Denial of Service attacks and other in Java

I am talking now about the question whether it makes sense to incorporate mechanisms into servlets to attacks (eg DDoS, but not only) ward. If I put an app in the cloud only option was to make software-what. Hardware support, as suggested, is in cloud computing via a third party there or not (I have to examine times). Regarding the hardware options I would like to widen the debate. Does it make sense :

• Just to set the hardware (perhaps because software can not detect?).
• Both secure hardware and software (perhaps because software only knows the rules, which followed the legitimate client and thus a violation is seen as an attack).

How high is the risk of an attack at all? Are there any statistics??

Quote:

---

Does it make sense :

• Just to set the hardware (perhaps because software can not detect?).
• Both secure hardware and software (perhaps because software only knows the rules, which followed the legitimate client and thus a violation is seen as an attack).

---

So basically I would consider the defense against DoS attacks rather than infrastructure task. If an application detects attacks, then yes the safety of another would take place. One would have to secure each application separately and maintain security as well. A secure infrastructure would protect all applications.

Quote:

---

Basically I would consider the defense against DoS attacks rather than infrastructure task. If an application detects attacks, then yes the safety of another would take place.

---

So I would not dare to make a safe application, but my infrastructure to make as a whole safer. If an application detects an attack (your case b), should be ensured that the software passes the information to make the whole infrastructure can benefit from it.

Absolutely! DoS / DDoS far away as possible should be captured from the real servers. It is usually not enough to add something static. If an attack takes place helps in many cases the first systems (FW, gateway, router, etc) to block the appropriate network segments. Otherwise also on the application side, a certain amount of protection to be installed. Here too, the application should operate a useful resource management, such as thread pools, meaningful queue lengths, etc. That helps, especially if the application actually "real" work (calculations, etc) makes.

I noticed that many of the term (D) DoS take closer than I do. I look at all forms of DoS attacks that can disrupt IT operations. My primary concern is to make a special application in software more secure and less vulnerable. Many understand this particular application under DoS attacks such as SYN flooding nonspecific. There are hardware and firewalls. I used that as a system architect for the application of little influence, which must be paid by the IT infrastructure, which many have already pointed out.

Attacks can specifically relate to an application. For example, regular calls to the website / web applications to be adjusted in order to overload the application itself. I doubt that hardware or firewall can detect such an attack, because they lack the knowledge of an improper call to a permissible distinction .. What I have now provided, that all calls are considered very early for errors that indicate attacks. The calls are dropped if something is not used correctly. Each of these errors is evaluated as the attempted attack. If repeated errors that address is blocked.