀 Next Steps
Now that you understand Hikube VM architecture:
🏃♂️ Immediate Start
→ Create your first VM in 5 minutes
📖 Advanced Configuration
→ Complete API Reference
For production, always use the replicated storage class and size your VMs with at least 2 vCPU to benefit from better performance.