In a lot of applications, users need to be able to upload files. For example, most social networks allow users to upload pictures.

We will modify the web application to allow users to upload pictures when they are viewing details about a specific phone. After some time, other users will be able to browse a gallery of user-uploaded pictures for each of the phones.

To do that, we will use the file upload capabilities of PHP.

It is important to make sure that the uploaded file is of the proper kind and size.

MySQL supports transactions. This allows software developers to create application which preserve data integrity even in the event of (software or hardware) failures.

The work to do is to use transactions as provided in MySQL to protect all collections of queries which should be treated as being atomic in nature.

Another important feature of MySQL is stored procedures. These procedures exist in the database and, when called from a PHP application, can greatly increase performance. Furthermore, stored procedures allow multiple applications, written in PHP or not, to behave in a consistent manner.

The work to do is simplify the checkout of cart items using stored procedures.

HTTP is stateless. Consequently:

When a web server is required to customize the content of a web page for a user, the web application may have to track the user’s progress from page to page. A common solution is the use of HTTP cookies. Another method include server side sessions.

PHP supports both sessions and cookies and, as such, allow us to transform the web application into one which supports multiple users concurrently. In other words, the work to do is to transform the web application into one where multiple users can log in and order phones at the same time. We will use PHP sessions to implement those new functionalities.

Additionally, we would like to implement a “Show popular phone of the hour”. We will use PHP cookies (with appropriate expiration dates) to implement this functionality in such a way that queries to the database are minimal.

Since version 5.0 of MySQL was released in October 2005, a lot of new features have been added to it to make it standards-compliant, secure and efficient. In 2011, MySQL 5.5 is still an open source software and is freely downloadable for various platforms. In this blog entry, I will focus on performance in MySQL.

MySQL provides two important utility statements, EXPLAIN and PROCEDURE ANALYSE, which can be used to understand where performance bottlenecks are. EXPLAIN is used to obtain information about how MySQL executes a SELECT statement. PROCEDURE ANALYSE suggests optimal data types for each column that may help reduce table sizes. Also, as MySQL exposes its inner workings to the database administrator, one can go very far in optimising performance.

One simple way to increase performance is to create indexes as they generally allow queries to run more quickly. Care must be taken to only create indexes when really needed as indexes consume space. In some special case, indexes can, in fact, decrease performance.

In the real world, web applications are used by concurrent users and it is important to use transactions and proper locking mechanisms in order to preserve the integrity of data. Of course, it is also important to understand that transactions and locking can have a negative impact on performance.

Whenever the same sequence of SQL statements is done over and over again, there is an opportunity to use a stored procedure which is just a set of SQL statements stored in the database server itself. Consequently, clients don’t need to issue the individual statements but can refer to the stored procedure instead. As written in the MySQL documentation, stored procedures can provide improved performance because less information needs to be sent between the server and the client. The tradeoff is that this does increase the load on the database server because more of the work is done on the server side and less is done on the client (application) side.

One recently added feature of MySQL is partitioning. This allows different portions of a table to be stored as separate files in different locations on disk. The rule used for doing the partitioning is specified by the user. Some queries can run much more quickly when a table is partitioned in virtue of the fact that data satisfying a given WHERE clause can be in one (or a few of the) partition(s), which obviously excludes any remaining partitions from the search.

Finally, given the prevalence of storage systems with very high I/O rates compared to our classical hard disks (e.g. solid state devices), MySQL has an adjustable I/O rate which can be tweaked to maximise performance.

We will cover some of these interesting aspects of MySQL in our coming Web Development with PHP and MySQL training starting on 25 July and 8 August.

One thing is sure: MySQL is far from being a toy.

The Linux kernel is open source software and, consequently, its source code (millions of lines of C) is available.

Sometimes, it is important (and interesting) to compile either a specific kernel module (typically because of a bug correction) or to compile the whole kernel. Of course, it is technically possible to download the latest Linux kernel, compile it and install it on an enterprise Linux distribution like RHEL or CentOS but this is not recommended.

Instead, Red Hat and CentOS recommend to obtain the source code of the kernel which is already in use in those distributions, configure it, compile it and install the compiled kernel without removing the previous one(s).

When using CentOS (for example), the first step is to obtain the source code of the kernel and unpack it correctly.

Then, depending on user requirements, the system administrator can:

• either build one kernel module (e.g. to correct a bug) and use that corrected module in the current kernel
• or rebuild the whole kernel and install it without removing existing kernels.

Network File System

Network File System (NFS) allows remote hosts to mount file systems over a network and interact with those file systems as though they are mounted locally. NFS is configured by editing /etc/exports. Note that the defaults settings for NFS shares are ro (read-only), sync, wdelay and root_squash. It is possible to use all_squash and specify anonuid and anongid. Hostnames should follow a specific format.

The mount command (when used with the appropriate options) can be used to connect to an NFS share. Naturally, one can also edit /etc/fstab so that NFS mounts are done automatically at boot time.

Samba

Samba is an important software as it allows Linux computers to easily communicate with computers running Windows.

Samba can:

• Serve directory trees and printers to Linux, UNIX, and Windows clients
• Assist in network browsing (with or without NetBIOS)
• Authenticate Windows domain logins
• Provide Windows Internet Name Service (WINS) name server resolution
• Act as a Windows NT-style Primary Domain Controller (PDC)
• Act as a Backup Domain Controller (BDC) for a Samba-based PDC
• Act as an Active Directory domain member server
• Join a Windows NT/2000/2003/2008 PDC

What Samba cannot do:

• Act as a BDC for a Windows PDC (and vice versa)
• Act as an Active Directory domain controller

Samba can be configured by editing /etc/samba/smb.conf but most people prefer using a graphical tool such as SWAT (it is important to add FLAGS = IPv4 in its xinetd configuration). When configuring Samba, it is possible to create a share:

[shared]
comment = This is a Linux share
path = /home/shared/
valid users = john, jane
public = no
writable = yes
printable = no
create mask = 0644

Valid users (e.g. john and jane in the above example) need to be assigned SMB passwords which can be distinct from their Linux password. The smbpasswd command need to be used.

Shares can then be accessed using Nautilus and even mounted in Linux using the cifs filesystem type. Samba comes with a number of useful command line tools which system administrators should be aware of.

Dynamic Host Control Protocol (DHCP)

Linux can be used as a DHCP server.

Quoting the official Red Hat Deployment guide, “DHCP is useful for automatic configuration of client network interfaces. When configuring the client system, the administrator chooses DHCP instead of specifying an IP address, netmask, gateway, or DNS servers. The client retrieves this information from the DHCP server. DHCP is also useful if an administrator wants to change the IP addresses of a large number of systems. Instead of reconfiguring all the systems, he can just edit one DHCP configuration file on the server for the new set of IP addresses. If the DNS servers for an organization changes, the changes are made on the DHCP server, not on the DHCP clients. When the administrator restarts the network or reboots the clients, the changes will go into effect.”

The preferred way to configure a DHCP server is to start from a sample dhcpd.conf.

Domain Name Service (DNS)

Linux provides Bind, a full-fledged DNS server. Bind can operate as a caching nameserver as well as a DNS server which is fully authoritative for a zone (or many zones). The preferred way to configure Bind is to start from sample named.conf and zones.

Struts contains a number of powerful UI (User Interface) components. In essence, each UI component:

• Generate (complex) HTML
• Links HTML form fields to Java properties
• Does automatic type conversion and validation

Out of the tens of UI components, four are very powerful while being relatively easy to use. They are:

• select
• radio
• doubleselect
• updownselect

Work to do

Write a Struts web application which showcases the four above UI components. Look at the image above for some inspiration…

Bonus

Want autocompletion like in Google Search? Look no further than the autocompleter tag. Try it!