Compaq ProLiant 1000 Compaq Backup and Recovery for Microsoft SQL Server 6.X - Page 50

Chart 11 - Backup of One Server Across Single 10BaseT Segment, network saturates

Page 50 highlights

Compaq Backup and Recovery for Microsoft SQL Server 6.x Page 50 The network operating system used on all servers in this section was Windows NT 4.0. The network protocol selected was TCP/IP. Networking hardware included: Compaq NetFlex-3 and Netelligent network interface cards. Compaq Netelligent 10Base-T Repeater. Compaq Netelligent 100Base-TX Repeater. CAT 5 UTP cables (supports both 10 Mbs & 100 Mbs). Both servers were cabled to the 10Base-T repeater for the 10 Mbit tests, then re-connected to the 100Base-T repeater for the 100 Mbit tests. The 100 Mbit technology chosen was 100Base-TX. Compaq supports other 100 Mbit Ethernet technology such as 100VG-Any LAN, but this was not tested. 100Base-TX supports full-duplex mode for bi-directional throughput up to 200 Mb/s. However, since most of the data transfer done here is unidirectional (from source server to destination server), benefits of full-duplex would be minimal and were not tested. Here we show the performance characteristics of backing up a database residing on one server to storage devices on another server, by transmitting a single data path across a 10Base-T network segment. Using ARCserve and its RAID option, the backup was done to a single 15/30-GB DLT tape drive as well as to an array of from 1 to 3 tape drives. SQL Server native dump was used to send data to a disk array. Chart 11 - Backup of One Server Across Single 10BaseT Segment The results obtained here are somewhat predictable. Given that the maximum transfer rate of a DLT tape drive (not counting compression) as well as a 10Base-T network are the same (1.25 MB/sec), one would expect to saturate the network wire with just a single DLT at the destination end. Furthermore, no increase in performance would be expected by striping multiple DLT's into a tape array using the ARCserve RAID Option. This is exactly what is shown in Chart-11, where the network saturates at

  • 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

Compaq Backup and Recovery for Microsoft SQL Server 6.x
Page
50
The network operating system used on all servers in this section was Windows NT 4.0.
The network
protocol selected was TCP/IP.
Networking hardware included:
Compaq NetFlex-3 and Netelligent network interface cards.
Compaq Netelligent 10Base-T Repeater.
Compaq Netelligent 100Base-TX Repeater.
CAT 5 UTP cables (supports both 10 Mbs & 100 Mbs).
Both servers were cabled to the 10Base-T repeater for the 10 Mbit tests, then re-connected to the
100Base-T repeater for the 100 Mbit tests.
The 100 Mbit technology chosen was 100Base-TX.
Compaq supports other 100 Mbit Ethernet technology such as 100VG-Any LAN, but this was not
tested.
100Base-TX supports full-duplex mode for bi-directional throughput up to 200 Mb/s.
However, since most of the data transfer done here is unidirectional (from source server to destination
server), benefits of full-duplex would be minimal and were not tested.
Here we show the performance characteristics of backing up a database residing on one server to
storage devices on another server, by transmitting a single data path across a 10Base-T network
segment.
Using ARCserve and its RAID option, the backup was done to a single 15/30-GB DLT tape
drive as well as to an array of from 1 to 3 tape drives.
SQL Server native dump was used to send data
to a disk array.
Chart 11 - Backup of One Server Across Single 10BaseT Segment
The results obtained here are somewhat predictable.
Given that the maximum transfer rate of a DLT
tape drive (not counting compression) as well as a 10Base-T network are the same (1.25 MB/sec), one
would expect to saturate the network wire with just a single DLT at the destination end.
Furthermore,
no increase in performance would be expected by striping multiple DLT’s into a tape array using the
ARCserve RAID Option.
This is exactly what is shown in Chart-11, where the
network saturates at