Cisco Cisco Unified Contact Center Enterprise 9.0(2) Release Note

Page of 75
 
Cisco Unified ICM/Unified CC Enterprise & Hosted Editions, Releases 7.0(0) SR1 – SR4 and 7.1(x)  Rev. 1.14 
 
 
Hardware and System Software Specification 
 
3BServer Hardware Configuration Guidelines 
©2008 Cisco Systems, Inc. 
 
6
o
  10,000 RPM (minimum) for all other nodes 
Š  ATA 
o
  7,200 RPM 
Configuration guidelines: 
•  A dedicated on-board or add-in RAID controller must be used with a minimum of 128 MB of battery backed cache. 
•  Increasing the number of physical drives increases the overall fault tolerance of the disk array. 
•  Use multi-channel controllers connected to discrete drive bays or backplanes. 
Note 
Multiple controller channels can be of significant advantage when there are multiple drive bays and 
backplane connections.  Each channel of the controller can connect to a separate backplane connection,  and 
arrays split between the channels and backplanes can take advantage of the increased throughput as well as 
increased resiliency. 
Š  Two channels per external storage enclosure are highly recommended. 
Š  Multiple external storage enclosures are desirable (when needed) for increased performance and fault tolerance.  
Š  External storage enclosures with dedicated RAID controllers are supported with MCS Server systems. 
Supported configurations: 
Š  Fibre Channel is supported only in a Point-to-Point topology deployment. 
Š  Dedicated on-board or add-in RAID Controllers are required to use any of the RAID levels supported. 
Š  RAID 1 (Mirroring and Duplexing) – This is the minimum recommended RAID level for all critical Contact Center 
components.  See 
 for details.  Mirroring is typically used for boot drives 
on all servers to prevent loss of data and down time in the event of a disk failure.  
Š  RAID 10 (A Stripe of Mirrors) – This is the required RAID level for all Medium to Large ICM/IPCC Logger and HDS 
nodes.  RAID 10 offers the highest performance needed to meet the demands of SQL Server and the Logger or HDS, while 
still maintaining a safe level of redundancy.  
Unsupported configurations: 
Š  Fibre Channel Arbitrated Loop (FC - AL) fabric topology 
Š  Software based RAID provided by the Operating System or other software  
Š  Proprietary RAID solutions 
Š  RAID 5 (Block Interleaved Distributed Parity) 
Š  RAID 0 (Striped Disk Array without Fault Tolerance) 
Š  RAID 0+1 (A Mirror of Stripes) 
Caution 
Typically used for redundancy in file server applications, RAID 5 has been observed in product testing to manifest 
considerable performance degradation in write intensive applications.  RAID 5 is thus not supported for new 
deployments.  Please contact Cisco regarding use of RAID 5 in upgrades from previous releases.  
RAID 0 is not supported due to the lack of fault tolerance. If one drive fails, then all data in the array is lost. RAID 
0+1 is not supported due to increased risks of data loss or down time in the event of a failure. 
Unqualified storage and backup options: 
Š  NAS (Network Attached Storage) or SAN (Storage Area Network) products have not been qualified by Cisco for use with 
the ICM and IPCC applications.  These applications are qualified and tuned for optimal operation on a dedicated storage 
solution.  NAS or SAN solutions are typically deployed in a shared environment where non-ICM/IPCC enterprise 
applications are contending for storage access, and it becomes extremely difficult to appropriately size the system or