HP 12000 HP VLS Solutions Guide Design Guidelines for Virtual Library Systems - Page 64

Example 3: VLS12000 EVA Sizing Without Deduplication, VLS Blueprints

Page 64 highlights

Figure 24 Testing Backups to Tape vs. Backups to a VLS If you run two tests, in both cases backing up approximately 13 GB of data from the five servers, the following results: • Backing up to physical tape takes 6 minutes 41 seconds • Backing up the same data to the VLS takes 3 minutes 34 seconds Example 3: VLS12000 EVA Sizing Without Deduplication When sizing the EVA array requirements for a VLS12000 without deduplication enabled, the key factor to take into account is the compressibility of the data since the hardware compression built into the VLS12000 nodes will reduce the data rate going to the EVA arrays by the compression factor. For example, a VLS12000 with 5 nodes has a 3000 MB/s backup performance capability. With 2:1 data compressibility this would only require 1500 MB/sec of EVA array performance. Therefore, you would need at least three EVA 4400/6xxx/8xxx arrays (each array with 8 shelves full of disks) to meet that requirement. Example 4: VLS12000 EVA Sizing With Deduplication When sizing the EVA array requirements for a VLS12000 with deduplication enabled, first calculate the base array performance required without deduplication (see above) and then for deduplication you double this performance to get the required array performance. This is because the VLS deduplication is post-process and therefore will add additional disk I/O requirements on top of the backup I/O. For example, a VLS12000 with 3 nodes and deduplication has a 1500 MB/s backup performance capability. With 2:1 data compressibility this would only require 750MB/sec of base EVA array performance plus another 750MB/sec for the deduplication, giving a total of 1500MB/sec array performance. Therefore, you would need at least six EVA 4400/6xxx/8xxx arrays (each array with 8 shelves full of disks) to meet that requirement. VLS Blueprints The following blueprints of VLS virtual tape libraries with deduplication and replication start from specifying company requirements, then defining the HP blueprint for the solution, and finally defining any solution caveats or ISV dependencies associated with the solution. This will help you make informed decisions and allow you to quickly assess areas of concern and possible implementations. 64 VLS Devices

  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22
  • 23
  • 24
  • 25
  • 26
  • 27
  • 28
  • 29
  • 30
  • 31
  • 32
  • 33
  • 34
  • 35
  • 36
  • 37
  • 38
  • 39
  • 40
  • 41
  • 42
  • 43
  • 44
  • 45
  • 46
  • 47
  • 48
  • 49
  • 50
  • 51
  • 52
  • 53
  • 54
  • 55
  • 56
  • 57
  • 58
  • 59
  • 60
  • 61
  • 62
  • 63
  • 64
  • 65
  • 66
  • 67
  • 68
  • 69
  • 70
  • 71
  • 72
  • 73
  • 74
  • 75
  • 76
  • 77
  • 78
  • 79
  • 80
  • 81
  • 82
  • 83
  • 84
  • 85
  • 86
  • 87
  • 88
  • 89
  • 90
  • 91
  • 92
  • 93
  • 94
  • 95
  • 96
  • 97
  • 98
  • 99
  • 100
  • 101
  • 102
  • 103
  • 104
  • 105
  • 106
  • 107
  • 108
  • 109
  • 110
  • 111
  • 112
  • 113
  • 114
  • 115
  • 116
  • 117
  • 118
  • 119
  • 120
  • 121
  • 122
  • 123
  • 124
  • 125
  • 126
  • 127
  • 128
  • 129
  • 130
  • 131
  • 132
  • 133
  • 134
  • 135
  • 136
  • 137
  • 138
  • 139
  • 140
  • 141
  • 142
  • 143
  • 144
  • 145
  • 146
  • 147
  • 148
  • 149
  • 150
  • 151
  • 152
  • 153
  • 154
  • 155
  • 156
  • 157
  • 158
  • 159
  • 160

Figure 24 Testing Backups to Tape vs. Backups to a VLS
If you run two tests, in both cases backing up approximately 13 GB of data from the five servers,
the following results:
Backing up to physical tape takes 6 minutes 41 seconds
Backing up the same data to the VLS takes 3 minutes 34 seconds
Example 3: VLS12000 EVA Sizing Without Deduplication
When sizing the EVA array requirements for a VLS12000 without deduplication enabled, the key
factor to take into account is the compressibility of the data since the hardware compression built
into the VLS12000 nodes will reduce the data rate going to the EVA arrays by the compression
factor. For example, a VLS12000 with 5 nodes has a 3000 MB/s backup performance capability.
With 2:1 data compressibility this would only require 1500 MB/sec of EVA array performance.
Therefore, you would need at least three EVA 4400/6xxx/8xxx arrays (each array with 8 shelves
full of disks) to meet that requirement.
Example 4: VLS12000 EVA Sizing With Deduplication
When sizing the EVA array requirements for a VLS12000 with deduplication enabled, first calculate
the base array performance required without deduplication (see above) and then for deduplication
you double this performance to get the required array performance. This is because the VLS
deduplication is post-process and therefore will add additional disk I/O requirements on top of
the backup I/O. For example, a VLS12000 with 3 nodes and deduplication has a 1500 MB/s
backup performance capability. With 2:1 data compressibility this would only require 750MB/sec
of base EVA array performance plus another 750MB/sec for the deduplication, giving a total of
1500MB/sec array performance. Therefore, you would need at least six EVA 4400/6xxx/8xxx
arrays (each array with 8 shelves full of disks) to meet that requirement.
VLS Blueprints
The following blueprints of VLS virtual tape libraries with deduplication and replication start from
specifying company requirements, then defining the HP blueprint for the solution, and finally defining
any solution caveats or ISV dependencies associated with the solution. This will help you make
informed decisions and allow you to quickly assess areas of concern and possible implementations.
64
VLS Devices