Businesses and consumers want to achieve more, faster, richer content and service delivery through any device, anytime, anywhere. The above features and requirements can be met through cloud computing, but only if the data center is able to overcome the challenges of network design.
Why is network complexity particularly important?
Today's data center networks have become too expensive and complex to achieve a cloud-computing vision, except in rare cases. Most IT managers acknowledge that the following situations exist in their data center network.
New application architecture. The old "Stateful client Server" Application mode limits the application traffic between the communication server and the client "South/north", but it cannot be scaled to meet the needs of tens of thousands of users, let alone millions of of users. Today's service-oriented network architecture and other Web2.0 architectures extend traffic through multiple dedicated servers, creating new requirements for the network.
New content. Now, from the training courses to the phone, everything has been digitized, and the content and storage of the mobile has become richer and more rapidly growing. Each innovation or performance improvement will result in greater network requirements.
New client terminals. The development of portable client terminals and wireless client terminals, coupled with the need for all servers and storage infrastructures to adapt to these client terminals, has resulted in increased network traffic and increasingly complex network connectivity.
Efforts to overcome these trends can only delay the problem, and may even increase network complexity. In this respect, I give the following examples to illustrate.
More network equipment. The ability of standard fast solutions is limited, new network hardware will only bring more interconnection and work, increase latency and cost, and increase network complexity.
Dedicated network. Data centers often address specific problems by adding private networks. Ethernet mobile data, Fibre Channel SAN connectivity Server storage array. But fragmented networks greatly enhance the complexity of virtualization and day-to-day network management.
Centralized resources and management. While incorporating infrastructure and operations into your own, outsourced, or coordinated data centers can improve utilization, efficiency, and control, it has an impact on the performance, availability, and security benefits of your network infrastructure.
Virtualization. is essentially another way to centralize resources. Virtualization can create storage, server, and network segmentation from fewer physical devices. However, the deployment and migration of virtualization workloads can be highly demanding for network performance and flexibility.
These problems are not only in the data center network, but also in the most demanding network environments. Depending on the level of equipment, they may even affect companies that do not intend to provide cloud computing services.
The economic way to simplify the network
Ideally, a centralized, essentially chunking network would become a single "logical" switch that would achieve physical redundancy (in terms of usability) and provide high performance, reduce latency, and conduct arbitrary conduction across the data center with a variety of traffic.
Today, only a very small number of data center networks can achieve this ideal, but most data centres can benefit from a new approach, which is about to be simplified as a core element. With new technology, designers can reduce network complexity and total cost of ownership by reducing the exchange layer, reducing interconnection, consolidating security infrastructures, standardizing a single operating system, and automating network operations across switching, routing, and Network service infrastructures. Whether companies agree on a highly virtualized, network-intensive approach to cloud computing, this simplification provides the foundation for today's high-performance networks and future rapid adaptation and growth.
Companies pay close attention to their operating costs. Reducing the number of switches can reduce operating costs, but savings can be achieved everywhere when the network layer is rationalized.
Effective design can integrate network infrastructure to reduce equipment, save space, power, cooling and management costs.
Network flattening facilitates the centralization and virtualization of security applications, replacing multiple SSL, VPN, Nat, and other devices. Consolidating security services can cut spending and make the network easier to manage, more efficient, and more secure.
Effective management can save money, and reducing the number and diversity of equipment can help simplify management. In addition, standards-based interfaces enhance the efficiency of the staff, use automated tools for network management, and provide day-to-day technical support.
The performance of Network simplification
In addition to reducing deployment and management costs, flat networks can improve network performance. Data packets from each switch or device increase the latency of the data path, so removing the swap layer reduces end-to-end latency. Flat networks also help organizations effectively use existing applications: for example, based on the application of a service-oriented architecture (SOA), a data center network can handle all the traffic from the server to the server generated by the application.
High-performance, low-latency networks are important for any company that plans to expand its virtualization program. Network migration planning is an important step in achieving the scale of virtualization and should be done in the first few phases. Improvements in cost and performance will quickly become a focus for organizations that intend to benefit from new computing, storage, and software technologies. In terms of long-term operational improvements, data centers and their operations will benefit from faster, simpler, more flexible, and cheaper foundations.
A single, highly scalable datacenter Network will help centralize server and storage resources, achieve more efficient load balancing, and better utilization. Whether your organization is planning a full range of cloud computing initiatives, or planning to increase revenue from virtualization, networked storage, and new application delivery models, first understand your network needs. A more flat, simpler network design will soon bring you a return on investment and help you deal with future challenges more quickly.