Introduction
Related work
Cloud-fog model
Fog classifier—host load balancer with energy aware placement algorithm
Data Center—VM Placement | Dynamic Cost Award | Dynamic Energy Aware |
---|---|---|
Dynamically VMs are placed with Fog classifier algorithm which has combination of amount of utilized energy and cloud host | The cost factor is measure by using number of active host and energy availability from renewable sources. Here the power can be selected from smart grid model | Initialize the energy count set as 0 and increase the value based on usage of data center based on cost it will be impacted |
Procedure – Cloudlet Accessing the VM Step 1: Select the number of Cloudlet task from resource pool Step 2: Obtain the VM_List and request send to cloud broker Step 3: Check the information set whether VM is available, if available request send to cloud data center and allocate VM otherwise consume the amount energy and cost Step 4: Data center is selected means record and monitor the energy if it is increase means fog classifier classify the node and allocate another VM based on availability Step 5: Then the server is selected by number of utilization index and VM process Step 6: The task is classified based on execution time and waiting time from resource pool selection. Round Robin policy is taken for allocating VM using scheduled manner Step 7: If the process is completed note the time taken for execution and energy consumption factor and reallocated the VM as empty |
Conditions
Experimental setup
VM – Types | VM—RAM | VM – Memory | Bandwidth of VM | VM CPU | Instructions per seconds |
---|---|---|---|---|---|
Small_VM | 512 MB | 3 GB | 100 mb/s | 3 | 200 |
Medium_VM | 1 GB | 3 GB | 100 mb/s | 3 | 500 |
Large_VM | 2 GB | 3 GB | 100 mb/s | 3 | 1000 |
X-Large_VM | 5 GB | 3 GB | 100 mb/s | 3 | 2000 |
Data Center | Data center utilization | Available energy (KW) | Energy aware request from server (KW) | Cloud information set (Pool) | Power Unit |
---|---|---|---|---|---|
DC1 | 10% | 500 | 200 | 50 | Power(Total) |
DC2 | 20% | 1000 | 500 | 100 | Power(Total) |
DC3 | 50% | 1500 | 1000 | 150 | Power(Total) |
DC4 | 100% | 2000 | 1500 | 200 | Power(Total) |
Parameter | Index |
---|---|
Router Power Consumption | 500KW |
Transponder Request Power Consumption | 250KW |
Number of Bandwidth | 1000KW |
Data Rate | 50Gpbs |
Distance from Each Cluster | 5 to 10 kms |
Span Time | 1.5 ms |
Number of Data centers | 4 |
Number of Nodes | 20 |
Number of user | 800 |
User downloading rates | {1,5,10,25,50,100}Mbps |
Number of VMs | 50 |
VM Popularity based on User Downloading rate | {1,5,10,25,50,100}% |
VMs placement
CloudSim experimental inputs
Data Center: |
System Architecture = × 86 Operating System = Linux VM Type = Amazon EC2 Resource Cost = 1.0 Memory Cost = 0.05 Storage Cost per usage = 0.01 Bandwidth at Resource Level = 0.1 |
Server: |
HOST-TYPES = {1,2,5,10,25} HOST-IPS = { 500,1000,1500,2000} HOST-Power = 500KW HOST-RAM = {1 GB,2 GB,5 GB} HOST-BW = 1Gbps HOST-STORAGE = 2 GB HPPROLIANTML110G4XEON3040(), HPPROLIANTML110G5XEON3075()–2 Servers are used |
Virtual Machine: |
VM_TYPES = 5 VM_MIPS = {1000,500,250,100,50} VM_PES = 250KW VM_RAM = 1 GB VM_BW = 100Mbps VM_SIZE = 2.5 GB |
VMTypes | Data Center | No. of Host | Energy Consumption (KW) | VM Migration | SLA(%) | Accuracy (%) | No. of Shutdown |
---|---|---|---|---|---|---|---|
Small_VM | DC1 | 100 | 12.01 | 245 | 6 | 94 | 122 |
DC2 | 200 | 13.88 | 331 | 6 | 94 | 145 | |
DC3 | 500 | 15.78 | 467 | 5 | 95 | 134 | |
DC4 | 1000 | 21.34 | 568 | 5 | 95 | 156 | |
Medium_ VM | DC1 | 100 | 18.91 | 255 | 6 | 94 | 123 |
DC2 | 200 | 20.21 | 278 | 4 | 96 | 144 | |
DC3 | 500 | 27.89 | 346 | 7 | 93 | 167 | |
DC4 | 1000 | 30.12 | 452 | 5 | 95 | 178 | |
Large_VM | DC1 | 100 | 21.98 | 435 | 6 | 94 | 125 |
DC2 | 200 | 23.67 | 467 | 6 | 94 | 165 | |
DC3 | 500 | 29.02 | 521 | 4 | 96 | 178 | |
DC4 | 1000 | 31.05 | 547 | 5 | 95 | 165 | |
X-Large_VM | DC1 | 100 | 23.45 | 325 | 5 | 95 | 112 |
DC2 | 200 | 27.89 | 431 | 4 | 96 | 143 | |
DC3 | 500 | 30.05 | 478 | 6 | 94 | 154 | |
DC4 | 1000 | 33.65 | 522 | 5 | 95 | 167 |
VMTypes | Data Center | No. of Host | Virtualized Load Balancer | Decision Tree Index | Fog Classifier | |||
---|---|---|---|---|---|---|---|---|
Accuracy (%) | No. of Shutdown | Accuracy (%) | No. of Shutdown | Accuracy (%) | No. of Shutdown | |||
Small_VM | DC1 | 100 | 81 | 65 | 85 | 89 | 94 | 122 |
DC2 | 200 | 82 | 67 | 86 | 67 | 94 | 145 | |
DC3 | 500 | 80 | 63 | 83 | 68 | 95 | 134 | |
DC4 | 1000 | 78 | 45 | 84 | 78 | 95 | 156 | |
Medium_VM | DC1 | 100 | 79 | 56 | 83 | 87 | 94 | 123 |
DC2 | 200 | 76 | 57 | 87 | 76 | 96 | 144 | |
DC3 | 500 | 76 | 58 | 85 | 85 | 93 | 167 | |
DC4 | 1000 | 79 | 67 | 87 | 67 | 95 | 178 | |
Large_VM | DC1 | 100 | 81 | 54 | 88 | 89 | 94 | 125 |
DC2 | 200 | 83 | 56 | 84 | 76 | 94 | 165 | |
DC3 | 500 | 82 | 43 | 85 | 79 | 96 | 178 | |
DC4 | 1000 | 81 | 67 | 85 | 87 | 95 | 165 | |
X-Large_VM | DC1 | 100 | 79 | 57 | 86 | 88 | 95 | 112 |
DC2 | 200 | 75 | 78 | 88 | 83 | 96 | 143 | |
DC3 | 500 | 76 | 76 | 85 | 87 | 94 | 154 | |
DC4 | 1000 | 79 | 72 | 87 | 87 | 95 | 167 |