It takes 3000 years for a single computer to solve chess. But if we keep adding a computer to solving the problem, wouldn’t that divide the time it takes by 1/N for each computer we add? For example if we launched maybe tens of thousands of smaller computers, maybe using a cloud platform like AWS, and had each computer work on a small sub-problem of chess, store the results in a cloud database so the other computers know what subproblems have already been solved or are being worked on, we could solve chess in less than a week.
Why wouldn’t this work? And if it could work why has no one attempted this yet?
For a few days now one of my websites (a local travel guide) has been getting a massive number of hits from Chinese IP addresses. This is to the point that my log files are 10 times bigger and my modsecurity log is over 100 times bigger than usual.
The IP addresses are very fluid. I currently have over 5,000 banned in Fail2Ban, and this is with a China-based IPv4 block in the htaccess.
The user agents for these hits are nearly all either LieBaoFast, MQQBrowser, Mb2345Browser or MicroMessenger.
The pages they are hitting are a dynamic search page e.g. search-accommodation?param=1¶m2=2 However, they are calling loads of parameters – which puts extra stress on the server.
So, my question is why? Basically what is going on? I can’t see scraping my (affiliate generated) accommodation listings are of benefit to anyone, and if it is some sort of DDOS that’s even weirder as this site isn’t even important to me?!
The problem I encounter is that I have thousands of items a day (e.g. transactions of toys that have been purchased in a store). My goal is to provide the users with an interface to fill in manually 10+ data fields on this transaction (e.g. toy category, is it the main toy or something you purchase as extra, what age is it for, etc.). Then, I would categorize the data by the information the users provided me. How would you recommend doing it?
I’ve had my website hacked. I have cleaned it and searched google site:site.com and made a list of all the results. Removed the actual links that are ok and made a list of the ones in the format of site.com/ad0-b1fermarte54eb17chb-1244425. I’ve submited these links to Google to remove and Google removed these links.
The problem here is that after my initial search, that resulted in 200 URLs which we’ve removed, other new links appeared out of nowhere. These links don’t actually work anymore and return a 404 page, but Google still has these in its results.
We’ve tried deleting our sitemap and re-submitting it. Google has crawled it, but again, it has not removed the dummy search results.
Looking in our google webmaster console, we saw that in the coverage tab the number of indexed links spiked from 230 (normal usage) to 10.900 links – I’m guessing these are all dummy links that were infected and now cleaned.
Any magic way of having Google remove dead links automatically? Or a way to force it to recrawl the entire website?
I have a website that is being hit with invalid URL requests by thousands of distinct IP addresses, never the same one used twice. Most of them are in a few ranges of IP addresses and often just go up sequentially.
Could this be a zombie botnet of compromised devices, or is it possible the attacker is spoofing these addresses?
The clustering of IP addresses into a handful of ranges seems inconsistent with what I would expect from random devices across the world being compromised and part of a botnet.
User agents are all legitimate and quite varied, but I know that is simple to spoof.
It doesn’t feel like a DDOS attack as it is “just” a few thousand per hour. If they really wanted to DDOS it seems like they would crank the volume up more. Once I adjusted some exception handling I was able to get my server to resume being responsive to legitimate use.
I suspect it is a malicious (poorly constructed) crawler/spider.
Is IP address spoofing easily done and common now in these scenarios?
in my project there is a form. One of the field is problematic for me. User should choose an option from a very long list (thousands of options). I wanted to do it by combobox, but I don’t know how searching should works.
We can’t display options after typing by user first letter, because list will be still very long .
I thought that we could display options after 3 letters.
User have to choose one name from the list (10 000 of names).
Is it clear? Do you have any ideas do it better?
The most important is that user must choose an option from the list (can’t type new name) and list is veeeery long.
I will do your data entries 100% accurately. I have experience in working in large batches of data entries. I can extract data from pdf, online doc or images to be extracted and entered into any application like Word or Excel. Services provided: 1.PDF to Word 2.Data scrape from website to Word 3.Image to Word 4.Image table to Excel 5.PDF table to Excel 6.Copy paste to Word / Excel Regard: USMAN
Category: Data Entry
This is basically a code development question that deals with frequent changes to one or more java files among many thousands.
A few years ago, as a software intern, I was given a large java / jsp code base and my mentor set me up using intellij IDE to build a WAR file for a web application (a student portal). I would then copy the War file to the tomcat container and launch it. I don’t recall, but I recall to the effect that all I had to do was to press that one button (build?) to generate the WAR file. It took a long while to generate this War file. Intellij used an Ant file to build the war file. The team later migrated to Maven, but that’s when my term was up.
Software development was very slow because I wanted to write code, compile and test code often in the web browser. There wasn’t unit code testing.
So here’s my question:
What does the industry do (for example FANG) where if I wanted to just modify ONE java file, how would I go about in a much faster fashion to compile that one file instead of recompiling every file to create a WAR file and then put it into the tomcat container and run it?
I just want to minimize the amount of time from inserting code to checking out how it behaves.
I tried to write my own batch script but taking into account all the thousands of java dependency was a nightmare. this this post. This is my first post since many years ago. I appreciate for someone to shed light on how I can approach this.
I wasn’t happy with using this IDE solution because there was a very senior software engineer who didn’t use an ide, isntead, used emacs on his mac. He got things done fast, but I didn’t understand how he would go modify a few files and got it to run on tomcat. I can’t imagine that he’d recompile everything. So please share with me some work flow using IDE and also preferably not to use an IDE.
I did my research and google but couldn’t find an examnple tutorial of how such would be done. One reason I ask this question is because if I ever get a software engineering job, I would not want to modify one file and take forever to recompile everything.. it just doesn’t sound like the way to go. This wasn’t taught in school.
If there is a git project that deals with the above.. I’d be grateful if someone can send me an exampile git project that deals with the above webapp scenario.
Thank you. Gordon
I’m building an application (both client and server side code) where probably thousands of devices will be deployed in field and these devices will be connected to internet via cellular connection. Each device will have a SIM in it. Now, these devices are embedded devices with very limited capability in terms of software options. The only comm protocol support by these are TCP, UDP and HTTP/s(server and client). You can’t even add any external libraries to support more protocols like MQTT etc. The development environment is very closed and only allows very specific and limited type of external libraries to be added.
So I’ve decided not to use UDP and TCP because they’re very very low-level and the development time and hence cost will be much high. So, I’m left with HTTP. The application requires to poll some data from the devices at defined intervals and sometimes on-demand as well(when the user requests). This suggests to put the server on devices, and just use GET/POST etc. to acquire the data periodically or whenever the user demands. But since these devices will be put in very remote and geographically different locations, we’ll have to assign a fixed IP(or a domain name) to each device so that the server can send requests. Now this architecture doesn’t seem very good because managing the IPs will be a huge headache.
The last option I have is to put a HTTP server on the actual physical server, and have each device ping periodically(like every 5 or 10 seconds ?) via POST request, for example, and as a response to this post request the server can ask for whatever it wants in json format, be it new data or any configuration changes etc. This ping will also act as a heartbeat. On the server side, due to some limitations I’m bound to use python. So I’m thinking of using Flask with some webserver like Nginx etc.
I need some suggestions if this is the only way I can implement this or are there any better solutions/architecture ? Also, will the flask+nginx be able to handle thousands of requests ? Or are there better options to implement the webserver in python ?