User manual

Chapter 3: CacheRaQ 4 Administration
46 Cobalt CacheRaQ 4 User Manual
• Number of bytes sent to clients from cache misses
• Total number of bytes sent to clients
• Average cache hit object size. This is the average document size for
documents which were delivered from the cache.
• Average cache miss object size. This is the average document size for
documents which had to be retrieved from their original servers.
• Cache hit rate. This number refers to the ratio of cache hits to total cache
requests.
• Bandwidth savings. This number reflects the amount of client HTTP traffic
(as a percentage) that was avoided because the documents were available in
the cache, thus saving network bandwidth.
• Average cache hit latency. This is the average amount of time (in seconds)
it took the CacheRaQ 4 to deliver a document from its cache to a client. The
time measured is the difference between the time that the CacheRaQ 4
receives the request initially to the time that the CacheRaQ 4 has finished
sending the complete object to the client.
• Average cache miss latency. This is the average amount of time it took the
CacheRaQ 4 to deliver a document that had to be retrieved from the original
server. It is calculated similarly to the cache hit latency (above), except that
only cache misses are considered.
• Weighted average request latency. This value is calculated by the
following formula:
(Average cache hit latency X Cache hit rate) + (Average cache miss latency
X Cache miss rate).
The cache miss rate is equal to 1 minus the cache hit rate.
• Request speedup. This number quantifies the CacheRaQ 4’s performance
improvement in terms of user-perceived latency. It is the ratio of cache miss
latency to the weighted average request latency.
The Summary Statistics output also includes three graphs. These graphs
illustrate:
1. the number of client HTTP requests versus the number of cache hits as a
function of time of day
2. the total amount of traffic (Kb) to clients versus the amount of traffic (Kb) to
clients from the cache as a function of time of day
3. the average cache hit latency versus the average cache miss latency as a
function of time of day