Amazons claim with Silk was that it would learn your browsing patterns and cache important pages to speed them up over time. I'm curious if the speed situation would improve a week or two into normal use?
Also, a 10% reduction in bandwidth is pretty tiny, I think Opera Mobile claims 60% with Turbo Mode on. What they need to work on is an automatic switching mode that predicts whether it will speed up or slow down the situation, and I think Opera is working on that as well.
Opera mini might do even better, it does some kind of pre-rendering of all of the content, and compresses everything, images are compressed lossy. Should have a positive impact on CPU usage, and from my experience, much lower memory usage (at least on a Nokia 5800, haven't had a crash due to all memory being occupied on my SGS2 so far).
Note that on the image quality is noticeably worse if using max compression, but quality doesn't seem to have been measured in this test.
I've got great experiences with it, especially when the wireless coverage is poor (non 3g or on the edge of the wifi range).
I haven't used Opera Mini in a while, but from what I remember it was 2-3x faster than regular mobile browsers. Plus it really did use only 10-20% of the data. I really doubt Silk can come even close to Opera Mini/Turbo mode.
One issue here is that the cache was cleared in between test runs. In real-life use, you wouldn't normally clear the cache, so that could be a factor in it being slower.
Is it just me or is this the opposite of what it claims? According to the text, the bandwidth (data sent) should be lower when the accelerated page loading is on but its lower when it is off.
Or is it that using amazon's silk even increases bandwidth usage?
The second grapth does appear to be reversed. If Amazon trully plans to make a phone too(although i see little reason for it at this time,unless they got some new services to offer soon) then Silk could be built for the phone and this is more of a beta testing. Have you checked if Silk saves time at the next step,it was supposed to do that ,if i remember corectly,to try to guess what the next page will be and preload it (ofc they could just preload on their servers all pages a certain page links to). You shoudl also check non US based sites/mirrors,could make a difference even for well hosted sites.
I tested this as well - I clicked through all of the pages of our Interlagos review, there's still a 14% load time penalty with Accelerated Page Loading enabled. Prefetching is definitely interesting but it doesn't appear to be aggressively enabled here (if at all).
Yep, they're data mining to get more customer info. Surprised? It's a $200 tablet and they're looking to make up some of the money they're "losing" for selling it so cheaply. Not that I think they're really losing money, probably about breaking even, but they'll do what they can to turn a few more bucks. I'm sure it's somewhere in the EULA that they're doing something like this, and every owner agreed to it when they started using the Fire. That's life. I guess users can hope that they're not going to do anything too malicious with that info.
Seems kinda obvious, but any chance of running some tests on the Amazon.com website? Might be rather difficult in terms of it being a very dynamic site a lot of the time, but certainly it would be the most interesting test case given that it's Amazon hosted.
Opera Mobile with turbo enabled would be a better comparison. Mini renders the page completely in the cloud, its a completely different way of browsing, vs Silk which just caches DNS requests and whatnot. And really, you only need Mini for slow, older phones, anything newer and Mobile is usually faster (except the odd extremely heavy page like The Verge) and Mini's browsing mechanism does break some pages.
Way to miss the point. The USER does not give a damn HOW the technology works. What they care about is - does it speed up browsing? - does it reduce data transferred?
Saying "well Amazon does a crappy job because they chose to use a different approach" is not an excuse. Amazon claimed to achieve a certain goal, and they don't succeed --- it really is as simple as that.
the first graph is either mis-labeled or the conclusions in the article are wrong. It currently shows loading anandtech with acceleration on 4.85s with it off 6.94s. That's over 30% faster with acceleration turned on.
The article states that it got slower with acceleration, but the numbers (unless labeled wrong) sure show otherwise.
They can't encrypt an open wifi network. The communication from that point to and from Amazons servers might be encrypted, but you're still open between your laptop and the wireless access point.
But they are only seeing encrypted data on that open WiFi. It's like accessing a HTTPS/SSL site on an open network, the data is still encrypted between host and server.
Sorry, but you are incorrect about this. Anyone can sniff your traffic on an open network, but if the traffic is encrypted from your tablet to amazon's servers, they can't read it.
Well I can't speak for Amazon's implementation, but the problem with HTTPS on an open wifi network was that logins were secured, but often cookies containing passwords were not. I can't say it would be the same, but don't assume its completely secure without some testing.
FYI Consumer Reports' testing agrees that Silk is slower.
"The biggest surprise to our tests might be that we've found little advantage to Amazon's Silk technology, which purports to speed up loading Web pages on the Kindle Fire's browser. Indeed, in our tests so far, many pages took a bit longer to load when the Silk's supposed accelerator was on." (Nov 18, 2011 6:00 PM)
Lets see how good and fast Amazon is at coming out with an update that: A ) Disables "accelerator" or B ) Fixes their data mining "accelerator" before it becomes bad press.
But really, for data mining, do THEY really need to do this? They already get lots of customer information from their BUYING habits.
I think the main thrust for Amazon will be the data mining. But as was mentioned, the real test will be whether or not the average user will see a speed increase over time.
Amazon should be able to examine web address trends over time and then cache those sites to send direct to the user when requested. Depending on your surfing patterns, you may or may not see any different....depends on whether you surf with the pack or go your own way.
While this approach might save some DNS lookup time, it probably cannot compensate for the additional latency. After all, the bytes have to go over the Internet from the source to AWS and then to your Kindle Fire -- 2 hops instead of one. Amazon need to be alot more aggressive in caching pages, which unfortunately cannot work with many of the dynamic content sites.
You are wrong. This approach CAN work well. I use a (conceptually similar, in practice different approach) on my iPhone and iPad --- I route all URLs through either Google's or Instapaper's HTML rewriting services. So what I am doing is - send a URL to Google/Instapaper - the remote server reads the page, reformats it, strips out crud on the sides of blogs, shrinks the images, etc - the new page is sent to me.
Even with all that work, this is NOTICEABLY faster than connecting to the page directly, over either 3G or WiFi.
The problem is not that the concept can't work; it's that Amazon appear to have a lousy implementation. It's conceivable, of course, that they hired the same capacity planners that Apple seems to hire (over and over again :-[ ) so their servers are currently massively overloaded, and things will improve as they do a better job of handling the load.
Not to mention the pages will be cached. There are already free web proxy's out there that do this same thing.
The more active the site (more viewers) the better the cache will be kept up to date, thus a better chance for you to hit against the cache and not wait on the refresh.
Could you run a traceroute between your site's IPs and Amazon's IP and see if network congestion could be a culprit?
You might also want to mention how ping/latency makes a big difference in load times (not just about Mbps). Regardless, yet another great article - don't be surprised if I sign off on YAGA ;)
I think its "unique" in being crappier than Opera Turbo and Skyfire (well tbh I've never tried the latter). Opera Turbo can compress pages up to 60%, instead of 10% like Silk.
I'd like to see in more details, what Amazon is really doing with the caching.
* I'd like to see if there is any tracing/page prediction is going on? * I'd like to know if there is any change in the ability to cache the parts locally (like Amazon does determine that a certain image while marked by the original server as non cashable is in fact always the same and so Amazon relays it to the browser with cache flags. * I'd also like to know if there is any cookies that are exchanged in addition to the ones from the original web site. * The ability to read the cookie stream of all request would be a gold mine
Hopefully Anand can get someone from Amazon explain for what kinds of browsing behavior one would expect an acceleration. If there is non Amazon is vulnerable for false advertisement claims.
I was under the impression that Silk was supposed to reduce round trips. In loading a website, a large number of http requests are made, which can be burdensome over high latency 3g. I was under the impression that SILK was trying to address this (kinda like SPDY, except SILK remains a fully compatible protocol as its more of a proxy).
How high is the latency on your 3g service? Would you be able to test on higher latency (real or simulated)? I think this is the key to evaluating the value of SILK.
It is possible that because your content was being loaded from 13 different servers, there were 13 sockets open at the same time, and data was being loaded in parallel. However, with only a single server, the number of sockets would also be probably only 1. If this theory is correct, that might also impact the page load times.
Can you find out how many sockets were open? It will probably require code modifications / rooting to find this out...
"It's interesting that even Reddit, a text heavy site, was able to see some benefits from Amazon's accelerated page loading feature." It seems like you didn't pick many sites that are hosted by Amazon, which is where SILK should really shine, and that's in their advertising. Googling around, I found Reddit, Foursquare, and Quora use EC2. I think it might be useful to incorporate this information the next time you test this, maybe on their next release as the service improves and/or other major sites make in on to AWS.
Does anyone know what was used to identify the IP addresses with silk on vs off. I've run a couple trace tests and don''t see a difference in the ip addresses used.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
49 Comments
Back to Article
tipoo - Monday, November 21, 2011 - link
Amazons claim with Silk was that it would learn your browsing patterns and cache important pages to speed them up over time. I'm curious if the speed situation would improve a week or two into normal use?tipoo - Monday, November 21, 2011 - link
Also, a 10% reduction in bandwidth is pretty tiny, I think Opera Mobile claims 60% with Turbo Mode on. What they need to work on is an automatic switching mode that predicts whether it will speed up or slow down the situation, and I think Opera is working on that as well.Fakkness - Monday, November 21, 2011 - link
I was also curious about this. How does it compare to Opera's service?pandemonium - Monday, November 21, 2011 - link
Seconded. I'd be curious to see how Opera stacks up to this; being my mobile browser of choice. I don't use turbo though since my data is unlimited.KZ0 - Tuesday, November 22, 2011 - link
Opera mini might do even better, it does some kind of pre-rendering of all of the content, and compresses everything, images are compressed lossy. Should have a positive impact on CPU usage, and from my experience, much lower memory usage (at least on a Nokia 5800, haven't had a crash due to all memory being occupied on my SGS2 so far).Note that on the image quality is noticeably worse if using max compression, but quality doesn't seem to have been measured in this test.
I've got great experiences with it, especially when the wireless coverage is poor (non 3g or on the edge of the wifi range).
hicksman - Tuesday, November 22, 2011 - link
This is a great article -- exactly why I come here. ThanksAgreed w/ the OP: I'd like to see benches without clearing cache
Lucian Armasu - Tuesday, November 22, 2011 - link
I haven't used Opera Mini in a while, but from what I remember it was 2-3x faster than regular mobile browsers. Plus it really did use only 10-20% of the data. I really doubt Silk can come even close to Opera Mini/Turbo mode.kmmatney - Tuesday, November 22, 2011 - link
One issue here is that the cache was cleared in between test runs. In real-life use, you wouldn't normally clear the cache, so that could be a factor in it being slower.xfinalheavenx - Monday, November 21, 2011 - link
Is it just me or is this the opposite of what it claims? According to the text, the bandwidth (data sent) should be lower when the accelerated page loading is on but its lower when it is off.Or is it that using amazon's silk even increases bandwidth usage?
xfinalheavenx - Monday, November 21, 2011 - link
^^^Talking about the "Accelerated Page Loading Bandwidth Graph"
Anand Lal Shimpi - Monday, November 21, 2011 - link
Fixed, sorry about that!jjj - Monday, November 21, 2011 - link
The second grapth does appear to be reversed.If Amazon trully plans to make a phone too(although i see little reason for it at this time,unless they got some new services to offer soon) then Silk could be built for the phone and this is more of a beta testing.
Have you checked if Silk saves time at the next step,it was supposed to do that ,if i remember corectly,to try to guess what the next page will be and preload it (ofc they could just preload on their servers all pages a certain page links to).
You shoudl also check non US based sites/mirrors,could make a difference even for well hosted sites.
Anand Lal Shimpi - Monday, November 21, 2011 - link
I tested this as well - I clicked through all of the pages of our Interlagos review, there's still a 14% load time penalty with Accelerated Page Loading enabled. Prefetching is definitely interesting but it doesn't appear to be aggressively enabled here (if at all).Take care,
Anand
jjj - Monday, November 21, 2011 - link
Well,it can get better.Thanks for the replly
MagickMan - Monday, November 21, 2011 - link
Yep, they're data mining to get more customer info. Surprised? It's a $200 tablet and they're looking to make up some of the money they're "losing" for selling it so cheaply. Not that I think they're really losing money, probably about breaking even, but they'll do what they can to turn a few more bucks. I'm sure it's somewhere in the EULA that they're doing something like this, and every owner agreed to it when they started using the Fire. That's life. I guess users can hope that they're not going to do anything too malicious with that info.Lonyo - Monday, November 21, 2011 - link
Seems kinda obvious, but any chance of running some tests on the Amazon.com website?Might be rather difficult in terms of it being a very dynamic site a lot of the time, but certainly it would be the most interesting test case given that it's Amazon hosted.
StormyParis - Monday, November 21, 2011 - link
Too bad you didn't compare to a different and more mature implementation of the same technology, ie Opera Mini.tipoo - Monday, November 21, 2011 - link
Opera Mobile with turbo enabled would be a better comparison. Mini renders the page completely in the cloud, its a completely different way of browsing, vs Silk which just caches DNS requests and whatnot. And really, you only need Mini for slow, older phones, anything newer and Mobile is usually faster (except the odd extremely heavy page like The Verge) and Mini's browsing mechanism does break some pages.name99 - Tuesday, November 22, 2011 - link
Way to miss the point. The USER does not give a damn HOW the technology works. What they care about is- does it speed up browsing?
- does it reduce data transferred?
Saying "well Amazon does a crappy job because they chose to use a different approach" is not an excuse. Amazon claimed to achieve a certain goal, and they don't succeed --- it really is as simple as that.
tipoo - Tuesday, November 22, 2011 - link
Test both then, Mini is usually the slower of the two, hence why I suggested Opera Mobile for newer phones. Nothing to get your panties in wads over.DarthPierce - Monday, November 21, 2011 - link
the first graph is either mis-labeled or the conclusions in the article are wrong. It currently shows loading anandtech with acceleration on 4.85s with it off 6.94s. That's over 30% faster with acceleration turned on.The article states that it got slower with acceleration, but the numbers (unless labeled wrong) sure show otherwise.
Anand Lal Shimpi - Monday, November 21, 2011 - link
Hmm try clearing your cache, there was an issue with the labels being reversed initially but it should reflect the article's findings correctly now.Take care,
Anand
schizoide - Monday, November 21, 2011 - link
One advantage of silk is that the connection to amazon's servers is encrypted. If you're on public wifi, that's a nice benefit.Beyond that, yeah, there's no reason to ever turn it on.
tipoo - Monday, November 21, 2011 - link
If you have server side encryption you can still be vulnerable on an open/public wifi network.schizoide - Monday, November 21, 2011 - link
Huh? How? The entire connection to amazon is encrypted.tipoo - Monday, November 21, 2011 - link
They can't encrypt an open wifi network. The communication from that point to and from Amazons servers might be encrypted, but you're still open between your laptop and the wireless access point.solipsism - Monday, November 21, 2011 - link
But they are only seeing encrypted data on that open WiFi. It's like accessing a HTTPS/SSL site on an open network, the data is still encrypted between host and server.schizoide - Monday, November 21, 2011 - link
Sorry, but you are incorrect about this. Anyone can sniff your traffic on an open network, but if the traffic is encrypted from your tablet to amazon's servers, they can't read it.tipoo - Tuesday, November 22, 2011 - link
Well I can't speak for Amazon's implementation, but the problem with HTTPS on an open wifi network was that logins were secured, but often cookies containing passwords were not. I can't say it would be the same, but don't assume its completely secure without some testing.Rigan - Tuesday, November 22, 2011 - link
That right there is enough to leave the thing turned on.steven75 - Tuesday, November 22, 2011 - link
Misunderstanding the security encryption? No.regread - Monday, November 21, 2011 - link
FYI Consumer Reports' testing agrees that Silk is slower."The biggest surprise to our tests might be that we've found little advantage to Amazon's Silk technology, which purports to speed up loading Web pages on the Kindle Fire's browser. Indeed, in our tests so far, many pages took a bit longer to load when the Silk's supposed accelerator was on." (Nov 18, 2011 6:00 PM)
Source: http://news.consumerreports.org/electronics/2011/1...
Belard - Tuesday, November 22, 2011 - link
Lets see how good and fast Amazon is at coming out with an update that:A ) Disables "accelerator" or
B ) Fixes their data mining "accelerator"
before it becomes bad press.
But really, for data mining, do THEY really need to do this? They already get lots of customer information from their BUYING habits.
DrApop - Monday, November 21, 2011 - link
I think the main thrust for Amazon will be the data mining. But as was mentioned, the real test will be whether or not the average user will see a speed increase over time.Amazon should be able to examine web address trends over time and then cache those sites to send direct to the user when requested. Depending on your surfing patterns, you may or may not see any different....depends on whether you surf with the pack or go your own way.
doesnotcompute - Monday, November 21, 2011 - link
While this approach might save some DNS lookup time, it probably cannot compensate for the additional latency. After all, the bytes have to go over the Internet from the source to AWS and then to your Kindle Fire -- 2 hops instead of one. Amazon need to be alot more aggressive in caching pages, which unfortunately cannot work with many of the dynamic content sites.name99 - Tuesday, November 22, 2011 - link
You are wrong. This approach CAN work well. I use a (conceptually similar, in practice different approach) on my iPhone and iPad --- I route all URLs through either Google's or Instapaper's HTML rewriting services. So what I am doing is- send a URL to Google/Instapaper
- the remote server reads the page, reformats it, strips out crud on the sides of blogs, shrinks the images, etc
- the new page is sent to me.
Even with all that work, this is NOTICEABLY faster than connecting to the page directly, over either 3G or WiFi.
The problem is not that the concept can't work; it's that Amazon appear to have a lousy implementation. It's conceivable, of course, that they hired the same capacity planners that Apple seems to hire (over and over again :-[ ) so their servers are currently massively overloaded, and things will improve as they do a better job of handling the load.
vol7ron - Thursday, November 24, 2011 - link
Not to mention the pages will be cached. There are already free web proxy's out there that do this same thing.The more active the site (more viewers) the better the cache will be kept up to date, thus a better chance for you to hit against the cache and not wait on the refresh.
solipsism - Monday, November 21, 2011 - link
What about trafic between Amazon.com? Surely that was faster with accelerated page loading?vol7ron - Monday, November 21, 2011 - link
Anand,Could you run a traceroute between your site's IPs and Amazon's IP and see if network congestion could be a culprit?
You might also want to mention how ping/latency makes a big difference in load times (not just about Mbps). Regardless, yet another great article - don't be surprised if I sign off on YAGA ;)
vol7ron
s44 - Tuesday, November 22, 2011 - link
How is this "unique" when Skyfire has been doing it on Android for ages?tipoo - Wednesday, November 23, 2011 - link
I think its "unique" in being crappier than Opera Turbo and Skyfire (well tbh I've never tried the latter). Opera Turbo can compress pages up to 60%, instead of 10% like Silk.Conficio - Tuesday, November 22, 2011 - link
These results are certainly surprising.I'd like to see in more details, what Amazon is really doing with the caching.
* I'd like to see if there is any tracing/page prediction is going on?
* I'd like to know if there is any change in the ability to cache the parts locally (like Amazon does determine that a certain image while marked by the original server as non cashable is in fact always the same and so Amazon relays it to the browser with cache flags.
* I'd also like to know if there is any cookies that are exchanged in addition to the ones from the original web site.
* The ability to read the cookie stream of all request would be a gold mine
Hopefully Anand can get someone from Amazon explain for what kinds of browsing behavior one would expect an acceleration. If there is non Amazon is vulnerable for false advertisement claims.
mabellon - Tuesday, November 22, 2011 - link
Hey Anand,I was under the impression that Silk was supposed to reduce round trips. In loading a website, a large number of http requests are made, which can be burdensome over high latency 3g. I was under the impression that SILK was trying to address this (kinda like SPDY, except SILK remains a fully compatible protocol as its more of a proxy).
Actually, after a quick search I came across Amazon hiring for new SDEs to work on SPDY in SILK (http://aws.amazon.com/amazonsilk-jobs/)
How high is the latency on your 3g service? Would you be able to test on higher latency (real or simulated)? I think this is the key to evaluating the value of SILK.
Thanks
steven75 - Tuesday, November 22, 2011 - link
...it seems pretty clear you shouldn't buy the Fire for web browsing anyway, as it's not pleasant.Harshad - Sunday, November 27, 2011 - link
It is possible that because your content was being loaded from 13 different servers, there were 13 sockets open at the same time, and data was being loaded in parallel. However, with only a single server, the number of sockets would also be probably only 1. If this theory is correct, that might also impact the page load times.Can you find out how many sockets were open? It will probably require code modifications / rooting to find this out...
AnnonymousCoward - Monday, November 28, 2011 - link
It ranks with:-The Killer NIC
-Rage Pro graphics decelerator
-physics hardware acceleration
InnerDivinity - Thursday, December 1, 2011 - link
"It's interesting that even Reddit, a text heavy site, was able to see some benefits from Amazon's accelerated page loading feature."It seems like you didn't pick many sites that are hosted by Amazon, which is where SILK should really shine, and that's in their advertising. Googling around, I found Reddit, Foursquare, and Quora use EC2. I think it might be useful to incorporate this information the next time you test this, maybe on their next release as the service improves and/or other major sites make in on to AWS.
gh2m - Friday, December 9, 2011 - link
Does anyone know what was used to identify the IP addresses with silk on vs off. I've run a couple trace tests and don''t see a difference in the ip addresses used.tcboy88 - Monday, December 12, 2011 - link
As we all know Silk is partially copying Opera TurboBut we are not sure about their comparison
Can you make a comparison between them?
Thanks