The brain of the cloud: discovering, grading, allocating, and optimizing compute, storage, and network resources.
The control layer receives service-layer requests and provisions resources (e.g. a VM with 4 GB RAM). It configures, monitors, and pools resources. Element managers configure single infrastructure types (storage array, network, compute). Unified managers provide a single interface via native APIs across all elements.
طبقة التحكم تستقبل طلبات من Service Layer وتخصص الموارد (مثل VM بـ 4GB RAM). Element Manager يدير مكوناً واحداً، Unified Manager يوفر واجهة موحدة عبر compute/storage/network.
Element manager = a department head; Unified manager = the CEO overseeing all departments.
Manages a single device type — config, expand capacity, troubleshoot, monitor.
Storage array manager for RAID, LUN masking, zoning.
Centralized management across compute, storage, network via element APIs.
Single pane of glass discovering and provisioning all resources.
Resource discovery inventories infrastructure on a schedule or on demand. Compute: blade servers, CPU speed, memory, virtual-to-physical mapping. Network: switch model, adapters, VLAN/VSAN IDs, QoS, topology, zones. Storage: system type, drive type, total/free/used capacity, RAID level, storage pools.
اكتشاف الموارد يجمع تفاصيل compute (blade servers, CPU, memory)، network (switch model, VLAN, VSAN, QoS, topology, zones)، وstorage (capacity, RAID, pools, drive type).
Discovery = taking stock of the warehouse; management = running it efficiently.
| Domain | Discovered Items |
|---|---|
| Compute | Blade servers, CPU speed, memory, VM-to-physical mapping |
| Network | Switch model, adapters, VLAN/VSAN, QoS, topology, zones |
| Storage | System type, drive type, capacity, RAID, pools, mappings |
Resource grading tiers resources by quality/performance (Gold = best, then Silver, Bronze). Allocation can be relative (shares/weights — proportional) or absolute (fixed reserved amount/limit).
تصنيف الموارد بمستويات جودة (Gold/Silver/Bronze) حسب الأداء. التخصيص إما نسبي (حصص/أوزان) أو مطلق (كمية محددة ثابتة).
Grading = airline classes (first/business/economy). Relative = splitting a pizza by shares; Absolute = a guaranteed fixed number of slices.
Service tiers ranking resource quality/performance.
Gold tier on fast SSD; Bronze on slow HDD.
Proportional allocation via shares/weights.
VM A gets twice the shares of VM B.
Fixed guaranteed/limited amount (reservation/limit).
Reserve exactly 2 GHz and 4 GB for a VM.
| Aspect | Relative | Absolute |
|---|---|---|
| Basis | Shares / weights | Fixed amount |
| Behavior | Proportional under contention | Guaranteed reservation/limit |
| Example | Platinum 2X vs Gold 1X shares | 2–4 GB RAM bounds; VM won't start if min unavailable |
| Tier | Drives | Tiering | RAID |
|---|---|---|---|
| Gold | Flash + FC + SATA | Yes | RAID 5 |
| Silver | Flash + FC + SATA | Yes | RAID 1+0 |
| Bronze | FC only | No | RAID 5 |
Compute optimizations: hyperthreading (one core appears as two logical CPUs), memory page sharing (deduplicate identical memory pages across VMs), dynamic memory (allocate RAM to VMs on demand), and VM load balancing (distribute VMs across hosts).
تقنيات تحسين الحوسبة: Hyperthreading (نواة تظهر كنواتين)، مشاركة صفحات الذاكرة المتطابقة، الذاكرة الديناميكية حسب الحاجة، وموازنة أحمال الأجهزة الافتراضية بين المضيفين.
Hyperthreading = one chef juggling two orders; page sharing = roommates sharing identical books; dynamic memory = adjustable portions; load balancing = spreading diners across tables.
One physical core presents as two logical cores.
8 cores → 16 logical CPUs.
Deduplicates identical memory pages across VMs.
Identical OS pages shared once in RAM.
Adjusts a VM's RAM allocation at runtime based on demand.
Memory grows/shrinks as the VM needs.
Distributes VMs across hosts to balance load.
Migrating VMs off an overloaded host.
Uses server-side flash to cache frequently accessed data.
SSD cache accelerating VM I/O.
Storage optimizations: thin/thick provisioning, storage pool rebalancing when drives are added, automated tiering (hot/cold data), and cache tiering (SSD cache layer in front of slower tiers).
تحسين التخزين: Thin/Thick provisioning، إعادة توازن Storage Pool عند إضافة أقراص، Tiering، وCache Tiering (SSD كطبقة كاش أمام HDD).
Tiering = keeping frequently used files on your desk; cache tiering = a fast 'inbox' tray for hot items.
Redistributes data when new drives are added to a thin pool.
Adding disks triggers pool rebalance before LUN expansion.
Auto-moves data between fast and slow tiers by access frequency.
Hot DB data on SSD tier, archives on HDD tier.
Uses fast media (SSD) as a cache in front of slower storage.
SSD cache accelerating an HDD pool.
Network optimizations: limit-and-share policies, traffic shaping (rate limits per tenant), link aggregation, NIC teaming, and multipathing. Resource provisioning allocates from graded pools per the service catalog/template.
تحسين الشبكة: Limit and Share، Traffic Shaping، Link Aggregation، NIC Teaming، Multipathing. Resource Provisioning يخصص الموارد من graded pools حسب service catalog.
Link aggregation = merging several lanes into a wider road; multipathing = multiple routes to the same destination for safety.
Network bandwidth policies limiting or sharing traffic per tenant/client.
Fair share across VMs on a host.
Controls bandwidth usage by limiting/prioritizing rates.
Capping backup traffic to 100 Mbps.
Combines multiple physical links into one logical link.
Four 10G links bonded into 40G.
Groups NICs for higher throughput and failover.
Two NICs teamed for redundancy.
Multiple physical paths to storage for redundancy/balancing.
Two HBAs reaching one LUN via two paths.
Allocating resources to workloads (manual or automated).
Auto-provisioning a VM with CPU, RAM, storage, network.
The provisioning workflow: (1) Discovery inventories available resources, (2) Grading categorizes pools by performance/capacity/availability, (3) Allocation assigns resources (relative shares or absolute bounds), (4) Provisioning creates VMs/volumes/networks per the service catalog, (5) Monitoring tracks performance, availability, and capacity.
سير عمل التخصيص: 1) Discovery لمعرفة الموارد، 2) Grading لتصنيفها (Gold/Silver/Bronze)، 3) Allocation (relative أو absolute)، 4) Provisioning حسب service catalog، 5) Monitoring للأداء والسعة.
Like hiring: inventory candidates (discovery), rank by skill (grading), assign roles (allocation), onboard (provisioning), and review performance (monitoring).
Predefined resource templates consumers can request.
Gold VM: 4 GB RAM, 80 GB HDD, RAID 5 storage.
A provisioned VM or service for a consumer.
Customer's running VM with assigned resources.
Aggregating physical resources for dynamic allocation.
CPU/memory pool shared across VMs.
An element manager handles a single infrastructure type: initial configuration (RAID, LUN masking, zoning, VLANs, guest OS install), capacity expansion (detect and add new resources to pools), troubleshooting, and monitoring performance, availability, capacity, and security.
Element Manager يدير مكوناً واحداً: إعداد أولي (RAID، LUN masking، zoning، VLANs)، توسيع السعة عند إضافة موارد، استكشاف الأخطاء، ومراقبة الأداء والتوفر والسعة والأمان.
A specialist mechanic for one type of vehicle — knows every detail of that component.
First-time setup of an infrastructure element.
Configure RAID 5, create LUNs, set up zoning.
Detect newly added hardware and add to pools.
New drives auto-added to storage pool.