Intensive packet domain mining engine (IPDME): A high-speed pre-processor for network intrusion detection

Abstract

Purpose

The service denial threats are regularly regarded as tools for effortlessly triggering online-based services offline. Moreover, the present occurrences reveal that these threats are being constantly employed for masking other vulnerable threats like disseminating malware, information losses, wire scams and mining bitcoins (Sujithra et al., 2018; Boujnouni and Jedra, 2018). In some cases, service denials have been employed to cyberheist financial firms which sums around $100,000. Documentation from Neustar accounts that is about 70 percent of the financial sector are aware of the threat, and therefore, incidents result in few losses, more than 35 percent of service denial attempts are identified as malware soon after the threat is sent out (Divyavani and Dileep Kumar Reddy, 2018). Intensive packet analysis (IPA) explores the packet headers from Layers 2 to 4 along with the application information layer from Layers 5 to 7 for locating and evading vulnerable network-related threats. The networked systems could be simply contained by low potent service denial operations in case the supplies of the systems are minimized by the safety modules. The paper aims to discuss these issues.

Design/methodology/approach

The initial feature will be resolved using the IPDME by locating the standard precise header delimiters such as carriage return line feed equally locating the header names. For the designed IPDME, the time difficulties in locating the initial position of the header field within a packet with static time expenses of four cycles. For buffering packets, the framework functions at the speed of cables. Soon after locating the header position, the value of the field is mined linearly from the position. Mining all the field values consequentially resolves the forthcoming restrictions which could be increased by estimating various information bytes per cycle and omitting non-required information packets. In this way, the exploration space is minimized from the packet length to the length of the header. Because of the minimized mining time, the buffered packets could be operated at an increasing time.

Findings

Based on the assessments of IPDME against broadly employed SIP application layer function tools it discloses hardware offloading of IPDME it could minimize the loads on the essential system supplies of about 25 percent. The IPDME reveals that the acceleration of 22X– 75X as evaluated against PJSIP parser and SNORT SIP pre-processor. One IPDME portrays an acceleration of 4X–6X during 12 occurrences of SNORT parsers executing on 12 processors. The IPDME accomplishes 3X superior to 200 parallel occurrences of GPU speeded up processors. Additionally, the IPDME has very minimal latencies with 12X–1,010X minimal than GPUs. IPDME accomplishes minimal energy trails of nearly 0.75 W using two engines and for 15 engines it is 3.6 W, which is 22.5X–100X less as evaluated to the graphic-based GPU speeding up.

Originality/value

IPDME assures that the system pools are not fatigued on Layer 7 mining by transmitting straightforwardly based on network intrusions without branching into the operating systems. IPDME averts the latencies because of the memory accesses by sidestepping the operating system which essentially permits the scheme to function at wired speed. Based on the safety perception, IPDME ultimately enhances the performance of the safety systems employing them. The increased bandwidth of the IPDME assures that the IPA’s could function at their utmost bandwidth. The service time for the threat independent traffic is enhanced because of minimization over the comprehensive latencies over the path among the network intrusions and the related applications.

Citation

Velliangiri, S. (2019), "Intensive packet domain mining engine (IPDME): A high-speed pre-processor for network intrusion detection", International Journal of Intelligent Unmanned Systems, Vol. 7 No. 4, pp. 135-149. https://doi.org/10.1108/IJIUS-03-2019-0018