Logging requests on high traffic websites - iis

I wonder how high traffic websites handle traffic logging, for example a website like myspace.com receives a lot of hits, I can imagine it would take a lot of space to log all those requests, so, do they log every single request or how do they handle this?

If you view source on a MySpace page, you get the answer:
<script type="text/javascript">
var pageTracker = _gat._getTracker("UA-6293770-1");
pageTracker._setDomainName(".myspace.com");
pageTracker._setSampleRate("1"); //sets sampling rate to 1 percent
pageTracker._trackPageview();
</script>
That script means they're using Google Analytics.
They can't just gauge traffic using IIS logs because they may sell ads to third parties, and third parties won't take your word for how much traffic you get. They want independent numbers from a separate company, and that's where Google Analytics comes in.
Just for future reference - whenever you've got a question about how a web site is doing something, try viewing the source. You'd be amazed at what you can find there in plain view.

We had a similar issue with out Intranet which is used by hundreds of people. The disk activity was huge and performance was being hurt.
The short answer is Asynchronous non-blocking logging.

probably like google analytics.
Use Javascript to load a page on a difference server, etc.

Don't how they track it since I don't work there. I am pretty sure that they have enough storage to record every little thing about their user if they wanted.
If I were them, I would use AwStats if I just wanted to know basic stuff about my users.
It is more likely that they have developed their own scripts for tracking their users. Stuff they would log
-ip_address
-referrer
-time
-browser
-OS
and so on. Then a script to see different data about the user varying by day, weeks, or months. As brulak said, something along the line of Analytics, but since they have access to actual database, they can learn much more about their users.

ZXTM traffic shaping and logging, speaking from experience here

I'd be extremely surprised if they didn't log every single request, yes, and operations with particularly high traffic volumes usually roll their own log-management solutions against the raw server logs, in some form or other -- sometimes as simple batch-type processes, sometimes as complete subsystems.
One company I worked for, back in the dot-com heyday, got upwards of twenty million pageviews a day; for that site (actually a set of them, running across a few dozen machines in all, as I recall), our ops team wrote a quite sophisticated, clustered solution in C that parsed, translated (into relational storage), compressed and distributed the logs daily. Log files, especially verbose ones, pile up fast, and the commercial solutions available at the time just couldn't cut it.

If by logging you mean for collecting server related information (request and response times, db and cpu usage per request etc) I think they sample only the 10% or 1% of the traffic. That gives the same results (provide developers with auditing information) without filling in the disks or slowing the site down.

Related

How to detect inbound HTTP requests sent anonymously via Tor?

I'm developing a website and am sensitive to people screen scraping my data. I'm not worried about scraping one or two pages -- I'm more concerned about someone scraping thousands of pages as the aggregate of that data is much more valuable than a small percentage would be.
I can imagine strategies to block users based on heavy traffic from a single IP address, but the Tor network sets up many circuits that essentially mean a single user's traffic appears to come from different IP addresses over time.
I know that it is possible to detect Tor traffic as when I installed Vidalia with its Firefox extension, google.com presented me with a captcha.
So, how can I detect such requests?
(My website's in ASP.NET MVC 2, but I think any approach used here would be language independent)
I'm developing a website and am
sensitive to people screen scraping my
data
Forget about it. If it's on the web and someone wants it, it will be impossible to stop them from getting it. The more restrictions you put in place, the more you'll risk ruining user experience for legitimate users, who will hopefully be the majority of your audience. It also makes code harder to maintain.
I'll post countermeasures to any ideas future answers propose.
You can check their ip address against a list of Tor Exit Nodes. I know for a fact this won't even slow someone down who is interested in scraping your site. Tor is too slow, most scrapers won't even consider it. There are tens of thousands of open proxy servers that can be easily scanned for or a list can be purchased. Proxy servers are nice because you can thread them or rotate if your request cap gets hit.
Google has been abused by tor users and most of the exit nodes are on Google black list and thats why you are getting a captcha.
Let me be perfectly clear: THERE IS NOTHING YOU CAN DO TO PREVENT SOMEONE FROM SCRAPING YOUR SITE.
By design of the tor network components it is not possible for the receiver to find out if the requester is the original source or if it's just a relayed request.
The behaviour you saw with Google was probably caused by a different security measure. Google detects if a logged-in user changes it's ip and presents a captcha just in case to prevent harmful interception and also allow the continuation of the session if an authenticated user really changed its IP (by re-logon to ISP, etc.).
I know this is old, but I got here from a Google search so I figured I'd get to the root concerns in the question here. I develop web applications, but I also do a ton of abusing and exploiting other peoples. I'm probably the guy you're trying to keep out.
Detecting tor traffic really isn't the route you want to go here. You can detect a good amount of open proxy servers by parsing request headers, but you've got tor, high anonymity proxies, socks proxies, cheap VPNs marketed directly to spammers, botnets and countless other ways to break rate limits. You also
If your main concern is a DDoS effect, don't worry about it. Real DDoS attacks take either muscle or some vulnerability that puts strain on your server. No matter what type of site you have, you're going to be flooded with hits from spiders as well as bad people scanning for exploits. Just a fact of life. In fact, this kind of logic on the server almost never scales well and can be the single point of failure that leaves you open to a real DDoS attack.
This can also be a single point of failure for your end users (including friendly bots). If a legitimate user or customer gets blocked you've got a customer service nightmare and if the wrong crawler gets blocked you're saying goodbye to your search traffic.
If you really don't want anybody grabbing your data, there are some things you can do. If it's a blog content or something, I generally say either don't worry about it or have summary only RSS feeds if you need feeds at all. The danger with scraped blog content is that it's actually pretty easy to take an exact copy of an article, spam links to it and rank it while knocking the original out of the search results. At the same time, because it's so easy people aren't going to put effort into targeting specific sites when they can scrape RSS feeds in bulk.
If your site is more of a service with dynamic content that's a whole other story. I actually scrape a lot of sites like this to "steal" huge amounts of structured proprietary data, but there are options to make it harder. You can limit the request per IP, but that's easy to get around with proxies. For some real protection relatively simple obfuscation goes a long way. If you try to do something like scrape Google results or download videos from YouTube you'll find out there's a lot to reverse engineer. I do both of these, but 99% of people who try fail because they lack the knowledge to do it. They can scrape proxies to get around IP limits but they're not breaking any encryption.
As an example, as far as I remember a Google result page comes with obfuscated javscript that gets injected into the DOM on page load, then some kind of tokens are set so you have to parse them out. Then there's an ajax request with those tokens that returns obfuscated JS or JSON that's decoded to build the results and so on and so on. This isn't hard to do on your end as the developer, but the vast majority of potential thieves can't handle it. Most of the ones that can won't put in the effort. I do this to wrap really valuable services Google but for most other services I just move on to some lower hanging fruit at different providers.
Hope this is useful for anyone coming across it.
I think the focus on how it is 'impossible' to prevent a determined and technically savvy user from scraping a website is given too much significance. #Drew Noakes states that the website contains information that when taken in aggregate has some 'value'. If a website has aggregate data that is readily accessible by unconstrained anonymous users, then yes, preventing scraping may be near 'impossible'.
I would suggest the problem to be solved is not how to prevent users from scraping the aggregate data, but rather what approaches could be used to remove the aggregate data from public access; thereby eliminating the target of the scrapers without the need to do the 'impossible', prevent scrapping.
The aggregate data should be treated like proprietary company information. Proprietary company information in general is not available publicly to anonymous users in an aggregate or raw form. I would argue that the solution to prevent the taking of valuable data would be to restrict and constrain access to the data, not to prevent scrapping of it when it is presented to the user.
1] User accounts/access – no one should ever have access to all the data in a within a given time period (data/domain specific). Users should be able to access data that is relevant to them, but clearly from the question, no user would have a legitimate purpose to query all the aggregate data. Without knowing the specifics of the site, I suspect that a legitimate user may need only some small subset of the data within some time period. Request that significantly exceed typical user needs should be blocked or alternatively throttled, so as to make scraping prohibitively time consuming and the scrapped data potentially stale.
2] Operations teams often monitor metrics to ensure that large distributed and complex systems are healthy. Unfortunately, it becomes very difficult to identify the causes of sporadic and intermittent problems, and often it is even difficult to identify that there is a problem as opposed to normal operational fluctuations. Operations teams often deal with statistical analysed historical data taken from many numerous metrics, and comparing them to current values to help identify significant deviations in system health, be they system up time, load, CPU utilization, etc.
Similarly, requests from users for data in amounts that are significantly greater than the norm could help identify individuals that are likely to be scrapping data; such an approach can even be automated and even extended further to look across multiple accounts for patterns that indicate scrapping. User 1 scrapes 10%, user 2 scrapes the next 10%, user 3 scrapes the next 10%, etc... Patterns like that (and others) could provide strong indicators of malicious use of the system by a single individual or group utilizing multiple accounts
3] Do not make the raw aggregate data directly accessible to end-users. Specifics matter here, but simply put, the data should reside on back end servers, and retrieved utilizing some domain specific API. Again, I assuming that you are not just serving up raw data, but rather responding to user requests for some subsets of the data. For example, if the data you have is detailed population demographics for a particular region, a legitimate end user would be interested in only a subset of that data. For example, an end user may want to know addresses of households with teenagers that reside with both parents in multi-unit housing or data on a specific city or county. Such a request would require the processing of the aggregate data to produce a resultant data set that is of interest to the end-user. It would prohibitively difficult to scrape every resultant data set retrieved from numerous potential permutations of the input query and reconstruct the aggregate data in its entirety. A scraper would also be constrained by the websites security, taking into account the # of requests/time, the total data size of the resultant data set, and other potential markers. A well developed API incorporating domain specific knowledge would be critical in ensuring that the API is comprehensive enough to serve its purpose but not overly general so as to return large raw data dumps.
The incorporation of user accounts in to the site, the establishment of usage baselines for users, the identification and throttling of users (or other mitigation approaches) that deviate significantly from typical usage patterns, and the creation of an interface for requesting processed/digested result sets (vs raw aggregate data) would create significant complexities for malicious individuals intent on stealing your data. It may be impossible to prevent scrapping of website data, but the 'impossibility' is predicated on the aggregate data being readily accessible to the scraper. You can't scrape what you can't see. So unless your aggregate data is raw unprocessed text (for example library e-books) end users should not have access to the raw aggregate data. Even in the library e-book example, significant deviation from acceptable usage patterns such as requesting large number of books in their entirety should be blocked or throttled.
You can detect Tor users using TorDNSEL - https://www.torproject.org/projects/tordnsel.html.en.
You can just use this command-line/library - https://github.com/assafmo/IsTorExit.

Difference between Ad company statistics, Google Analytics and Awstats on adult sites

I have this problem. I have web page with adult content and for several past months i had PPC advertisement on it. And I've noticed a big difference between Ad company statistics of my page, Google Analytics data and Awstats data on my server.
For example, Ad company tells me, that i have 10K pageviews per day, Google Analytics tells me, that i have 15K pageviews and on Awstats it's around 13K pageviews. Which system should I trust? Should i write my own (and reinvent a wheel again)? If so, how? :)
The joke is, that i have another web page, with "normal" content (MMORPG fan site) and those numbers are +- equal in all three systems (ad company, GA, Awstats). Do you think it's because it's not adult oriented page?
And final question, that is totally offtopic, do you know about Ad company that pays per impression and don't mind adult sites?
Thanks for the answers!
First, you should make sure not to mix up »hits«, »files«, »visits« and »unique visits«. They all have a different meaning and are sometimes called differently. I recommend you to look up some definitions if you are confused about the terms.
awstats has probably the most correct statistics, because it has access to the access.log from the web server. Unfortunately, a cached site (maybe cached by the browser, a proxy from an ISP or your own caching server) might not produce a hit on the web server. Especially if your site is served with good caching hints which don't enforce a revalidation and you are running your own web cache (e.g. Squid) in front of your site, the number will be considerable lower, because it only measures the work of the web server.
On the other hand, Google Analytics is only able to count requests from users which haven't blocked Google Analytics and have JavaScript enabled (but they will count pages served by a web cache). So, this count can be influenced by the user, but isn't affected by web caches.
The ad-company is probably simply counting the number of requests which they get from your site (probably based on their access.log). So, to get counted there, the add must not be cached and must not be blocked by the user.
So, as you can see, it's not that easy to get a single correct value. But as long as you use the measured values in comparison to those from the previous months, you should get at least a (nearly) correct rate of growth.
And your porn site probably serves a high amount of static content (e.g. images from the disk) and most of the web servers are really good at serving caching hints automatically for static files. Your MMORPG on the other hand, might mostly consist of some dynamic scripts (PHP?) which don't send any caching hints at all and web servers aren't able to determine those caching headers for dynamic content automatically. That's at least my explanation, without knowing your application and server configuration :)

Is there such a thing as a reverse CDN? (content 'retrieval' network)

Our clients upload a serious amount of data from all over the world and we'd like to do our best to make that as painless as possible. Our clients upload 2GB worth of files over their sometimes very 'retail' broadband packages (with capped upload speeds) that draw out upload times to 24-48 hours. At any given time we have 10 or more concurrent uploads and peek periods we can have 100 concurrent uploads. So we decided to consider ways to reduce latency and keep our clients traffic local... so just as a CDN has download servers in various locations, we'd like upload servers.
Any experience or thoughts?
We're not a huge company but this is a problem worth solving so we'll consider all options.
What about putting some servers physically closer to your clients ?
Same ISP, or at the very least in the same countries. Then you just collect it on schedule. I don't imagine that they're getting top speeds when there's 100 of them uploading to you either, so the sooner you can get them completed the better.
Also, do they need to upload this stuff immediately ?? Can some of them post DVD for whatever isn't time sensitive ? I know it sux dealing with media in the post.... so it's hardly ideal.
A reverse CDN sort of situation would only really happen if you had multiple clients using torrents and seeding their uploads (somehow) to one of your servers.
You haven't really said if this is a problem for you, or your clients. So, some more info is going to get you a better answer here.
2GB per what time period? Hour? Day?
If your operation is huge, I wouldn't be too surprised if Akamai or one of the other usual CDN suspects can provide this service to you for the right price. You might get your bizdev folks (or purchasing) in touch with them.

How long to retain an archive of web server traffic logs?

We've currently got four web servers in a farm generating IIS web logs about 100Mb per day. These can be compressed pretty effieciently down to somewhere around 5% of their size.
We are planning to use waRmZip to move them off the servers and onto a SAN. After a week or so we can be confident we don't have any technical issues to investigate so the only other thing would be using them for trend analysis as a compliment to Google Analytics.
What retention periods do people recommend? Are there any legal requirements to keep this data?
Legal requirements will depend on your country, how much you're logging, and quite possibly the nature of your business. Talk to your company's lawyers - legal advice on SO is likely to be worth what you pay for it.
If you're only storing 5MB per day, you should be able to store them for basically as long as you want without worrying on the technical front.
Please consider the sensitivity of your web log data as well. I have no idea whether access to your web apps would be considered sensitive if made public, but you need to realize that your web logs contain the necessary information to potentially identify individuals (esp. in conjunction with other information available elsewhere). Your privacy policies should reflect how long you retain these logs and what purposes to which they will be put. Google, I think, recently decided to anonymize their logs after 9 months to help protect user privacy. Granted, their situation is a little different since they collect so much information, but you need to consider your customer's needs as well as your own when determining how long and in what form to keep your logs.
I tend to keep mine forever. That's mainly for trend analysis because Google misses some visitors (non-JavaScript ones).

Using Datamining/Statistics for Log Monitoring

I have a large set of log files that I want to characterize or possibly add some kind of decision tree or some kind of analytics. But I don't know exactly what. What kind of analysis have you done with log files, a lot of log files.
For example, so far I am collecting how many requests are made to a particular page for a given log file.
Servlet = 60 requets
Servlet2 = 70 requests, etc.
I guess right there, filter by only the most popular requests. Also, might do something like 60 requests given a 2 hour period. 60 / 160 minutes.
Deciding what analysis to do depends on what decisions you're trying to make based on that analysis. For example, I currently monitor logs for exceptions reported by our application (all exceptions in the client application are logged with the server) to decide what should be high priority client bugs to investigate. I also use log searching software to monitor for any Exceptions reported by our server software which may need more immediate investigation. On top of the logs generated by everything anyway, I also use some monitoring software to track usage of our web server and database server which records usage stats etc. in a database. The final aim of this is to predict future usage levels and purchase more hardware as appropriate to keep up with demand.
Two (free) tools I've been using are:
Hyperic for monitoring, it's pretty easy to set up and might be able to start logging a lot of data you may be interested in, ie requests per second on a web server.
Splunk for searching log files, it's very easy to get set up and work with and gives you excellent searching capabilities over your log files. If you're working with log files right now and haven't tried out splunk I definitely recommend it. I have noticed a couple of moments of 100% cpu whilst using it on our main production server so stopped running it on that machine recently, just a word of warning.
Not sure what your aim is with this analysis, mine has been very much about looking for any errors I should know about, and planning for future capacity needs. If you're interested in the latter I'd also recommend The Art of Capacity Planning.

Resources