CE content engine - cache engine

[vladrac-ccna]

Hello everyone.

Im a bit new to the subject, and got a few questions.

First we have 2 CEs running in our network and they are connected to the router just before the fw that goes out to the internet (ie all internet traffic goes through this router).

Our INTERNET traffic is pretty high reaching over 70mbps PEAK with a high average as well. (67% of it is HTTP)

However traffic towards the CE seems to be pretty low (1-3mbps).

Savings are around 20-35%.

I cant decypher the miss-reasons (is there a place to look for more information about this?)

not_in_cache: 177842

dmbuf_low: 0

none_get_method: 333757

ftp_not_anonymous: 3

http_not_anonymous: 0

suspicious_url: 473626

ie_5_ims: 0

has_if_match: 0

has_invalid_if_range: 1

has_if_unmodified_since: 0

has_invalid_range: 9

has_more_than_supported_range: 3

has_pragma_no_cache: 512188

has_authorization: 6190

has_cache_control_no_cache: 24056

is_https: 52357

invalid_ims: 4

cert_check_fail: 7

second_validation: 0

invalid_ims_reply: 301

ims_200_reply: 11154

xfs_open_error: 2153

has_unknown_length_transfer_pending: 20

object_in_cache_older_than_clients: 5288

object_in_cache_expired_cannot_verified: 2281

different_protocol: 0

other_error: 1001

I'm not sure why this traffic is this low and not sure why the savings are that low too.

Any place to get more information? Ive already looked at Cisco ACNS Software Configuration Guide

for Locally Managed Deployments , but it didnt give me all the answers.

Thanks,

Vlad

[pradeepde]

This behavior is normal. The CE is caching contents therefore the disk is full. This is not a problem and you can confirm this to the customer. The CE automatically removes older content to create disk space.

[vladrac-ccna]

Sorry, my question is not related to the HD, but to the difference between the actual INTERNET flow and the traffic that is being redirected to the CE.

If anyone has any experience with CEs, please let me know.

Vlad

[andrew.burns · Accepted solution]

Hi,

Content caching isn't an exact science - the saving you get depends entirely on the kind of content passing through the cache and to some extent on how it's configured.

If traffic towards the cache looks low then you need to look at the mechanism for directing traffic at the cache (wccp, proxy, etc) - you might have an access-list that only directs a subset of traffic to the cache, or maybe not all clients use the cache as a proxy?

If you need more detail on the miss reasons then look here:

http://www.cisco.com/en/US/products/sw/conntsw/ps491/products_command_reference_chapter09186a008067d292.html#wp5053357

As an example the miss-reason "has_pragma_no_cache" means that the web page itself has an intruction not to cache the page.

Finally - a saving of 20-35% is actually pretty good.

HTH

Andrew.

[vladrac-ccna]

Thx Andrew for your reply.

There's no ACL on the redirect and only a few user are using it as a proxy. For the rest of the user it should work as transp proxy.

Thx for the link, pretty good.

Vlad