Servers, hyper-converged infrastructure, and the virtualization technology that powers every cloud VM.
The physical layer consists of the real hardware in the data center. Servers come as rack-mounted units or as blades inserted into a shared chassis that provides power, cooling, and networking.
البنية الفيزيائية هي الأجهزة الحقيقية في مركز البيانات. السيرفرات تأتي بأشكال مختلفة: Rack مستقل، وBlade رفيع يوضع داخل هيكل (Chassis) مشترك.
Rack servers are like standalone books on a shelf; blades are like thin books sharing one common spine (the chassis).
A self-contained server with its own power and cooling, mounted in a standard 19-inch rack.
A 2U server bolted into a data-center rack.
A stripped-down server module that relies on a shared chassis for power, cooling, and networking.
Multiple blades sliding into one blade enclosure.
The enclosure housing multiple blades and providing shared power, cooling, and I/O.
A 10U chassis holding 16 blade servers.
| Aspect | Rack Server | Blade Server |
|---|---|---|
| Power & cooling | Self-contained | Shared via chassis |
| Density | Lower | Higher |
| Cabling | More cables | Reduced (shared backplane) |
| Cost (small scale) | Lower upfront | Higher (needs chassis) |
| Best for | Small / mixed loads | Dense, large deployments |
Hyper-Converged Infrastructure is a software-centric system delivering compute, storage, and networking in a tightly integrated platform. It typically consists of multiple server nodes, direct-attached storage, a hypervisor on each node, an intelligent software layer for orchestration, and built-in data protections (mirroring, replication, erasure coding).
HCI يجمع الحوسبة والتخزين والشبكة والهايبرفايزر في نظام واحد مُدار بالبرمجيات. يتكون من عقد سيرفرات متعددة مع تخزين مباشر وطبقة برمجية ذكية تدير الموارد.
An all-in-one printer (print + scan + fax) instead of three separate machines.
Software-defined platform combining compute, storage, network, and virtualization in one node/cluster.
Nutanix or VMware vSAN clusters.
Manages and orchestrates resources across all HCI nodes.
Software that pools storage and compute across the cluster.
Virtualization abstracts physical resources into virtual ones. Infrastructure virtualization covers system, storage, and network. Process virtualization creates isolated execution environments for applications (e.g. JVM, .NET, Docker containers, Solaris Zones).
الافتراضية تخلق نسخة وهمية من المورد فوق العتاد الحقيقي. نوعان: افتراضية البنية (سيرفر/تخزين/شبكة) وافتراضية العمليات (JVM، Docker، Zones).
One apartment building (physical server) divided into many separate apartments (VMs).
Virtualizing hardware resources (servers, storage, network) into pools.
Splitting one physical server into several VMs.
Creating a virtual execution environment for a single application/process.
Java VM, .NET runtime, Docker containers, Solaris Zones.
Abstracts compute, storage, and network so multiple VMs run on one physical machine.
VMware ESXi, Xen, Microsoft Hyper-V.
A hypervisor (Virtual Machine Monitor) is the software layer that creates and runs virtual machines, allocating physical resources to each. Type 1 (bare-metal) runs directly on hardware; Type 2 (hosted) runs on a host OS.
الهايبرفايزر (VMM) هو طبقة برمجية تدير الأجهزة الافتراضية وتوزع موارد العتاد بينها. نوعان: Type 1 على العتاد مباشرة، وType 2 فوق نظام تشغيل.
A building manager who divides one building's resources (water, power) fairly among all tenants (VMs).
Software that creates, runs, and manages virtual machines.
VMware ESXi, Microsoft Hyper-V, KVM.
Runs directly on hardware; high performance.
VMware ESXi, Xen, Hyper-V installed on the server itself.
Runs on top of a host operating system.
VirtualBox, VMware Workstation.
VMM multiplexes VMs on CPUs using time slicing.
Round-robin scheduler sharing CPU among VMs.
VMM partitions and maps virtual memory to physical RAM.
Each VM gets isolated memory mappings.
Virtualizes I/O devices so VMs access storage and network.
Virtual NICs and virtual disks presented to VMs.
| Aspect | Type 1 (Bare-metal) | Type 2 (Hosted) |
|---|---|---|
| Runs on | Hardware directly | A host OS |
| Performance | Higher | Lower (extra OS layer) |
| Use case | Data centers / production | Desktops / testing |
| Example | ESXi, Hyper-V, KVM | VirtualBox, VMware Workstation |
VMs provide isolation (a failure in one VM doesn't affect others), encapsulation (the whole VM is a set of files), and portability (move VMs between hosts easily).
من أهم فوائد الأجهزة الافتراضية: العزل بين الأجهزة، وتغليف الجهاز كملفات، وسهولة نقله بين السيرفرات.
Each VM is a sealed lunchbox: contents don't spill into others (isolation), packed as one box (encapsulation), and easy to carry anywhere (portability).
Each VM is independent; a crash or attack in one does not affect others.
A virus in one VM cannot reach neighboring VMs.
The entire VM (state, disk, config) is stored as a set of files.
Copying a VM is just copying its files.
VMs can be moved or migrated between physical hosts.
Live-migrating a VM to another server with no downtime.
Three approaches to CPU virtualization: full virtualization (guest OS unmodified, uses binary translation), paravirtualization (guest OS modified to cooperate via hypercalls), and hardware-assisted (CPU extensions like Intel VT-x / AMD-V help).
هناك ثلاثة أساليب لافتراضية المعالج: الكاملة (محاكاة دون تعديل النظام)، وشبه الافتراضية (تعديل النظام ليتعاون)، والمدعومة بالعتاد (المعالج نفسه يساعد).
Full = translator does all the work; Para = both parties learn a shared language; Hardware-assisted = a built-in real-time translator chip.
| Aspect | Full | Paravirtualization | Hardware-Assisted |
|---|---|---|---|
| Guest OS modified? | No | Yes | No |
| Technique | Binary translation | Hypercalls | CPU extensions (VT-x/AMD-V) |
| Performance | Moderate (overhead) | High | High |
| Compatibility | High (any OS) | Lower (needs modified OS) | High |
CPU virtualization uses time slicing — the VMM schedules VMs on physical CPUs (round-robin, work-conserving). When #VCPUs > #physical CPUs, the system oversubscribes CPU resources. Memory virtualization maps each VM's virtual address space to physical RAM with VMM-controlled isolation.
افتراضية المعالج تستخدم تقسيم الوقت (Time Slicing) مع جدولة Round Robin. يمكن Oversubscribe أي عدد VCPUs أكبر من المعالجات الفعلية.
Time slicing is like one cashier serving many customers in quick turns; oversubscription is booking more appointments than rooms, assuming not everyone shows at once.
Dividing CPU time into small slices shared among VMs.
Each VM gets a few milliseconds of CPU in turn.
Allocating more virtual CPUs than physical cores.
8 VMs each with 2 vCPUs on a 4-core host.
Mapping each VM's virtual memory to physical memory through the hypervisor.
Two VMs each 'see' 4GB while sharing 6GB of physical RAM.
Each VM is stored as a set of files — this is encapsulation in practice.
كل جهاز افتراضي يُخزن كمجموعة ملفات: ملف الإعدادات، القرص الافتراضي، اللقطات (Snapshots)، وملفات السجل.
A VM folder is like a complete 'machine in a box' made of files.
Defines the VM's hardware settings (CPU, RAM, NICs).
.vmx file in VMware.
Stores the VM's operating system and data.
.vmdk / .vhdx file.
Captures the VM's state at a point in time for rollback.
Snapshot before installing an update.
Records VM activity and events for troubleshooting.
vmware.log.
Each physical subsystem maps to a virtualization type: CPU → processor virtualization (time slicing), Memory → memory virtualization (address mapping), I/O devices → I/O virtualization (virtual NICs/disks), Storage → storage virtualization (abstract logical view of physical disks).
كل مورد في السيرفر يمكن افتراضيته: المعالج (time slicing)، الذاكرة (mapping)، أجهزة I/O (NICs، أقراص)، والتخزين (logical view فوق الأقراص الفيزيائية).
Like dividing a house into rooms — each room (subsystem) gets its own virtual space with shared utilities.
VMM multiplexes VMs on CPUs using scheduling.
Round-robin time slicing across 4 VMs.
Presents virtual I/O devices to VMs.
Virtual NIC and virtual disk to each VM.
Abstracts network resources into virtual switches/VLANs.
vSwitch connecting VMs on one host.
Provides logical view of physical storage.
LVM pooling disks into one volume.
Thanks to encapsulation and portability, VMs support live migration (move a running VM between hosts), suspend/resume, snapshots (point-in-time state capture), and clones (full copy as a new VM). Migration is independent of physical hardware.
بفضل Encapsulation و Portability، يمكن نقل VM بين hosts (live migration)، تجميدها (suspend)، أخذ snapshot للحالة، أو استنساخها (clone) كنسخة مستقلة.
Moving a packed suitcase (VM files) from one apartment to another without unpacking.
Move a running VM to another host with minimal downtime.
vMotion moving a VM during maintenance.
Freeze VM state to disk; resume later on same or different host.
Pausing a VM before host reboot.
Full copy of a VM creating an independent duplicate.
Clone a golden-image VM for new deployments.
| Type | Purpose | Independent? | Note |
|---|---|---|---|
| Snapshot | Point-in-time rollback | No (depends on parent) | Not a full backup |
| Clone | Duplicate VM | Yes | Full independent copy |
| Backup | Disaster recovery | Yes | Stored off-system |