humor race and misunderstanding Essay -- essays research papers

Humor, Race and Misunderstanding â€Å"Comedy is simply a funny way of being serious.† Peter Ustinov   Ã‚  Ã‚  Ã‚  Ã‚  In â€Å"On Being Crazy†, DuBois defines crazy as behavior based on illogical perceptions. Both the narrator and wayfarer agree that the wayfarer walking in the mud and becoming muddy rather than walk along side a dirty nigger who is clean is crazy. DuBois illustrates that society prefers to cut off its nose to spite its face instead of fixing the nose. Twain builds much of the Adventures of Huckleberry Finn in the same manner. His commentary regarding the social injustice is buried within the story’s humor. While it isn’t readily available on the surface, if one is so inclined to look, it can be easily found. In the Adventures of Huckleberry Finn, Twain satirizes the idiocy and cruelty of society in general. The language of the book, despite its accurate reflection of 19th century dialect, in and of itself is an illustration of misunderstanding. Nigger. Although it has acquired additional connotations in today’s society, at base, it still evokes an intense emotional and psychological response. Even now, in print or in speech, the word nigger causes one to cringe. Today, no one would tolerate being referred to as a nigger let alone called a nigger to their face. However, Jim and every other black, free or slave, endured that insult daily because it was a part of the social fabric. Even those with ‘good intentions’ used nigger as a utility word to desc...

Analysis on Pyramus and Thisbe Essay

â€Å"Leaving the person I love in danger and continuing to live on is the same as being dead.† -Hyuga, Natsume. Pyramus and Thisbe were a couple that preferred dying instead of living life without each other or considering that his or her beloved was in danger. Pyramus and Thisbe were deeply in love with each other, but could not be together for their parents had prohibited them from being with each other. The young lovers wanted to be together so badly and so they decided to run away together ignoring the demands of their parents. All bad choices are followed by consequences and that ´s what happened with these couple. As they were running away a horrible creature appeared to Thisbe making her run while her scarf fell. The creature crabbed the scarf and left it in the ground. As soon as Pyramus arrived to the place where they had to meet, he saw the scarf full of blood. He could not accept the tragedy and killed himself. While he committed suicide, Thisbe saw there was no more danger and decided to walk back to meet Pyramus. When she got there she saw Pyramus dead and grabbing on to her scarf. She could not accept what had happened and neither a life without him and so she killed herself. The myth of Pyramus and Thisbe is a tragic story but with important lessons. The tragedy of Pyramus and Thisbe would never have happened if they had taken good choices and if they had not supposed things they thought were true. Pyramus and Thisbe both decided to disobey their parents and ignore the consequences that would come after. They were so deeply in love with each other and thought the worse that could happened to them was that their parents would find them together. In that moment neither of them thought of the dangerous things that could happen to them or even about death. Both young lovers took a bad choice and as always their consequences were not far away. They both ran away from home and as a consequence Thisbe first encountered danger and then both had to do things that were against their own will. Both had to kill themselves because they could not imagine a life without each other. Life always has consequences for the choices and actions we take. If only they would have thought of the consequences of their actions, maybe they would have had a life together. The Bible says:  Ã¢â‚¬Å"Children, Obey your parents in the Lord, for this is right. Honor your father and your mother-which is the first commandment with a promise- that it may go well with you and that you may enjoy long life on the earth.† The lord promises reward to those who obey their parents and that’s what we have to do every day of our lives. If we take good decisions then there is no doubt that our consequences will be of reward. The myth of Pyramus and Thisbe became a tragic story not only because of the consequences of their actions but also because they supposed things that were not true. When Pyramus saw Thisbe ´s scarf in the floor with blood he supposed she was dead when really she was not. Instead of looking for her and starting to ask for help, he thought of the worse things that could happen to her beloved. He ended up thinking that the horrible beast ate her and with that in his mind he took his life away. If only he had waited for her to come back and not suppose of something that was not true, it would not have ended in such a tragedy. If Pyramus had not killed himself then Thisbe would not have killed herself too and they would have lived happily ever after. It’s a lesson that we have to learn from, it is better for us not to suppose about something we think is true. It is better to ask so we don ´t imagine things like with think they are. In conclusion, Pyramus and Thisbe both received a tragic ending because of the bad decisions they took. We both had to die with the idea that it was their fault their beloved died. Even though it may seem so tragic, it has two important lesson from which we have to learn. We have to see their errors and make better decisions so we do not fall into the same tragedies. Learning from their mistakes we can be better people and even take better decisions that will not give us bad consequences but instead rewards. When we are not sure that we are making good decisions we can always ask others or use the bible to help us see what is wrong, right and what will bring us a joyful life here on earth.

The Request For Proposal Process - 2846 Words

Cynthia Holden IS 578 The Request for Proposal Process Enterprises seeking a consulting firm or product vendor for management, strategy, outsourcing, software implementation, change management, or a host of other reasons issue a request for proposal (RFP) to elicit bids for services or products from vendors in the form of a responding proposal. The RFP process is a means of bringing structure and transparency to the procurement decision and the process allows both the client and potential vendors to vet each other to determine if they are suitable business partners. Before an organization issues the formal RFP to potential vendors, the organization completes several steps. Steps to the Formal RFP One of the most imperative steps†¦show more content†¦And the weight they will assign to each of the criteria. Drafting and Publishing the RFP Documenting the requirements for the engagement is essential, however the requirements need not be exhaustive as to limit the creativity of the respondents. The RFP team needs to take care that the RFP does not include the solution to the issue. In doing so, the team is essentially saying to the respondents, â€Å"We know what is best.† Consequently, some firms may conclude that the potential client is difficult to work with and not worth the effort of responding to the RFP. This limits the pool of respondents and the client may miss an opportunity to collaborate with a consulting firm or product vendor whose ideas and or products are beneficial to their organization’s project success. Requirements are just a portion of the RFP, it should also include, organizational background, short project description and project objectives: In my experience this is often the lengthiest section of the RFP, as it describes the characteristics that define a successful outcome of the project in the team’s estimation. It is worth noting that in general, specific, closed questions are easier to evaluate and score than open-ended ones. However, open-ended questions allow for creativity and may add to the evaluation time, as well as clarity on what the consulting firm or product has to offer. The RFP commonlyShow MoreRelatedPerforming A Cost Benefit Analysis1635 Words   |  7 Pagesstability are all reflected in vendor cost and would be reflected in the cost-benefit analysis and will aid in the vendor selection process. Request for information is the phase in the project where vendor information is requested. A request for information may include information such as rates, experience, and relevant reasons for the vendor’s application. A request for information or RIF is a good way to establish the vendor’s communication ability with potential clients. Electronic Health RecordsRead MoreProposal : A Proposal For Proposal1064 Words   |  5 Pagesbusiness, a request for proposal is a contract between two partners, or companies that one side offers the proposal while the other side will help to take care and mange that proposal to make it more successfully. This agreement is considered as a type of document which investors announce that funding is available for a particular project. Companies or investors have money, however the project is over their control because it requires a long and complex process. Therefore, a request for proposal is a needRead MoreThe Steps in the Acquisition Pr ocess for a Negotiated Best Value Source Selection1344 Words   |  6 PagesAcquisition is a process that requires teamwork with each individual and/or group working together to ensure that the customer is provided the greatest overall benefit in response to their requirements (best value). Acquisition planning should start once the need for supplies or services are identified. Integrated Product Teams (IPT) should be used when needed to help develop the acquisition strategy. The use of these teams will reduce false starts and delays resulting from unclear scopes ofRead MoreKey Issues and Concerns in the Development of a Proposal Essay1299 Words   |  6 PagesKey Issues and Concerns in the Development of a Proposal Key Issues and Concerns in the Development of a Proposal Introduction Status of the Opportunity The first steps taken after reviewing a Request for Proposal involves determining the status of the opportunity. (Johnson-Sheehan, 2008). In short one must determine if an opportunity or problem exists. The Request for Proposal under review indicates that Government officials are soliciting a vocational initiative that will be funded toRead MoreCorwin Corporation Case Analysis1005 Words   |  5 PagesCorporation is an internationally known rubber products manufacturer with a reputation for quality. Corwin’s management is conservative and favors expanding markets for existing product over new product development. The company receives frequent requests to manufacture specialty products. A strict management policy and a risk adverse culture results in a 90% no bid on specialty product inquiries. However, Corwin selected to respond to a bid from one of its customers. The project was a complete failureRead MoreProject Management Of The Project1246 Words   |  5 PagesThe propos al writing process is multifaceted, encompassing both the actual submission of a request for proposal as well as the management of the implementation of received funds. Through class lectures and assigned readings, we’ve discussed the various aspects that make up the process of grant writing. Through this essay, I will explain this process as it leads up the management of the project once funded. Furthermore, this essay will explain how to ensure that the proposal submitted addresses keyRead MoreAbstract. This Document Will Briefly Discuss Subjects And1361 Words   |  6 PagesAbstract This document will briefly discuss subjects and answer questions related to the development of a Request for Proposal (RFP). First, the purpose of the Management Requirements Section will be discussed and its importance to the RFP. As well, different methods will be reviewed that ensure suppliers provide buyers with the right resources, in addition to managing the risk associated with low quality resources. Further, the complexity of training logistics is discussed in comparison buyerRead MoreMock Request for Proposal1105 Words   |  5 PagesRequest for Proposals Ethan Binkley ENG/221 April 23, 2013 Request for Proposals (RFP) for: A a t Administrative Assistant Temps Microsoft Office Training Services Administrative Assistant Temps (AAT) 555 N. Business Park Avenue Nowhere NY 95952 Section 1: INTRODUCTION TO THE COMPANY 1.1 Introduction Administrative Assistant Temps (AAT) is an agency designed to fill short term, administrative and secretarial positions for companies around the globe. We have beenRead MoreRequest For Proposal Rfp ( Rfp )1738 Words   |  7 Pages REQUEST FOR PROPOSAL (RFP) PROJECT NAME COMPANY NAME STREET ADDRESS CITY, STATE ZIP CODE DATE TABLE OF CONTENTS 1. SUMMARY AND BACKGROUND 2 2. PROPOSAL GUIDELINES 2 3. PROJECT PURPOSE AND DESCRIPTION 3 4. PROJECT SCOPE 4 5. REQUEST FOR PROPOSAL AND PROJECT TIMELINE 5 6. BUDGET 5 7. BIDDER QUALIFICATIONS 6 8. PROPOSAL EVALUATION CRITERIA 6 1. SUMMARY AND BACKGROUND Center for Advance Research in Engineering (Care) is currently accepting proposals to Reverse Engineer the availableRead MoreCorwin Corporation Case Analysis997 Words   |  4 PagesCorporation is an internationally known rubber products manufacturer with a reputation for quality. Corwin’s management is conservative and favors expanding markets for existing product over new product development. The company receives frequent requests to manufacture specialty products. A strict management policy and a risk adverse culture results in a 90% no bid on specialty product inquiries. However, Corwin selected to respond to a bid from one of its customers. The project was a complete failure

Does Life Exist Elsewhere in the Cosmos

The search for life on other worlds has consumed our imaginations for decades. Humans feed on a constant supply of science fiction stories and movies such as  Star Wars, Star Trek, Close Encounters of the Third Kind, which all cheerfully suggest that they are out there. People find aliens and the possibility of alien life are fascinating topics and wondering if aliens have walked among us is a popular pastime. But, do they really exist out there? Its a good question. How the Search for Life Is Done These days, using advanced technology, scientists may be on the verge of discovering places where life not only exists but may well be thriving. Worlds with life on them may be all over the  Milky Way Galaxy. They could also be in our own solar system, in places that arent exactly like the life-friendly habitats that exist here on Earth. Its not just a search about life, however. Its also about finding places that are hospitable to life in all its many forms. Those forms may be like the life that exists on Earth, or they could be very different. Understanding the conditions in the galaxy that enable the chemicals of life to assemble together in just the right way.   Astronomers have found more than 5,000 exoplanets in the galaxy. These are worlds circling other stars. There are many more candidate worlds to be studied. How do they find them? Space-based telescopes such as the Kepler Space Telescope look for them using specialized instruments. Ground-based observers also look for extrasolar planets using very sensitive instruments attached to some of the worlds largest telescopes.   Once they find worlds, the next step for scientists is to figure out if they are habitable. That means, astronomers ask the question: can this planet support life? On some, conditions for life could be quite good. Some worlds, however, orbit too close to their star, or too far away. The best chances for finding life lie in the so-called habitable zones. These are regions around the parent star where liquid water, which is necessary for life, could exist. Of course, there are many other scientific questions to be answered in the search for life.   How Life is Made Before scientists can understand if life exists on a planet, its important to know how life arises.  One major sticking point in discussions of life elsewhere is the question of how it gets started. Scientists can manufacture cells in a laboratory, so how hard could it really be for life to spring up under the right conditions? The problem is that they are not actually building them from the raw materials. They take already living cells and replicate them. Thats not the same thing at all. There are a couple of facts to remember about creating life on a planet: Its NOT simple to do.  Even if biologists had all the right components, and could put them together under ideal conditions, we cant yet make even one living cell from scratch. It may very well be possible someday, but not now.Scientists dont really know how the first living cells formed. Sure they have some ideas, but no one has replicated the process in a lab.   What they do know are the basic chemical building blocks of life. The elements that formed life on our planet existed in the primordial cloud of gas and dust from which the Sun and planets arose. That would include the carbons, hydrocarbons, molecules, and other pieces and parts that make up life. The next big question is how it all came together on early Earth to form the first one-celled life forms. Theres not a complete answer to that one, yet. Scientists know conditions on early Earth were conducive to life: the right mix of elements was there. It was just a matter of time and mixing before the earliest one-celled animals came about. But, what was it that spurred all the right things in the right place to form life? Still unanswered. Yet, life on Earth — from the microbes to the humans and plants — is living proof that it is possible for life to form. So, if it happened here, it could happen elsewhere, right? In the vastness of the galaxy, there ​should exist another world with conditions for life to exist and upon that tiny orb life would have sprung up. Right? Probably. But no one knows for sure yet. How Rare is Life in our Galaxy? Given that the galaxy (and universe) for that matter, is rich with the basic elements that went into creating life, its very likely that yes, there are planets with life on them. Sure, some birth clouds are going to have slightly different mixes of elements, but in the main, if were looking for carbon-based life, theres a good chance its out there. Science fiction likes to talk about silicon-based life, and other forms not familiar to humans. Nothing rules that out. But, theres no convincing data showing the existence of any life out there. Not yet. Attempting to estimate the number of life forms in our galaxy is a bit like guessing the number of words in a book, without being told which book. Since there is a large disparity between, for instance, Goodnight Moon and Ulysses, it is safe to say that the person doing the guessing doesnt have enough information. That may seem a bit depressing, and its not the answer everybody wants. After all, humans LOVE science fiction universes where other life forms are teeming out there. Chances are, there is life out there. But, just not enough proof. And, that raises the question, if there IS life, how much of it is part of an advanced civilization? Thats important to think about because life could be as simple as a microbial population in an alien sea, or it could be a full-blown space-faring civilization. Or somewhere in between.   However, that doesnt mean there isnt any. And, scientists have devised thought experiments to figure out how many worlds might have life in the galaxy. Or the universe. From those experiments, theyve come up with a mathematical expression to give an idea about how rare (or not) other civilizations may be. Its called the Drake Equation and looks like this: N R*  Ã‚ · fp  · ne  Ã‚ ·fl  Ã‚ ·fi    · fc  Ã‚ · L. where N is the number you get if you multiply the following factors together: the average rate of star formation, the fraction of stars that have planets, the average number of planets that can support life, the fraction of those worlds that actually develop life, the fraction of those that have intelligent life, the fraction of civilizations that have communications technologies to make their presence known, and the length of time that theyve been releasing them.   Scientists plug numbers in for all these variables and come up with different answers depending on what numbers are used. It turns out there could be just ONE planet (ours) with life, or there could be tens of thousands of possible civilizations out there.   We Just Dont Know — Yet! So, where does this leave humans with an interest in life elsewhere? With a very simple, yet unsatisfying conclusion. Could life exist elsewhere in our galaxy? Absolutely. Are scientists certain of it? Not even close. Unfortunately, until humanity actually makes contact with a people not of this world, or at least begin to fully understand how life came to exist on this tiny blue rock, the questions about life elsewhere arent going to be answered. Its most likely that scientists will find evidence of life in our own solar system first, beyond Earth. But, that search requires more missions to other places, such as Mars, Europa, and Enceladus. That discovery may come about much faster than the discovery of life on worlds around other stars.   Edited by Carolyn Collins Petersen.

The Islamic State Of Iraq And Syria - 1542 Words

East Asia is globalizing at a rapid speed, and due to its large influence on the rest of the world it is important to analyze the progression occurring here. Currently, the war against terrorism is a growing concern and countries around the world have come together to meet for a consensus about the negativity surrounding terrorism, specifically a unification against ISIS. ISIS, which stands for the Islamic State of Iraq and Syria, is a terror organization that has claimed responsibilities for the recent bombings in Paris, Belgium and Pakistan (1). The Obama Administration has shown great attention and focus towards ending ISIS and creating a unification between allied nations. On November 15 2015, President Obama met with the presidents of†¦show more content†¦On a local scale, communities in East Asia are aware and understand the threat of ISIS. Early in 2015, 2 Japanese hostages were captured by ISIS and were later killed by the terror group (3). During this time the Japan ese government worked diligently with international alliances and the local communities to assure that anything that could be done, would be done. Unfortunately, the hostages were killed, but such a scenario presents the global-local nexus of Japan. The Japanese people were obviously alarmed by the situation, and the government provided them with up-to-date notifications of the situation and reassurance that their government was involved. This type of interaction transpired into Japan’s global interactions with President Obama and French president Francois Hollande (3). These nations voiced their opinion against the acts of ISIS and showed a consensus that such an organization needed to be ended. Unfortunately, many East Asian nations besides Japan have been affected by ISIS. In Indonesia there are about seven Islamist extremist groups that acknowledge their actions to ISIS. A recent bloody strike reported on January 14th in the capital’s downtown district left eight d ead, validating the presence of ISIS in East Asia’s door steps. A group in the Philippines recently pledged their allegiance to ISIS in a short clip. A Starbucks in Thailand was affected by two suicide bombers and four additional explosions occurred

Jane Eyres Self

Jane Eyre?s Self-Discovery Essay The novel Jane Eyre, by Charlotte Bronte consists of continuous journey through Janes life towards her final happiness and freedom. Janes physical journeys contribute significantly to plot development and to the idea that the novel is a journey through Janes life. Each journey causes her to experience new emotions and an eventual change of some kind. These actual journeys help Jane on her four figurative journeys, as each one allows her to reflect and grow. Jane makes her journey from Gateshead to Lowood at the age of ten, finally freeing her from her restrictive life with her aunt, who hates her. Jane resented her harsh treatment by her aunt. Mrs. Reeds attitude towards Jane highlights on of the main themes of the novel, the social class. Janes aunt sees Jane as inferior, who is less than a servant. Jane is glad to be leaving her cruel aunt and of having the chance of going to school. At Lowood she wins the friendship of everyone there, but her life is difficult because conditions are poor at the school. She has come to be respected by the teachers and students, largely due to the influence of her teacher, Miss Temple, who has taken a part as a mother, governess, and a companion. Jane has found in Miss temple what Mrs. Reed always denied her. Also at Lowood Jane confront another main theme of the novel, the natural violence, which is depicted by Bronte then typhus kills many of the students including Janes best friend, Helen Burns. This scene is especially important, because it makes Jane stronger, which is appropriate, as mentally strong people cope with violence in a more rational way. As Jane grows up and passes the age of eighteen, she advertises herself as a governess and is hired to a place called Thornfield. Although journeying into the completely unknown, Jane does not look back, only forward to her new life and her freedom at Thornfield. This particular journey marks a huge change in Janes life; its a fresh start for her. Another important journey Jane makes is from Gateshead back to Thornfield having visited her aunt Reed on her deathbed. By then Jane realizes that she loves Rochester. A key theme is raised here, Jane fierce desire to love and to be loved. She feels alone and isolated when she has no friends around her. This is a sharp contrast compared to other characters search for money and social position. These contrasting themes strengthened with every journey she makes. When returning to Thornfield Jane is unhappy, but keeps her promise to Mr. Rochester and his daughter. She believes at this point that Mr. Rochester is going to marry Blanche Ingram, and that she will have to leave Thornfield and never see Mr. Rochester again. However, Mr. Rochester offers his hand in marriage to Jane, but her happiness is short-lived after finding out that he is still married to Bertha. Although so many terrible things are happening to her, her spirit remains unbroken. Jane flees from Thronfield and Mr. Roches ter. Jane hearing Rochesters voice calling her prompts her final journey from St. John to Thornfield. Jane and Rochesters relationship blossoms once again, but differently than before. In the past, Jane felt like an inferior to Rochester because he was her employer and was wealthy. Jane now feels at perfect ease, Rochester has become a better man because of his disabilities. Ultimately, these four journeys mirror Janes four emotional journeys. She transforms from an immature child to an intelligent adult. Jane also changes from innocent and nave to mature and strong-willed person. All of her experiences teach her how to love and feel loved and to discover her true family roots.

Open Flow Essay Example

Open Flow Essay Load Balancing Surya Prateek Surampalli Information Technology Department, Southern Polytechnic State University [emailprotected] edu Abstract—in high-traffic Internet today, it is often desirable to have multiple servers that represent a single logical destination server to share the load. A typical configuration comprises multiple servers behind a load balancer that would determine which server would serve the request of a client. Such equipment is expensive, has a rigid set of rules, and is a single point of failure. In this paper, I propose an idea and design for an alternative load-balancing architecture with the help of an OpenFlow switch connected to a NOX controller that gains political flexibility, less expensive, and has the potential to be more robust to failure with future generations of switches I. Introduction In today’s increasingly internet-based cloud services, a client sends a request to URL or a logical server and receives a response from a potentially multiple servers acts as a logical address server. Google server is said to be the best example, the request is sent to server farm as soon as the client resolves the IP address from the URL [1]. Load balancers are expensive that acts as a reverse proxy and distributes network or application traffic across a number of servers. Load balancers are used to increase capacity (concurrent users) and reliability of applications. They improve the overall performance of applications by decreasing the burden on servers associated with managing and maintaining application and network sessions, as well as by performing application-specific tasks [1]. We will write a custom essay sample on Open Flow specifically for you for only $16.38 $13.9/page Order now We will write a custom essay sample on Open Flow specifically for you FOR ONLY $16.38 $13.9/page Hire Writer We will write a custom essay sample on Open Flow specifically for you FOR ONLY $16.38 $13.9/page Hire Writer Since load balancers are not basic equipment and run custom software, policies are rigid in their choices. Specific administrators are required and also the arbitrary policies are not possible to implement. Since running policy and the switch are connected it is reduced to a single point of failure [2]. The order of magnitude will cost less than a commercial load-balancer if architecture with an OpenFlow switch is implemented which is controlled by the commodity server and also provides flexibility for writing patterns which allow the controller to be applied arbitrary political [1]. If the next generation of OpenFlow switches has the capability of connecting to multiple controllers, there is a chance of making the system much robust to abortion by keeping the any server behind the which that acts as the controller [1]. II. Background A. Load Balancing Load balancing helps make networks more efficient. It distributes the processing and traffic evenly across a network, making sure no single device is overwhelmed [1]. Web servers, as in the example above, often use load balancing to evenly split the traffic load among several different servers. This allows them to use the available bandwidth more effectively, and therefore provides faster access to the websites they host [3]. Whether load balancing is done on a local network or a large Web server, it requires hardware or software that divides incoming traffic among the available servers. Networks that receive high amounts of traffic may even have one or more servers dedicated to balancing the load among the other servers and devices in the network. These servers are often called (not surprisingly) load balancers [1]. Load balancing can be performed using dedicated hardware devices such as load balancers or having intelligent DNS servers. A DNS server can redirect traffic data centre with a heavy load or redirect requests made by customers for a data centre that is less network stretches from clients. Many data centres use of expensive hardware load balancing equipment that makes in distributing the network traffic across multiple machines to avoid congestion on a server. A DNS server resolves a hostname to a single IP address where the client sends the request. To the outside world there is a logical address that resolves a host name [3]. This IP address is not associated with a single machine, but is the type of service a client request. DNS can resolve a host name to a load balancer within a data centre. But this could be avoided for safety reasons and to avoid attacks on the device. When a client request comes to the load balancer, the request is redirected according to the policy. B. OpenFlow Switch An OpenFlow switch is a software program or hardware device that forwards packets in a software-defined networking (SDN) environment. OpenFlow switches are either based on the OpenFlow protocol or compatible with it [1]. In a conventional switch, packet forwarding (the data plane) and high-level routing (the control plane) occur on the same device. In software-defined networking, the data plane is decoupled from the control plane. The data plane is still implemented in the switch itself but the control plane is implemented in software and a separate SDN controller makes high-level routing decisions. The switch and controller communicate by means of the OpenFlow protocol. The OpenFlow switch on the other hand uses an external controller called NOX to add rules into its flow table. C. NOX Controller NOX is a network control platform, that provides a high-level programmatic interface upon which network management and control applications can be built. In brevity, NOX is an OpenFlow controller [3]. Therefore, NOX applications mainly assert flow-level control of the network meaning that they determine how each flow is routed or not routed in the network. The OpenFlow switch is connected to the NOX controller and communicates over a secure channel using the OpenFlow protocol. The current design of OpenFlow only allows one NOX controller per switch. The NOX controller decides how packets of a new flow should be handled by the switch. When new flows arrive at the switch, the packet gets redirected to the NOX controller which then decides whether the switch should drop the packet or forward it to a machine connected to the switch. The NOX controller can also delete or modify existing flow entries in the switch. The NOX controller can execute modules that describe how a new flow should be handled. This provides us an interface to write C++ modules that dynamically add or delete routing rules into the switch and can use different policies for handling flows. D. Flow Table A flow table entry of an OpenFlow switch consists of a header fields, counters and actions. Each flow table entry stores Ethernet, IP and TCP/UDP header information. This information includes destination/source MAC and IP address and source/destination TCP/UDP port numbers. Each flow table entry also maintains a counter of number of packets, and bytes arrived per flow. A flow table entry can also have one or more action fields that describe how the switch will handle packets that match the flow entry. Some of the actions include sending the packet on all output ports, forwarding the packet on an output port of a particular machine and modifying packet headers (Ethernet, IP and TCP/UDP header). If a flow entry does not have any actions, then the switch drops all packets for the particular flow. Each Flow entry also has an expiration time after which the flow entry is deleted from the flow table. This expiration time is based on the number of seconds a flow was idle and the total amount the time (in seconds) the flow entry has been in the flow table. The NOX controller can chose a flow entry to exist permanently in the flow table or can set timers which delete the flow entry when the timer expires. III. Load-Balancer Design Load balancing architecture comprises an OpenFlow switch with a control device of NOX and server machines connected to output ports of the switch server. The OpenFlow switch uses an interface to connect to the Internet. Each server has a static IP address and NOX controller maintains a list of currently connected to the OpenFlow switch servers. Each server is running web server emulation on a well known port. [pic] Figure1. Load-balancer architecture using OpenFlow switch and NOX controller The hostname of server to IP address is resolved by each client and a request is sent to that IP address on the known port number. If you consider the above diagram, when a packet is reached to the switch from the client, the header information of the packet is compared with the entries of the flow table. If the header information of the packet corresponds to an inlet of the flow, the counter for the number of packets, the byte count is incremented, and the actions associated with the input of the flow are performed on the packet. If no match is found, the switch forwards the packet to NOX. NOX decides how the packet for this flow should be handled by the switch. NOX and then inserts a new article in the cash flow of the switch using the OpenFlow protocol. To achieve load-balancing features, the modules should be written in C++ that is executed by NOX controller. NOX should perform the function of handle () when a new flow arrives at the switch. This function sets the load balancing policy and adds new rules in the flow table of the switch. All client requests should be destined for the same IP address, then whatever the module is executed by NOX, should add rules for each flow which can modify the destination MAC and IP address of the packet with a server’s MAC and IP address. The switch will forward the packet to the server output port after modifying the packet header. When servers return a packet to the client, the module adds an entry flow that changes the source IP address with the IP address of the host that the client sends its request. So the client should always receive packets from the same IP address. If the client connection / server connection is closed or remains idle for 10 seconds, then the inactivity timer expires causing the input stream to be deleted from the cash flow of the switch. This allows input stream recycling Servers wait for a NOX to register and then report their current load on some schedule similar to the Listener Pattern. NOX in a separate thread listening on a UDP socket for heartbeats with reported by server loads and maintains a table with the current loads of all servers. When applying for a new stream is received, it chooses the server with the lowest and the load current increases to the low current server. This prevents flow of all flows routed to the same server as the server reports a new load. It also breaks ties by turning it into a round robin until the servers report their actual load heartbeat. Flow Algorithm Require: Flow, path 1: sourceHost = LocateSource(flow); 2: destinationHost = LocateDestination(flow); : layer = setToplayer(); 4: currentSwitch = LocateCurrentSwitch(); 5: direction = 1; //upward 6: path = null; //list of switches 7: return search (); This algorithm works as follows. When the OpenFlow controller receives a packet from a switch, it switches the control to the load balancer. Line 1 to 6 introduces the initialization for necessary variables. The load balancer ? rstly a nalyses the packet’s match information including the input port on the switch that receives the packet as well as the packet’s source address and destination address. Then it looks up those addresses using its knowledge about the network topology. Once the source and destination hosts are located, the load balancer calculates the top layer that the ? ow needs to access. We use the search direction ? ag. The ? ag has two values: 1 for upward and 0 for downward. It is initialized to 1. A path is created for saving a route grouped by a list of switches later. Line 7 calls search () that performs the search for paths recursively. In the method search (), It ? rstly adds current switch into path. It returns the path if current search reaches the bottom layer. It reverses the search direction if current search reaches the top layer. Then it calls a method 1: search () { 2: path. add(curSwitch); 4: if isBottomLayer(curSwitch) then 5: return path; 6: end if 7: if curSwitch. getLayer ( ) == layer then 8: direction = 0; //reverse 9: end if 10: links = findLinks(curSwitch, direction); 11: link = findWorstFitLink(links); 12: curSwitch = findNextSwitch(link); 13: return search (); 14:} that returns all links on current switch that are towards current search direction. Only one link is chosen by picking up the worst-? t link with maximum available bandwidth. And then the current switch object is updated. The method search () is called recursively layer by layer from the source to destination. At last the path will be return to the load balancer. The path information will be used for updating ? ow tables of those switches in the path. Flow Scheduling The Flow scheduling functionality works as follows. Each OpenFlow switch maintains its own ? ow table. Whenever any packet comes in, the switch checks the packet’s match information with the entries in its ? ow table. The packet’s match information includes ingressPort, etherType, srcMac, dstMac, vlanID, srcIP, dstIP, IP protocol, T CP/UDP srcPort, TCP/UDP dstPort. If it ? nds a match, it will send out the packet to the corresponding port. Otherwise it will encapsulate the packet in a PACKET IN message and send the message to the controller. As a module of the OpenFlow controller, the load balancer will handle the PACKET IN message. It ? nds a proper path by executing a search with the DLB algorithm described in Algorithm 1. The path is a list of switches from source to destination of the packet. Then the load balancer creates one FLOW MOD message for each switch in the path and sends it to the switch. This message will have the packet’s match information as well as a output port number on that switch. The output port number is directly calculated by the path and network topology. If one switch receives a FLOW MOD message, it will use it to update its ? ow table accordingly. Those packets buffered on ports of that switch may ? nd their matches in the updated ? ow table and be sent out. Otherwise the switch will repeat this process. IV. Future Work The OpenFlow specification includes an optional feature that would allow multiple NOXs to make active connections to the switch. In the case then of the NOX failing, another machine could assume the role of the NOX and continue routing traffic. Naturally the system would need to detect the failure, have a mechanism to remember any state associated with the current policy, and all servers would have to agree on who the new NOX was. These requirements naturally lend themselves to the Paxos consensus algorithm in which policy and leader elections can be held and preserved with provable progress [3]. We have implemented Paxos in another research project and could add it to our server implementation at the controller/signaler layer. As long as at least half of the nodes in the cluster stay up, state will be preserved and traffic should continue to flow. V. Conclusion It is possible to achieve similar functionality to a commercial load balancer switches using only physical commodities. The OpenFlow switch provides the flexibility to implement the arbitrary policy in software and politics separate the switch itself. Since the policy is decoupled from the switch, we can avoid the machine implementation of the policy of a single point of failure and the creation of a more robust system. References [1] OpenFlow Switch Specification. Version 0. 8. 9 (Wire Protocol 0x97). Current maintainer: Brandon Heller ([emailprotected] edu). December 2, 2008. [2] Web caching and Zipf-like distributions: evidence and implications. Breslau, L. Pei Cao Li Fan Phillips, G. Shenker, S. Xerox Palo Alto Res. Center, CA. INFOCOM 1999. [3] Paxos Made Simple. Leslie Lamport [4] M. Al-Fares, A. Loukissas, and A. Vahdat. A Scalable, Commodity Data Center Network Architecture. ACM SIGCOMM, 2008. [5] C. E. Leiserson. Fat-trees: Universal networks for hardware-ef? cient supercomputing. IEEE Transactions on Computers, 1985. [6] T. Benson, A. Anand, A. Akella, and M. Zhang. Understanding Datacenter Traf? c Characteristics. SIGCOMM WREN workshop, 2009. [7] HOPPS, C. Analysis of an Equal-Cost Multi-Path Algorithm. RFC 2992, IETF, 2000. [8] W. J. Dally and B. Towles. Principles and Practices of Interconnection Networks. Morgan Kaufmann Publisher, 2004. [9] S. Kandula, S. Sengupta, A. Greenberg, P. Patel and R. Chaiken. The Nature of Data Center Traf? c: Measurements Analysis. ACM IMC 2009. [10] N. McKeown, T. Anderson, H. Balakrishnan, G. Parulkar, L. Peterson, J. Rexford, S. Shenker, and J. Turner. OpenFlow: Enabling Innovation in Campus Networks. ACM SIGCOMM CCR, 2008. [11] R. N. Mysore, A. Pamporis, N. Farrington, N. Huang, P. Miri, S. Radhakrishnan, V. Subramanya, and A. Vahdat. PortLand: A Scalable, Fault-Tolerant Layer 2 Data Center Network Fabric. ACM SIGCOMM, 2009. [12] Beacon OpenFlow Controller https://OpenFlow. stanford. edu/display/Beacon/Home. [13] B. Lantz, B. Heller, and N. McKeown. A Network in a Laptop: Rapid Prototyping for Software-De? nded Networks. ACM SIGCOMM, 2010. [14] Y. Zhang, H. Kameda, S. L. Hung. Comparison of dynamic and static load-balancing strategies in heterogeneous distributed systems. Computers and Digital Techniques, IEE, 1997. [15] OpenFlow Switch Speci? cation, Version 1. 0. 0. http://www. OpenFlow. org/documents/OpenFlow-spec-v1. 0. 0. pdf. [16] N. Handigol, S. Seetharaman, M. Flajslik, N. McKeown, and R. Johari. Plug-n-Serve: Load-balancing web traf? c using OpenFlow. ACM SIGCOMM Demo, 2009. [17] R. Wang, D. Butnariu, J. Rexford. OpenFlow-Based Server Load Balancing Gone Wild. Hot ICE, 2011. [18] M. Koerner, O. Kao. Multiple service load-balancing with OpenFlow. IEEE HPSR, 2012. Figure 2: Load-Balancer block diagram architecture using OpenFlow switch and NOX controller. Open Flow Essay Example Open Flow Essay Load Balancing Surya Prateek Surampalli Information Technology Department, Southern Polytechnic State University [emailprotected] edu Abstract—in high-traffic Internet today, it is often desirable to have multiple servers that represent a single logical destination server to share the load. A typical configuration comprises multiple servers behind a load balancer that would determine which server would serve the request of a client. Such equipment is expensive, has a rigid set of rules, and is a single point of failure. In this paper, I propose an idea and design for an alternative load-balancing architecture with the help of an OpenFlow switch connected to a NOX controller that gains political flexibility, less expensive, and has the potential to be more robust to failure with future generations of switches I. Introduction In today’s increasingly internet-based cloud services, a client sends a request to URL or a logical server and receives a response from a potentially multiple servers acts as a logical address server. Google server is said to be the best example, the request is sent to server farm as soon as the client resolves the IP address from the URL [1]. Load balancers are expensive that acts as a reverse proxy and distributes network or application traffic across a number of servers. Load balancers are used to increase capacity (concurrent users) and reliability of applications. They improve the overall performance of applications by decreasing the burden on servers associated with managing and maintaining application and network sessions, as well as by performing application-specific tasks [1]. We will write a custom essay sample on Open Flow specifically for you for only $16.38 $13.9/page Order now We will write a custom essay sample on Open Flow specifically for you FOR ONLY $16.38 $13.9/page Hire Writer We will write a custom essay sample on Open Flow specifically for you FOR ONLY $16.38 $13.9/page Hire Writer Since load balancers are not basic equipment and run custom software, policies are rigid in their choices. Specific administrators are required and also the arbitrary policies are not possible to implement. Since running policy and the switch are connected it is reduced to a single point of failure [2]. The order of magnitude will cost less than a commercial load-balancer if architecture with an OpenFlow switch is implemented which is controlled by the commodity server and also provides flexibility for writing patterns which allow the controller to be applied arbitrary political [1]. If the next generation of OpenFlow switches has the capability of connecting to multiple controllers, there is a chance of making the system much robust to abortion by keeping the any server behind the which that acts as the controller [1]. II. Background A. Load Balancing Load balancing helps make networks more efficient. It distributes the processing and traffic evenly across a network, making sure no single device is overwhelmed [1]. Web servers, as in the example above, often use load balancing to evenly split the traffic load among several different servers. This allows them to use the available bandwidth more effectively, and therefore provides faster access to the websites they host [3]. Whether load balancing is done on a local network or a large Web server, it requires hardware or software that divides incoming traffic among the available servers. Networks that receive high amounts of traffic may even have one or more servers dedicated to balancing the load among the other servers and devices in the network. These servers are often called (not surprisingly) load balancers [1]. Load balancing can be performed using dedicated hardware devices such as load balancers or having intelligent DNS servers. A DNS server can redirect traffic data centre with a heavy load or redirect requests made by customers for a data centre that is less network stretches from clients. Many data centres use of expensive hardware load balancing equipment that makes in distributing the network traffic across multiple machines to avoid congestion on a server. A DNS server resolves a hostname to a single IP address where the client sends the request. To the outside world there is a logical address that resolves a host name [3]. This IP address is not associated with a single machine, but is the type of service a client request. DNS can resolve a host name to a load balancer within a data centre. But this could be avoided for safety reasons and to avoid attacks on the device. When a client request comes to the load balancer, the request is redirected according to the policy. B. OpenFlow Switch An OpenFlow switch is a software program or hardware device that forwards packets in a software-defined networking (SDN) environment. OpenFlow switches are either based on the OpenFlow protocol or compatible with it [1]. In a conventional switch, packet forwarding (the data plane) and high-level routing (the control plane) occur on the same device. In software-defined networking, the data plane is decoupled from the control plane. The data plane is still implemented in the switch itself but the control plane is implemented in software and a separate SDN controller makes high-level routing decisions. The switch and controller communicate by means of the OpenFlow protocol. The OpenFlow switch on the other hand uses an external controller called NOX to add rules into its flow table. C. NOX Controller NOX is a network control platform, that provides a high-level programmatic interface upon which network management and control applications can be built. In brevity, NOX is an OpenFlow controller [3]. Therefore, NOX applications mainly assert flow-level control of the network meaning that they determine how each flow is routed or not routed in the network. The OpenFlow switch is connected to the NOX controller and communicates over a secure channel using the OpenFlow protocol. The current design of OpenFlow only allows one NOX controller per switch. The NOX controller decides how packets of a new flow should be handled by the switch. When new flows arrive at the switch, the packet gets redirected to the NOX controller which then decides whether the switch should drop the packet or forward it to a machine connected to the switch. The NOX controller can also delete or modify existing flow entries in the switch. The NOX controller can execute modules that describe how a new flow should be handled. This provides us an interface to write C++ modules that dynamically add or delete routing rules into the switch and can use different policies for handling flows. D. Flow Table A flow table entry of an OpenFlow switch consists of a header fields, counters and actions. Each flow table entry stores Ethernet, IP and TCP/UDP header information. This information includes destination/source MAC and IP address and source/destination TCP/UDP port numbers. Each flow table entry also maintains a counter of number of packets, and bytes arrived per flow. A flow table entry can also have one or more action fields that describe how the switch will handle packets that match the flow entry. Some of the actions include sending the packet on all output ports, forwarding the packet on an output port of a particular machine and modifying packet headers (Ethernet, IP and TCP/UDP header). If a flow entry does not have any actions, then the switch drops all packets for the particular flow. Each Flow entry also has an expiration time after which the flow entry is deleted from the flow table. This expiration time is based on the number of seconds a flow was idle and the total amount the time (in seconds) the flow entry has been in the flow table. The NOX controller can chose a flow entry to exist permanently in the flow table or can set timers which delete the flow entry when the timer expires. III. Load-Balancer Design Load balancing architecture comprises an OpenFlow switch with a control device of NOX and server machines connected to output ports of the switch server. The OpenFlow switch uses an interface to connect to the Internet. Each server has a static IP address and NOX controller maintains a list of currently connected to the OpenFlow switch servers. Each server is running web server emulation on a well known port. [pic] Figure1. Load-balancer architecture using OpenFlow switch and NOX controller The hostname of server to IP address is resolved by each client and a request is sent to that IP address on the known port number. If you consider the above diagram, when a packet is reached to the switch from the client, the header information of the packet is compared with the entries of the flow table. If the header information of the packet corresponds to an inlet of the flow, the counter for the number of packets, the byte count is incremented, and the actions associated with the input of the flow are performed on the packet. If no match is found, the switch forwards the packet to NOX. NOX decides how the packet for this flow should be handled by the switch. NOX and then inserts a new article in the cash flow of the switch using the OpenFlow protocol. To achieve load-balancing features, the modules should be written in C++ that is executed by NOX controller. NOX should perform the function of handle () when a new flow arrives at the switch. This function sets the load balancing policy and adds new rules in the flow table of the switch. All client requests should be destined for the same IP address, then whatever the module is executed by NOX, should add rules for each flow which can modify the destination MAC and IP address of the packet with a server’s MAC and IP address. The switch will forward the packet to the server output port after modifying the packet header. When servers return a packet to the client, the module adds an entry flow that changes the source IP address with the IP address of the host that the client sends its request. So the client should always receive packets from the same IP address. If the client connection / server connection is closed or remains idle for 10 seconds, then the inactivity timer expires causing the input stream to be deleted from the cash flow of the switch. This allows input stream recycling Servers wait for a NOX to register and then report their current load on some schedule similar to the Listener Pattern. NOX in a separate thread listening on a UDP socket for heartbeats with reported by server loads and maintains a table with the current loads of all servers. When applying for a new stream is received, it chooses the server with the lowest and the load current increases to the low current server. This prevents flow of all flows routed to the same server as the server reports a new load. It also breaks ties by turning it into a round robin until the servers report their actual load heartbeat. Flow Algorithm Require: Flow, path 1: sourceHost = LocateSource(flow); 2: destinationHost = LocateDestination(flow); : layer = setToplayer(); 4: currentSwitch = LocateCurrentSwitch(); 5: direction = 1; //upward 6: path = null; //list of switches 7: return search (); This algorithm works as follows. When the OpenFlow controller receives a packet from a switch, it switches the control to the load balancer. Line 1 to 6 introduces the initialization for necessary variables. The load balancer ? rstly a nalyses the packet’s match information including the input port on the switch that receives the packet as well as the packet’s source address and destination address. Then it looks up those addresses using its knowledge about the network topology. Once the source and destination hosts are located, the load balancer calculates the top layer that the ? ow needs to access. We use the search direction ? ag. The ? ag has two values: 1 for upward and 0 for downward. It is initialized to 1. A path is created for saving a route grouped by a list of switches later. Line 7 calls search () that performs the search for paths recursively. In the method search (), It ? rstly adds current switch into path. It returns the path if current search reaches the bottom layer. It reverses the search direction if current search reaches the top layer. Then it calls a method 1: search () { 2: path. add(curSwitch); 4: if isBottomLayer(curSwitch) then 5: return path; 6: end if 7: if curSwitch. getLayer ( ) == layer then 8: direction = 0; //reverse 9: end if 10: links = findLinks(curSwitch, direction); 11: link = findWorstFitLink(links); 12: curSwitch = findNextSwitch(link); 13: return search (); 14:} that returns all links on current switch that are towards current search direction. Only one link is chosen by picking up the worst-? t link with maximum available bandwidth. And then the current switch object is updated. The method search () is called recursively layer by layer from the source to destination. At last the path will be return to the load balancer. The path information will be used for updating ? ow tables of those switches in the path. Flow Scheduling The Flow scheduling functionality works as follows. Each OpenFlow switch maintains its own ? ow table. Whenever any packet comes in, the switch checks the packet’s match information with the entries in its ? ow table. The packet’s match information includes ingressPort, etherType, srcMac, dstMac, vlanID, srcIP, dstIP, IP protocol, T CP/UDP srcPort, TCP/UDP dstPort. If it ? nds a match, it will send out the packet to the corresponding port. Otherwise it will encapsulate the packet in a PACKET IN message and send the message to the controller. As a module of the OpenFlow controller, the load balancer will handle the PACKET IN message. It ? nds a proper path by executing a search with the DLB algorithm described in Algorithm 1. The path is a list of switches from source to destination of the packet. Then the load balancer creates one FLOW MOD message for each switch in the path and sends it to the switch. This message will have the packet’s match information as well as a output port number on that switch. The output port number is directly calculated by the path and network topology. If one switch receives a FLOW MOD message, it will use it to update its ? ow table accordingly. Those packets buffered on ports of that switch may ? nd their matches in the updated ? ow table and be sent out. Otherwise the switch will repeat this process. IV. Future Work The OpenFlow specification includes an optional feature that would allow multiple NOXs to make active connections to the switch. In the case then of the NOX failing, another machine could assume the role of the NOX and continue routing traffic. Naturally the system would need to detect the failure, have a mechanism to remember any state associated with the current policy, and all servers would have to agree on who the new NOX was. These requirements naturally lend themselves to the Paxos consensus algorithm in which policy and leader elections can be held and preserved with provable progress [3]. We have implemented Paxos in another research project and could add it to our server implementation at the controller/signaler layer. As long as at least half of the nodes in the cluster stay up, state will be preserved and traffic should continue to flow. V. Conclusion It is possible to achieve similar functionality to a commercial load balancer switches using only physical commodities. The OpenFlow switch provides the flexibility to implement the arbitrary policy in software and politics separate the switch itself. Since the policy is decoupled from the switch, we can avoid the machine implementation of the policy of a single point of failure and the creation of a more robust system. References [1] OpenFlow Switch Specification. Version 0. 8. 9 (Wire Protocol 0x97). Current maintainer: Brandon Heller ([emailprotected] edu). December 2, 2008. [2] Web caching and Zipf-like distributions: evidence and implications. Breslau, L. Pei Cao Li Fan Phillips, G. Shenker, S. Xerox Palo Alto Res. Center, CA. INFOCOM 1999. [3] Paxos Made Simple. Leslie Lamport [4] M. Al-Fares, A. Loukissas, and A. Vahdat. A Scalable, Commodity Data Center Network Architecture. ACM SIGCOMM, 2008. [5] C. E. Leiserson. Fat-trees: Universal networks for hardware-ef? cient supercomputing. IEEE Transactions on Computers, 1985. [6] T. Benson, A. Anand, A. Akella, and M. Zhang. Understanding Datacenter Traf? c Characteristics. SIGCOMM WREN workshop, 2009. [7] HOPPS, C. Analysis of an Equal-Cost Multi-Path Algorithm. RFC 2992, IETF, 2000. [8] W. J. Dally and B. Towles. Principles and Practices of Interconnection Networks. Morgan Kaufmann Publisher, 2004. [9] S. Kandula, S. Sengupta, A. Greenberg, P. Patel and R. Chaiken. The Nature of Data Center Traf? c: Measurements Analysis. ACM IMC 2009. [10] N. McKeown, T. Anderson, H. Balakrishnan, G. Parulkar, L. Peterson, J. Rexford, S. Shenker, and J. Turner. OpenFlow: Enabling Innovation in Campus Networks. ACM SIGCOMM CCR, 2008. [11] R. N. Mysore, A. Pamporis, N. Farrington, N. Huang, P. Miri, S. Radhakrishnan, V. Subramanya, and A. Vahdat. PortLand: A Scalable, Fault-Tolerant Layer 2 Data Center Network Fabric. ACM SIGCOMM, 2009. [12] Beacon OpenFlow Controller https://OpenFlow. stanford. edu/display/Beacon/Home. [13] B. Lantz, B. Heller, and N. McKeown. A Network in a Laptop: Rapid Prototyping for Software-De? nded Networks. ACM SIGCOMM, 2010. [14] Y. Zhang, H. Kameda, S. L. Hung. Comparison of dynamic and static load-balancing strategies in heterogeneous distributed systems. Computers and Digital Techniques, IEE, 1997. [15] OpenFlow Switch Speci? cation, Version 1. 0. 0. http://www. OpenFlow. org/documents/OpenFlow-spec-v1. 0. 0. pdf. [16] N. Handigol, S. Seetharaman, M. Flajslik, N. McKeown, and R. Johari. Plug-n-Serve: Load-balancing web traf? c using OpenFlow. ACM SIGCOMM Demo, 2009. [17] R. Wang, D. Butnariu, J. Rexford. OpenFlow-Based Server Load Balancing Gone Wild. Hot ICE, 2011. [18] M. Koerner, O. Kao. Multiple service load-balancing with OpenFlow. IEEE HPSR, 2012. Figure 2: Load-Balancer block diagram architecture using OpenFlow switch and NOX controller.