At present, software-defined networking (SDN) has become a fashionable topic in the industry. SDN technology and its possible impact have been highly valued by academics and industry, and there are also different understandings and understandings of SDN. Recently, from a well-known social news site, Reddit, we heard the opinions of frontline engineers on ADN. They agreed that there is still a long way to go before we can really purchase SDN technology on a large scale. In Reddit's discussion post, network engineers demanded that such bad marketing practices be stopped, and they expressed real SDN needs to suppliers. We invited technical writer David Geer to find some network engineers on Reddit to discuss why they are not ready to purchase SDN technology. The information we get is a sharp commentary on Reddit style - full of questions about what network engineers really want to get from SDN. Voice from Reddit: Network personnel say "no" to SDN Geer first raised the following question: "Now SDN is already in the real product. Before seriously considering trial SDN, what do the regular network engineers, network managers and administrators of large enterprises (non-giant companies) need to get from SDN in the next 1-2 years?" The first answer seems to come from a network engineer who is busy at work: "It needs to collaborate and integrate more cloud options, and it has a dynamic infrastructure that can self-repair and continually optimize performance through the use of collaborative strategies. This will enhance the competitiveness of the enterprise by binding solid quality commitments." These words made the very personal network engineers on Reddit arguing fiercely. An engineer replied: "As soon as I see the word synergy, I immediately know that the rest of this article is junk." This made me want to laugh. What's more interesting is that there is the following words: Engineers hope that SDN can reduce their infrastructure costs - but is this possible? There is such a reply later: "I need SDN to lower the price of the hardware. Other than that, it doesn't do much to me." Another engineer agrees with this statement: “Reducing equipment prices while making my distribution system cleaner and cheaper.†Of course, there must be a response from the supplier (but at least he is telling the truth): "What price do you want? Have you seen the white box switch? I am from Cumulus, so I prefer white box devices, but if you buy a white box device, you can install multiple operating systems like a server. †And our reporter wants to know more about improving the cost-effectiveness of the network through SDN. Finally, an engineer joined the exchange, and his views differed from many SDN supporters—he preferred traditional network technology. "The term SDN has been abused by the entire industry. In fact, many people are not talking about SDN, but more are saying that a certain vendor is already implementing SDN because they claim that their device management API is SDN-enabled. This phenomenon, I am a fashionable network engineer, because as early as SDN popular and I have been implementing SDN (write the library and tools to manage network device configuration according to their own needs). You may think, 'but you can yourself Use software to develop protocols and control traffic. 'I really don't worry about this. Most network engineers don't worry about this. The existing protocol is better and faster. I don't have any traffic flow problems. I need to use SDN. Solve. Until now, most people who say SDN are finally some system administrators, they think that they find a new method is an agreement that has not been understood for more than 20 years." What we really need is configuration management and mapping! Many people think that SDN is ultimately automation and management, and one engineer wants to know more about this: "I have a question. Does SDN support automated generation of network diagrams? This is my biggest concern in the next 1-2 years." However, the response he received was not optimistic. "LLDP (Link Layer Discovery Protocol) can be supported, but it is not complicated." The next responder is not interested in SDN at all: "No, definitely not. You can automate the generation of network diagrams, but they are a bunch of garbage. LLDP can't help you generate useful network diagrams." Another engineer’s point of view is to bring some hope: "It really doesn't work - but, as long as the ideas are correct, we (network engineers) should be able to get consistent documentation and configuration. Some of them are related to SDN - like it sees the network as a network, not a series of isolated devices that make up the network "" Finally, someone came up with a very interesting but very real reflection of SDN: "I don't have such a rich experience, I can't give a smarter point in a short time (not much more in 20 years) - so, for me, a big truth: automatic allocation. Controlled by a simple click operation Network-side distribution operations. There is also automatic allocation of devices such as CPE (Customer Field Device) / PE (Supplier Boundary) / NID (Network Interface Device). This business case is to reduce operating costs because it can reduce the completion of these tasks. Network personnel." Then another engineer put forward ideas about management, exchange, and cost-effectiveness: "What I want is NetFlow in all locations. I hate SNMP and it lacks a 100% CDP environment. I must have been broken by Cisco." I am very satisfied with the price/performance of most Layer 3 switches, but the price of the access layer equipment is too high. We didn't get the features we needed at all, and we had to buy a lot of things we didn't want. This is a major concern because our access layer equipment is far more than other levels. Due to the growth of the network, we have to buy a core device every 5-10 years, not for the pursuit of new features, and usually do not even consider the price issue (buy it if needed), but in the process of expansion and growth, we continue to Ground stacking access switches. It is hoped that there will be some equipment with ultra-stable, extremely low overload and zero support costs, and they must have extremely stable monitoring and management functions. UbiquiTI's switch product line is now the only vendor product closest to my requirements. Unfortunately, I don't think anyone is really willing to enter this low-end market. We abandoned the standard switch in the data center and switched to Mellanox's Infiniband device. Its performance is extremely high, almost close to zero overload, and it is very easy to use. SDN must focus on environments outside the data center. SDN has always focused too much on data centers, but engineers have been trying to use SDN in more environments for years. Reddit's discussion also covers this issue. “I think the problem is that when people start talking about and really designing SDN, their view of the network is limited to the data center. All the discussions revolve around the data center. Although I agree that SDN has the potential to redefine the functionality of the data center network, Leaving other aspects, you can't build and design an automated network platform. I understand that the general idea is to be able to quickly and efficiently allocate new application services and then allow the network to be self-built, but many other types of networks can also benefit from The concept, they apply to today's customers, can bring them value." “You mean, because you can use any hardware, you can make suppliers compete for your business? Is this the way you want SDN to reduce hardware prices? Or, do you mean lower support costs?†Later, Cumulus had another reply: “Cumulus Networks offers SDN methods (stacking, network virtualization, and automation). We are the support and support structure for Google-style data centers; for businesses, we are 'support' and physical switches.†SDN is not new, and engineers like traditional networks. We can provide many different materials braided sleeve, such as PET, PP, Nomex, etc., different materials have different characteristics, suitable for different environments and application scenarios, customers can choose different materials and specifications according to their needs. Applied to the home scene, PET Braided Sleeve is the most cost-effective choice, PET Expandable Braided Sleeve is made of polyester monofilament, which has a wonderful expandability, flame resistance, durable and heat insulation.PET Expandable sleeve is very smooth, a lot of shining colors available.Excellent flexibility, easy to be bent, durable, expandable.PET Expandable Braided Sleeve is widely used in the field of protecting computer cable, audio or video cable, automobile wiring harness, wire etc.we can provide different colors, different colors of PET weaving sleeve, it can not only protect the wire, but also make the messy wire and cable look more neat and beautiful, and applied to the automotive, military, automation equipment and other industrial fields, The Nomex,Carbon fiber and other high-performance materials are good choice. Nylon Braided Sleeves,Kevar Cable Sleeving,Optical Cable Sleeve,PET Cable Sleeving Dongguan Liansi Electronics Co.,Ltd , https://www.liansielectronics.com