Troubleshooting guide

1-24
Cisco Wide Area Application Services Configuration Guide
OL-26579-01
Chapter 1 Configuring Application Acceleration
Enabling and Disabling the Global Optimization Features
The changes are saved to the device or device group.
To configure video acceleration from the CLI, use the accelerator video global configuration command.
Configuring SMB Acceleration
The SMB application accelerator handles optimizations of file server operations. It can be configured to
perform the following file server optimizations:
• Read Ahead optimization—The SMB accelerator performs a read-ahead optimization on files that
use the oplock feature. When a client sends a read request for a file, it is likely that it may issue more
read requests for the same file. To reduce the use of network bandwidth to perform these functions
over the WAN on the file server, the SMB accelerator performs read-ahead optimization by
proactively reading more file data than what has been initially requested by the client.
• Directory listing optimization—A significant portion of the traffic on the network is for retrieving
directory listings. The SMB accelerator optimizes directory listings from the file server through
prefetching. For directory prefetching, a request from the client is expanded to prefetch up to 64 KB
of directory listing content. The SMB accelerator buffers the pre-fetched directory listing data until
the client has requested all the data. If the directory listing size exceeds 64 KB then a subsequent
request from the client is expanded by the SMB accelerator again to prefetch content up to 64 KB.
This continues until all the entries of the directory are returned to the client.
• Metadata optimization—The SMB accelerator optimizes fetching metadata from the file server
through metadata prefetching. Additional metadata requests are tagged along with the client request
and are sent to the file server to prefetch more information levels than what was requested by the
client.
• Named Pipe optimization—The SMB accelerator optimizes frequent requests from Windows
Explorer to the file server to retrieve share, server, and workstation information. Each of these
requests involves a sequence of operations that include opening and binding to the named pipe,
making the RPC request, and closing the named pipe. Each operation incurs a round trip to the file
server. To reduce the use of network bandwidth to perform these functions over the WAN on the file
server, the SMB accelerator optimizes the traffic on the network by caching named pipe sessions and
positive RPC responses.
• Write optimization—The SMB accelerator performs write optimization by speeding up the write
responses to the client by acknowledging the Write requests to the client whenever possible and, at
the same time, streaming the Write request over the WAN to the server.
• Not-Found Metadata caching—Applications sometimes send requests for directories and files that
do not exist on file servers. For example, Windows Explorer accesses the Alternate Data Streams
(ADS) of the file it finds. With negative Not-Found (NF) metadata caching, the full paths to those
nonexistent directories and files are cached so that further requests for the same directories and files
get local denies to save the round-trips of sending these requests to the file servers.
• DRE-LZ Hints—The SMB accelerator provides DRE hints to improve system performance and
resources utilization. At the connection level, the SMB accelerator uses the BEST_COMP latency
sensitivity level for all connections, as it gives the best compression. At the message level, the SMB
accelerator provides message-based DRE hints for each message to be transmitted over the WAN.
• Microsoft optimization—The SMB accelerator optimizes file operations for Microsoft applications
by identifying lock request sequences for file name patterns supported by Microsoft Office
applications.