• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

WAN Optimization

Joined
Sep 22, 2008
Messages
878
We have an Exinda device that was put in place with the idea that we could use it to manage bandwidth utilization and control some of our internet based applications. I always thought there had to be a similar device on the other side of the WAN that supported QOS, caching, etc. We are just one location, how effective can an appliance like this be?

That said, it is no longer supported. Should I be looking at any WAN Optimizing device for upgrade or would I specifically want application acceleration?
 
The main problem with a WAN optimizer is that it only works on repeated traffic, because they cache the bits of packets they see repeated. Meaning that it may help with one particular website people access all the time, but it will only help with the parts that do not change. So unless people are looking at the exact same part of a web site day in and day out you probably wont even notice the help. If it is a document management site for example, it might help with the graphics of the site but do nothing for the files being downloaded, and since the parts that are cached are so miniscule you probably wont event notice.

Most people think WAN optimization magically make the internet faster, they do not.
 
Wont help with single site, but we've rolled out Silverpeak appliances to all of our sites, and they're very impressive.
 
I have a lot of experience with the Riverbed Steelhead lineup and its very impressive, speeding up almost everything we do across our MPLS WAN. However, it does indeed need a device on each end of the communication.
 
I have a lot of experience with the Riverbed Steelhead lineup and its very impressive, speeding up almost everything we do across our MPLS WAN. However, it does indeed need a device on each end of the communication.

Well riverbed is a file compressor. It does it on the fly and decompresses on the other end. It will package anything that can compress hyper fast and then decompress it just as fast on the other end. It isnt based on repeating frames or packets. It is based on compressible data.
 
Well riverbed is a file compressor. It does it on the fly and decompresses on the other end. It will package anything that can compress hyper fast and then decompress it just as fast on the other end. It isnt based on repeating frames or packets. It is based on compressible data.

Have you used them? It's more than that.
 
The Exinda has some functionality to choke bandwidth used by things like pandora, p2p, etc. Is this actually working or is it just saying traffic between appliances can only take x% of traffic

Is there such a thing for just a single office to manage bandwidth?
 
While we had our Exinda installed a single device shaped traffic successfully.
 
There is a mix of things going on in this thread.

Things like riverbed are WAN optimizers for things like remote offices where there is a device on both ends. Have used them and they work very well for that, especially on low bandwidth remote office connections.

Then there are traffic shapers that are more about QoS and managing bandwidth allocation to different devices and apps. Many routers will do this these days as well and prioritize traffic for certain services/apps/devices/etc.

Then there are things like caching reverse proxies that are more about removing load from INTERNAL servers for hosting web applications etc. These devices cache the websites and services being served from your location to lessen the load on other internal servers that are serving similar content (for example a web server serving the same page to 1000s of visitors).
 
The main problem with a WAN optimizer is that it only works on repeated traffic, because they cache the bits of packets they see repeated.

Not true at all. The devices also compress and optimize windowing (and various other things). You can see big performances increases regardless of caching.

You do need devices on both sides, though, in the case of Exinda, I believe it does some shaping/policing on any traffic you want, which would work with a single device. This isn't optimization though.

Riverbed is the best in this space and it does much more than has been stated in this thread.
 
Thanks for your help guys this helps clear it up. What it looks like to me is we are paying out the wazoo for a QOS appliance.
 
As already mentioned there are various types of optimizations where most needs a second device on the other side of the line to unwrap the optimizations.

Examples for these methods are:

- Realtime compression (using LZO or LZF or similar, this way the outside MTU must be insidemtu+1 and is performed per packet).

- Buffered compression (as above but fully utilizing the available MTU - no need to alter outside MTU).

- Altering tcp-stack options (like send and receive windows etc).

- Various application based optimizations (CIFS is a common example where the WAN optimizer can fool the sender to send faster than it otherwise would due to latency between the hosts).

etc...

When it comes to http you can use something called deflating which means that the payload can be compressed if the client supports this (most clients do).

You can do this in an appliance but today its often better to do this straight on the server itself.

In case you run Apache this is very easy to setup:

<IfModule mod_deflate.c>
DeflateBufferSize 8192
DeflateCompressionLevel 9
DeflateMemLevel 9
DeflateWindowSize 15
# AddOutputFilterByType DEFLATE text/html text/plain text/xml
AddOutputFilterByType DEFLATE text/html text/plain text/xml text/css application/javascript application/x-javascript application/x-httpd-php application/rss+xml application/atom_xml
</IfModule>

As an addition it can be healthy to optimize the http further by altering various cache-settings such as:

<IfModule mod_headers.c>
Header unset Pragma
Header unset Last-Modified
Header unset ETag
</IfModule>

<IfModule mod_expires.c>
ExpiresActive on
ExpiresByType image/gif "access plus 1 days"
ExpiresByType image/jpeg "access plus 1 days"
ExpiresByType image/png "access plus 1 days"
ExpiresByType text/css "access plus 1 days"
ExpiresByType application/javascript "access plus 1 days"
ExpiresByType application/x-javascript "access plus 1 days"
FileETag none
</IfModule>
 
Yes but how much more do I need to explain? There is only so much attention span to walls of text about a product.

Specific application optimizations and latency mitigation are some more things it does than just compress repeatable text. That alone is a life savor going across continents. SMB is very chatty and even if you had a 45Mb link, a single transfer might not go above 6Mb with a latency of 100ms or higher. Riverbed's help with that too. Things are "snappy".
 
The Riverbed Steelhead product doesn't need to see the same data more than once to improve performance. It's able to do compression and the applications/protocols it has specific optimizations for can get "LAN-like" performance because of the latency improvements it can make. For repeated data, it's about data patterns, not just the same web page, file, etc. These data patterns can show up in various transports. In addition, they do QoS (egress and ingress) with application identification.
 
Back
Top