F = T ∇ Sτ

A phenomenon of document-oriented databases is quite interesting - many software developers face problems there this kind of databases is an excellent choice, but these developers don't use them and reinvent the wheel using relational or object-oriented databases. Difficult to say why it happens - because of ignorance, fear of performance problems or desire to reinvent own wheel, but this situation widely spread over the world. Fortunately, the common sense is prevailing, and NoSQL movement prove it. I have a serious experience with IBM Lotus Notes/Domino , and one of the most interesting features for me on first stages of its studying was saving application design in the documents. Thus the deployment of Lotus Notes database is incredibly easy - one just have to copy NSF-file to another location and it's ready for use (not always, but for trivial cases it's enough). Sometimes it can be a really useful feature, especially during prototyping, but not so many document-orie...

The short answer - yes . The reasons are below: There are well-knows systems (like YouTube) that shows Python suitability: YouTube Architecture The most performance problems are related with not a language speed, but with communication and databases speed. Let me show the example from my own experience: I've worked on project that process a huge amount of data (about inserting 100-200k records a day, and selecting from about 100-200 million records 10-20 times per second). The bottleneck was a database, not the language speed. All indexes were pretty complex, and selecting can take up to 30-90 seconds which was inappropriate. We had to change architecture: use data pool, caches, AMQP (with RabbitMQ server), and after that we don't know any problems at all with performance. Right tools means everything. For example, if you have to serve enormous web-requests, just use the suitable webserver like Tornado . Especially it is useful for serving Comet -based web applications. A...

I have published several articles on eellc.ru website. I hope you'll find them useful and interesting: Software System Requirements Questionnaire SQL in large-scale systems Architecture of large-scale systems Django speed, stability and security The website doesn't allow to post comments, so you can leave your remarks, suggestions and complaints here. Enjoy!

Backup is one of the most crucial operations for providing availability of the computer system. Although writing backup scripts is an easy task, there are few points with LVM and SELinux which are described below. LVM is a logical volume manager for the Linux kernel. One of the key features of LVM is generating snapshots for logical volumes. This allows the administrator to create a new block device which presents an exact copy of a logical volume, frozen at some point in time. It is very useful for backups - we can make backups without stopping volumes. Snapshots can be created with 'lvcreate' command with '-s' option. SELinux is an implementation of a flexible mandatory access control architecture in the Linux operating system. It uses file labelling with special security context. During a usual file archiving this context could be lost, so the administrator should use special archivers like 'star'. So, the administrator should use following procedure for a...