دانلود مقاله مرکز توجه شبکه تعریف شده با نرم افزار
چکیده
مقدمه
پیشینه
مطالعات مرتبط
شهود و نمای کلی SDN Spotlight
شناسایی اختلالات ارسال
شناسایی حلقه
تجزیه و تحلیل دقت تشخیص
نتایج تجربی و ارزیابی
نتیجه گیری
منابع
Abstract
Introduction
Background
Related work
Intuition and overview of SDN Spotlight
Forwarding anomaly detection
Loop detection
Detection accuracy analysis
Experimental results and evaluation
Conclusion
References
چکیده
عیب یابی در شبکه های مبتنی بر نرم افزار هنوز یک کار دست و پا گیر است که می تواند توجه انسان را تحت تأثیر قرار دهد. ناهنجاری‌های مختلف، مانند خرابی نصب، قوانین بی‌نظم و حلقه‌ها، حتی زمانی که جدیدترین روش‌های تشخیص استفاده می‌شوند، مورد توجه قرار نمی‌گیرند. در این مقاله، به موضوع تأیید خط‌مشی‌های SDN با بررسی فعال صفحه داده می‌پردازیم. SDN Spotlight به عنوان یک چارچوب تشخیص ناهنجاری ارائه می شود که سعی می کند خرابی های نصب، تضاد قوانین و حلقه ها را شناسایی کند. برخلاف کارهای اخیر، مانند Monocle و Pronto، SDN Spotlight زنجیره ای از قوانین را با استفاده از یک بسته کاوشگر تأیید می کند. این رویکرد همچنین تعداد قوانین نظارت را کاهش می دهد که تأثیر مستقیمی بر صرفه جویی در مصرف حافظه TCAM و به حداقل رساندن زمان تطبیق بسته ها دارد. SDN Spotlight دو مشکل را حل می کند: تأیید قانون و تأیید رفتار حمل و نقل. در چارچوب SDN Spotlight، ما دو رویکرد مختلف برای تشخیص ناهنجاری ارسال را معرفی می‌کنیم: Hedge-SDN Spotlight و Open-SDN Spotlight. علاوه بر این، ما یک الگوریتم تولید کاوشگر کارآمد و سریع را ابداع می کنیم که یک بسته کاوشگر واحد در هر زنجیره از قوانین تولید می کند. برخلاف سایر کارهای مرتبط، Hedge-SDN Spotlight هنگام شناسایی حلقه‌ها و خرابی‌های ارسال، مثبت و منفی کاذب را نشان نمی‌دهد. نتایج آزمایش نشان می‌دهد که SDN Spotlight بسیار سریع‌تر از روش SDNProbe و SDN traceroute است، در برخی موارد تا ۷ برابر سریع‌تر.
توجه! این متن ترجمه ماشینی بوده و توسط مترجمین ای ترجمه، ترجمه نشده است.
Abstract
Troubleshooting in SDN-based networks is still a cumbersome task that can overwhelm human attention. Various anomalies, such as installation failure, disordered rules, and loops, remain unnoticed even when the most recent detection methods are used. In this paper, we address the issue of verifying SDN policies by actively probing the data plane. SDN Spotlight is presented as an anomaly detection framework that tries to detect installation failures, rule conflicts, and loops. In contrast to recent work, such as Monocle and Pronto, SDN Spotlight verifies a chain of rules using a single probing packet. This approach also reduces the number of monitoring rules, which has a direct effect on saving TCAM memory usage and minimizing the packet matching time. SDN Spotlight addresses two problems: verifying rule installation and forwarding behavior verification. Within the SDN Spotlight framework, we introduce two different approaches for forwarding anomaly detection: Hedge-SDN Spotlight and Open-SDN Spotlight. Furthermore, we devise an efficient and fast probe generation algorithm that generates one single probing packet per chain of rules. As opposed to other related work, Hedge-SDN Spotlight does not yield false positives and false negatives when detecting loops and forwarding failures. The results of the experiment demonstrate that SDN Spotlight is much faster than the SDNProbe and SDN traceroute method, in some cases by a factor of up to seven times as fast.
Introduction
Nowadays, more devices are interconnected, data centers are expanding, businesses encourage bring your own device (BYOD) policies, the Internet of Things (IoT) is on the rise, and end-users span over multiple consumer devices.
As a result of this growth, traditional networks have become complex, hard to manage, prone to errors and logical flaws [1], resulting in time-consuming management and fault handling [2]. Computer network troubleshooting tends to be labor-intensive and requires intricate work. Network outages can seem almost unavoidable and are due to a number of reasons. Typical reasons for an outage include, but are not limited to, human errors such as faulty configurations or malicious activity, equipment malfunction, and force majeure events.
Conclusion
SDN policy verification and loop detection are important research topics. Although several policy violation detection approaches exist in the literature, most of them try to detect violations among the rules that are installed on the forwarding devices. Such approaches cannot detect failures of rule installation or physical port errors. Moreover, few recent studies have attempted to address the anomaly detection issue by using the common idea of installing a test rule per target rule. This leads to an excessive increase in the size of OpenFlow tables and, unfortunately, to wastage of already scarce TCAM memory, and an increase in the packet matching time. In this paper, we propose an efficient probe-based detection framework that uses a negligible number of test rules. Our method detects firmware OpenFlow rule installation, forwarding anomalies, and loops. Moreover, physical failures such as port failure or unstable links can be discovered by our suggested mechanism. We introduce the concept of catch-rules to hook the probing packet and forward it to the controller for analysis. Moreover, the proposed method supports all types of rules: forward, drop, and set. Our framework consists of two main methods: Hedge, and Open. The Hedge and Open methods have the same goal and can be seen as alternative methods that each have advantages and disadvantages. The Hedge approach is both free of false positives and false negatives, but this comes at the expense of an extra overhead in terms of installing more catch rules, while the Open approach has a more lightweight detection procedure, with a low likelihood of false-positive and false-negative results. The results of the experiment are very promising and show that this method can be used in a production environment. As future work, we would like to extend our method in order to not only detect the root cause of failures, but also to resolve them while automatically taking the network invariants into account. Moreover, we are working on optimizing the algorithm for parsing network nodes and using a binary search approach to detect the root cause of a failure.