Packet Classification: Problem and Resolution

Packet Classification: Problem and Resolution INTRODUCTION This chapter covers the introduction to packet classification. Problems for packet classification, objectives to overcome the problem involved in packet classification, motivation to do the project on packet classification and also the organization of the project. The development of the internet grows for every year, because of the easy access of the internet. The gain of the internet can be obtained through the smartphones, netbooks, notebooks. For processing the packets, network processor is used, and it will carry out the tasks as to convert the packets into fragments, reassembling these fragments, forwarding, encryption and packet classification. Due to increased line rates, pressure is increased on line rates and it in turn pressure on network processor. The pressure can be relieved in two ways: By inserting more processing cores and it increases power consumption. Increasing the clock speed, but it creates difficulty due to the physical limitation in the silicon. So that it can be relieved in two different ways: Insert the clock gating, which reduces the power consumption. Insert the buffer, to form the pipelining and it also increases the speed. 1.1  PROBLEM DEFINITION Network processors are getting more strain, due to more use of internet and the strain needs to be reduce. To give the security for the network packets. To minimize the power required for packet classification. To achieve high speed and also high throughput for packet classification. 1.2  OBJECTIVES Understood the concept of hypercut algorithm for packet classification and also analyzed the flow chart for packet classification. Understood the verilog code and Xilinx tool. Wrote the verilog code for hypercut algorithm. The simulation results for hypercut algorithm is verified in Xilinx tool. Clock gating circuit is inserted in the architecture of the classifier, it reduces the power consumption. Pipelining concept is used in proposed architecture of the classifier and the simulation results are verified in Xilinx. 1.3  MOTIVATION The network traffic is doubling for each six to nine months. Also traditional algorithms are not supporting the increasing network traffic on core and edge devices. Large number of rulesets: Due to increased access of the network, the more services need to be implement in network device, so that more number of rules are needed. It creates the difficulty for classifying the packets. Flexibility: Traditional algorithms are particularly designed for IPV4, so that novel solutions are required to manage both IPV4 and IPV6 addresses. Scalability: As the network services are increasing, there is requirement to add or delete the rules. So that, good scalability is required for packet classification. 1.4  Organization of the thesis The thesis contains 6 chapters Chapter 1, it will covers the introduction of packet classification, problems involved in packet classification, objective to the packet classification and motivation for choosing the packet classification. Chapter 2, it will covers the basics of existing packet classification and also the basics of proposed packet classification. Chapter 3, it will covers the method used to do the packet classification, proposed architecture and also it tells how proposed architecture is better compared to previous algorithms. Chapter 4, it discusses the simulation the simulation results obtained for existing and proposed architecture of the classifier. Chapter 5, it covers the conclusion and future scope of the project. Chapter 6, It lists the reference papers used for literature review of the packet classification. Chapter 2 BASICS It covers the basics of packet classification. It also explains the structure of packet header, brief introduction to internet, the mode of information transmission through the internet, OSI layers, the type of matching, software and hardware implementation of packet classification , Clock gating and pipelining are also discussed. It also covers the review of different packet classification algorithms, by reading this the user can select the algorithm, which is best suit for his application. The internet is a global system. It is consists of inter connected computer networks, which uses the protocols(TCP/IP), to match several billion devices all over the world. It is also termed as networks of network. Access of internet is a process of connecting mobile devices, computers and computer terminals to the internet. Internet access will enables the users to access the internet services such as email and world wide web. Using various technologies, internet service providers will access the internet. A packet is a formatted unit of data, which is carried by the packet switched network. By formatting of the data, the bandwidth of communication medium can be increased. The structure of the packet contains the two varieties of data Control information User data 2.1  Control information This will provide the information, on where to send the data. Example, It provides the source and destination IP addresses, sequencing information and error codes. Fig 2.1 shows the structure of the packet Fig 2.1 structure of the packet 2.2  Payload The maximum size of the packet is 64 K bytes. The payload of the packet is variable. Example IPV4 typically adds the 20 bytes of payload to every packet. The packet is passed through the network using three devices such as hub, switch and router. 2.3  The modes of information transmission through the internet 2.3.1  Hub Hub is a central device, for which all other devices are connected. It is called the star system. It is very simple, when any device sends the data, it will send the data to all other devices and all other devices needs to decide whether the data is belonging to them, if it is not belonging to them, they will ignore it. It will present in physical layer. Fig 2.3.1 shows the structure of the hub. . Fig 2.3.1 shows the structure of the hub 2.3.2  Switch The switch is smarter compared to hub. First it creates the table, which records the IP/MAC addresses of the devices(PC’s) connected together. At the start, when any device sends the data, that time switch will not be knowing the destination IP addresses. So it will forwards the packets to all other devices, which are connected to it and it also records the IP address of the device. Next when packet belonging to those destination IP addresses comes, it will directly forward the packets to destination devices It is present in data link layer of OSI layer. Fig 2.3.2 shows the structure of the switch. 2.3.3  Router It is the smartest device compared to hub and switch. The router will record the address of all the devices which are connected to it. The router will read the information present in packet header and it will decide , where the packet needs to be sent and how to process the packet. It provides the security. While in hub, switch the destination IP address is known, hacker may hack the destination device, it will consider both destination and source IP address of the devices and it will decide whether the source device is hacker or not. If it is hacker, it will deny the packet. Fig 2.3.3 shows the structure of router. Fig 2.3.3 shows the structure of router. 2.4 OSI LAYERS It consist of seven layer Application layer Presentation layer Session layer Transport layer Network layer Data link layer Physical layer The Fig 2.4 for OSI model is shown below as Fig 2.4 OSI layers Application layer: This layer will provide the interface to application programmes. Presentation layer: In this layer , it converts the data from system specific format to the format which is suitable for application. It also provides encryption and compression. Session layer: Which facilitates the starting, managing and ending of connection between the two nodes. Ex: For a video session, it will synchronize the related stream of data such as audio and video. Transport layer: It will break the data into segments It will decide about how much information can be sent to email server and how much information can be received back. The responsibility of transport layer are: Reliability Flow control Fragmentation Network layer: Here the segments are broken into packets by adding the source and destination IP address to them. Next the packets are sent to data link layer. Here router is working in this layer. Data link layer: In this layer, the packets are broken into the frames, which are created for the specific network. The frames are assigned the address of two nodes, the data is moving in between. Physical layer: The frames given by the data link layer are converted into bits in physical medium. UDP: User datagram protocol It is light weight and connectionless. Advantages: The packet size is small. UDP header- 8 bytes TCP header- 20 bytes There is no requirement to create and maintain the connection. It has more control over the data Disadvantages: It does not provide error recovery. It does not compensate for lost data packets. Packets can arrive at out of order, so that data loses meaning. There is no control of congestion. Transmission control Protocol: It is reliable and connection based. Advantages: It delivers the acknowledgements. It provides retransmission. It provides in order delivery. It will delays the transmission when the network is busy. It provides error recovery. Disadvantages: It has bigger header. It doesn’t always get sent out quickly. It is the side effect of congestion. It has bigger overhead. UDP is message oriented It sends the data in distinct chunks. Ex: Email. For multimedia applications, UDP is used, because of these reasons as: It has less overhead. Data loss can be masked. UDP is used in small transmission. It is also used in bandwidth intensive applications, that tolerate packet loss. TCP is stream oriented It can be used in continuous flow of data. Ex: Phone conversation. For text communication, TCP is better. Ex: File transfers, Remote access. TCP is used when delivery acknowledgement are needed. In physical layer, information is transmitted in bit stream using hub. In data link layer information is transmitted in frames using switches. In network layer information is transmitted in packets using router. A router is a device that forwards the packet. A router is connected between two networks namely LAN’s or WAN’s. network processors are specialized CPU, which is optimized to support the implementation of network protocols at maximum speed. The function of network processor is to carry out the tasks such as packet separation, reassembly, encryption and classification. Packet classification is the process of categorizing the packets into flows in internet router. Packet will be classified in network layer. Packet has five fields as shown in fig Source IP address: It indicates the IP address of the sender . Destination IP address: It indicates the IP address of the destination. Source port: It indicates the port number of sender. Destination port: It indicates the port number of destination. Protocol: Which specifies the type of transport packet being carried. The incoming packet to router will matches the specific rule if the distinct field in the packet will match corresponding field in the rule. There are three matches Exact match: The values present in rule field header are same as the values present in packet header. Prefix match: The values of rule field header are prefix for header fields of the packet. Range match: The packet header field values must be lie in the range which is specified by the rule. 2.5  The types of packet classification algorithms Packet classification algorithm can be implemented in two major types Software based Hardware based 2.5.1  Software Implementation This can be used with general purpose processor and network processor. The software based algorithm can be divided into two types as Field dependent Algorithm Field independent Algorithm Field Independent Algorithm: For each field in the rule, these algorithms will build the index table separately. Ex: RFC Field dependent Algorithm: In these algorithm, the fields of the rule will be matched in dependent manner and there is no need to group the result in final stage. The memory requirement for these algorithms is less than field independent algorithms. Ex: Hypercut, Hicut 2.5.2  Hardware based implementation This is used with ASIC or with FPGA. This implementation is used with internet routers for the high speed that supports to handle the packet. The reasons to use software implementation Programmability: ASIC architectures has small Programming capacity, Because ASICs have special design. Special chips: To accelerate the packet processing speed, special chips called TCAMS are used. The proposed algorithm uses clock gating circuit to reduce power consumption and pipelining to increase the speed. 2.6  Clock gating Clock gating is a technique, which is used in synchronous circuits to minimize the power consumption. This technique is used to prune the clock, it disables the port of the circuitary, so that flip flops present in the circuitry will not switch the states. When switching is absent, the dynamic power consumption is reduced, but the leakage currents are present. Clock gating works by taking the enable signal of the circuitry, so that flip flops or devices present in latches will not switch the states, so that switching power reduces. So it is necessary to have enable conditions in order to get benefit from clock gating. The clock gating saves the power. Clock gating can be added in two ways: By writing the RTL code, the synthesis tool automatically translates the RTL code into clock gating logic. In order to gate the clock of specific modules or registers manually clock gating circuit can be inserted by instantiating library the specific ICG cells. Using automated the clock gating tools, clock gating is inserted in semi automatic fashion. These tools will insert ICG cells to RTL code or directly add the enable conditions to RTL code. 2.7  Pipelining It is group of data processing elements, which are connected in series, so that output of one element is the input to next element. We build a pipeline by dividing the complex operation into simple operation. Here instead of taking bulk thing and executing it, the bulk thing is break up into smaller pieces and process it one after another. For example Consider a calculation c= log(|a+b|), which consist of three operations, which are shown in fig 2.7. Fig 2.7 Pipelining example Add a and b to get a+b, it takes 40ns. Take the magnitude, we get as |a+b|, it takes 35 ns. Take the log we get as log(|a+b|), it takes 60ns. Consider a situation when we need to carry out for 100 such pairs. Without pipelining , it would take a total of 100*135= 13500ns. By realization, it is found that it is whole sequential process. Let the values evaluated to be a1 to a100 and we need to add values to be b1 to b100. In first evaluation, ( a1+b1)is calculated, In next evaluation, |a1+b1|,(a2+b2) is calculated, in third evaluation log|a1+b1|,|a2+b2|, ( a3+b3) is evaluated. After the first output data that is log|(a1+b1)|, the subsequent outputs are log|(a2+b2)|, log|(a3+ b3)| will now start arriving at a gap of 60ns . All the 100 inputs can be applied in 199*60=5940ns and the total time taken to evaluate 100 data will be 5940+180= 6120ns. This time is half compared without pipelining. This process of evaluation is called pipelinlng. 2.8 Literature review Algorithms are classified in 4 classes: Basic structures Geometry based Heuristics Hardware based 2.8.1  Basic structures a. Linear search: This algorithm[1], is very simple. It contains all the rules. Here each packet is matched opposite to all the rules until the corresponding fields of the packet should match to the rule. Although, it is simple, it is not widely used. Because, it takes the large time for matching with the rule. Consider N is the number of rules, â€Å"the worst case space and time complexity is O(N),where O is the order and N is the number of rules. Fig below shows the linear structure. Fig 2.8.1.a Linear search algorithm b. Hierarchial trie: It is an extension part of the binary trie. By using the individual bits of the search key, the branches of the trie can be traversed. In the d dimensional hierarchial trie[2], first bulid the one dimensional hierarchial trie which is called F1 trie. Foe each prefix P in the F1 trie, there is a recursively (d-1) dimensional hierarchial tries are present(Tp). For example, if the data structure is 2 dimensional the only one F1 trie is present. Hierarchial tries are also termed as multilevel tries or backtracking tries or tries of trie. Womens Roles in Anime: The Reflection of Japanese Society Womens Roles in Anime: The Reflection of Japanese Society What do the portrayals of women in Japanese animation reveal about the role of women in Japanese society today? The role of women in Japanese culture is an important topic amongst when discussing contemporary Japan. While it is important to consider traditional methods of analysis that come from statistics and surveys, you also need to look at how popular culture defines the modern Japanese woman. Japanese animation can provide a tool to analysis of the role of women in Japanese society. Through looking at women and they are portrayed in Japanese animation over the decade, we can begin to see the role of women in contemporary Japanese society. The purpose of this discussion is to determine a connection between Japanese animation (a form of popular culture) and the role of women (whether this role is shifting or remaining conservative) in contemporary Japanese society. In order to understand the link between Japanese animation and society, it is important to consider the meaning of these terms and how they relate to each other. Martinez defines popular culture as the culture of the masses; which does apply to the Japanese culture. Nevertheless, other opinions raise the point to how popular culture is a problematic term when translating it into the Japanese context. According to Hidetoshi Kato, a Japanese scholar, the term popular culture as taishu bunka; translates into mass culture. However, Martinezs definition is a middle of the road definition for the terms offered by Kato. In addition, Martinez defines the anthropology of popular culture as the study of the interaction between the apparently separate realms of the material and the symbolic. Other anthropologist such as Roger Buckley agrees with Martinez, stating that popular culture should tell us something about contemporary Japanese behavior. The images presented by Japanese animation including its related cousin, manga, or Japanese comics have helped shape the identity of modern Japanese women. Japanese animation, or anime, is useful when studying the role of women in society. Popular culture serves to reflect and instigate change in Japanese society, by observing the changes and themes in anime, these trends are identifiable. Popular culture also helps one to understand the dynamics of Japanese society and culture. While the behaviors identified are not Japanese, such as conformity, loyalty and deference8, there are certain subtle undertones that make themes and characters that make them Japanese. Lastly, Japanese animation industry caters for both sexes across a wide age group. This is important, because unlike the animation industry in the United States whereupon childrens programming is largely aimed at young boys. Japanese animation makes it possible to gain an insight into the role of women in society because both sexes views are represented. This provides a foundation for observing how the role of women in Japanese society can be studied through examining their role in Japanese animation. Japanese animation is interesting to study when observing the roles of gender because of the way it written for diverse tastes and it differs from Western animation. Anime has often been described as edgy, provocative, and documentary-like; these industry buzzwords describe anime as a different expression of animation, when it is not. In reality, the Japanese have embraced animation as an expression that goes beyond the standard set by Disney. Anime includes animation catered for all age groups. Anime content ranges from superb works to trashy soap operas or pornography; however, they play an important role in Japans popular culture by providing a legacy of past ideals. Respected writer Frederik Schodt has split the anime available into two distinct halves as defined by the gender, Boys comic [shonen] anime carefully balance suspense with humor, dramatic stories of sports, adventure, ghosts, science fiction, and school life. Girls comic anime [shojo] also strive for balance but are distinguished by their tales of idealized love. Given the range to choose from, it is important to analyze a number of sources to see how anime reflects the position of women in Japanese society, but to examine them with an open mind. Reaction to anime has been filled with distaste. Film critics have both embraced and criticized Japanese animation. Mamorum Oshiis theatrical adaption of Shirow Masamunes Ghost in the Shell, regarded as a movie that questions what it is that makes us human by many, has been lauded as a spectacular-looking Japanese animated film, but like so many of its kind, involves a confusing narrative and peculiar metaphysics that reduce interest16. To avoid generalizations it is important to look at works from a several genres to how anime portrays Japanese women in society. To observe the role of women as portrayed in anime with a female-specific target audience, Komodo no Omocha (1996) will be used to analyzed. To see how women are being represented in anime that is targeted at a male demographic, Dragon Ball Z (199?) will form the basis of analysis. Finally, a recent work from Japans most respected and successful animation studio today, Studio Ghibli, will be analyzed Mononoke Hime (1997). The importance of studying Studio Ghiblis work is to see how women are portrayed by an animation studio whose works reach a very wide audience, that typically cross the gender and age boundaries. While this analysis will not prove to be the final answer, it will prove an interesting investigation when considering what images of femininity have been projected onto Japanese society over the past decade. Kodomo no Omocha was a popular series that aired on Japanese television in the mid 1990 is, compared to many shojo anime, Kodomo no Omocha is radical, though it is by far not the only one. The story is a comedy fused with typical shojo elements love interests, and pretty artwork. The female lead is a young girl named Sana, who lives with her eccentric, but loving, mother and Rei, her personal manager. Unlike other female characters that will be analyzed, Sana is loud, outspoken and is not intimidated by anyone. She is works hard to do her best at her job (she is a famous star of a childrens TV program), friends and family. While she exhibits traditionally female qualities, she is not the stereotypical wholly subservient female character. Even though she is only a child, she stresses the importance of her job. Her mother is another strong female role within the anime. She is intelligent, in control and the men are more likely to be subservient to her than anything else is. Other fema les in the show lean towards conservative Japan, such as when Sanas female teacher cries and runs to another male teacher for assistance whenever the class acts up. The male characters are traditionally empowered, though like many a shojo anime, have soft sides to their personalities. The depiction of the two main female characters, Sana and her mother, offer an insight into the modern Japanese woman as having the capacity to be strong, yet gentle and compassionate. Dragon Ball Z offers an interesting insight into the portrayal of women in anime aimed at a male demographic. The reason this is an interesting series to look at is its popularity and the way it portrays gender roles. Women are portrayed in few roles in this anime. The female roles are dismally backwards. Dragon Ball Z has portrayed the role of women as subservient/secondary at their core. It has pointed out that fundamental beliefs about gender roles are difficult to change in any society, including Japans, The show is very male-centric, emphasizing strength, discipline and hard work as the key to being successful in the world. On the other hand, this is a martial arts anime, so this should not be a surprising aspect. One of the women who appear regularly is the protagonists female friend Bulma, an intelligent scientist. Bulma is a gifted inventor and often creates something that will aid her group of friends in whatever plight they are in. Bulma is portrayed as a confident, intelli gent woman who, while not of equal status, remains an important part of the team. However, when she is put in a position of danger, she falls into the damsel in distress stereotype all too common in male-orientated anime. She is rescued one way or another, but is often the least of the groups priorities she is often rescued as an after-thought, as opposed to being a genuine critical concern. This is a good example of what Eri Izawa determines as the Unequal Relationship genre, where women are second to men in a world dominated by patriarchy. Hayao Miyazaki and Isao Takahata founded studio Ghibli in 1985. The importance with studying work by Studio Ghibli is because their productions (which are theatrical) have such widespread appeal in Japanese society. A large number of Japans populace sees the stories that are created here. The films cross the divides of age and gender. The movie that is I am looking at is Mononoke Hime. It grossed approximately US$150 million at the box office, beaten only by Titanic. In Mononoke Hime, the story centers on three principle characters San, the Wolf Princess, Lady Eboshi and prince Ashitaka, the pillar between which these two women are placed. One of interesting strengths of this film is the way in which the opposing characters, San and Lady Eboshi, are portrayed. Both are strong, powerful women, sure of themselves and their world. Lady Eboshi is a superlative example of the new woman of Japan. She is in charge of a profitable ironworks that employs people regarded as beneath society (prostitutes and lepers), leading by compassion, returning dignity and purpose to their lives. San is the other extreme; she lives in the forest of the Deer God, raised by the great wolf gods who dwell there. San is a strong female character who follows a more natural way of life; as opposed to the entrepreneurial mindset of Lady Eboshi, San is dedicated to preserving the forest and the creatures that dwell within. Miyazaki is portraying the capacity of women to be strong, competent and successful. San represents the capacity of women to be strong in the dwelling of the traditional while Lady Eboshi is an example of how women can be successful and entrepreneurial in the face of modernization. From three examples, it is easy to see these reflections on the role of women are present in contemporary Japanese society. Martinez admits that it is difficult to discover what the true role of Japanese women in Japanese society are in a perceived country where men are still dominant [and] Japanese women are held to be gentle, submissive and beautiful. Historically, Japanese womens suffrage was achieved in 1945 and it is noted that while Japanese women may be able to find jobs, it is far more difficult to find careers, even today. However, public opinion is changing. There is clear that the role of women in Japan is no longer the traditional housewife/mother. Surveys by the Prime Ministers Office in 1987 and 1995 show a shift in public opinion of women, while in 1987 over 50% of the men surveyed agreed with the traditional role of women in Japan, by 1995 the percentage had dropped to 33%. The state of Japanese animation over the past decade reveals a relative parallel between the co ntent of Japanese animation and these results. The role of women in anime such as Kodomo no Omocha gives girls a role model unlike the traditionally submissive, quiet woman in the face of Sanas eccentric, enthusiastic nature that is rewarded by success. Studio Ghiblis films have portrayed women as equal, confident and able to take charge in their lives. Anime aimed at boys retain the portrayal of women as bystanders, whether it be the token damsel in distress or cheerleader. However, that is changing; Japanese animation aimed at the boys has begun to portray women beyond this traditional stereotype. Several productions such as Ranma 1/2 and Love Hina offered a cast of characters that serve to put women on equal ground as men in some cases, though there is still the re-enforcement of traditional gender roles. Japanese animation is an established form of pop culture. It is consumed by both sexes and across all age groups, and offers a series of representations of women. While it has become apparent through observing trends in Japanese animation and contemporary Japanese society is an apparent shift, however, it is important to identify that the traditional role of women will probably always manifest itself in pop culture. This is because the traditional role of women is an option, just as choosing a career or being entrepreneurial are choices. What is encouraging to see is that those choices are being represented, either directly or indirectly, in anime. It speaks well for the future of Japanese society as alternative roles of women are being portrayed for both sexes in the culture of the masses. Reference: Martinez, D.P. (1998). The Worlds of Japanese popular culture: gender, shifting boundaries [UK, Cambridge UP, p. 3]. Hidetoshi, K. (1989). Some Thoughts on Japanese popular culture. Richard Gid Powers Hidetoshi, K. (1989). Handbook of Japanese popular culture. Greenwood Press. Buckley, R. (1990). Japan today. Cambridge: UP. Schodt, F.L. (1983). Manga! manga! the world of japanese comics. USA: Kodansha. Izawa, E. (2001). Japanese manga and animation: gender relations in manga and anime. Retrieved from http://www.uncc.edu/~medmoto/3209/anime/gender.html (abridged) Robertson, J. (1998). Takarazuka: sexual politics and popular culture in modern japan. University of California. Dimensions of Japanese society: gender, margins and mainstream. Great Britain: Macmillan Press Ltd.