How full-scale modeling and simulation achieved Quality-of-Service?


This blog focuses on how full-scale modeling and simulation is  used to achieve the Quality-of-Service such as packet loss, throughput, bit rate etc. QoS (Quality of Service) refers to all such mechanisms or tools ensuring critical applications requiring real-time data (usually audio or video) get sufficient dedicated bandwidth, with priority over other applications. It is extremely critical to the success of the project. For example, in the manufacturing sector where machines are beginning to leverage the network to provide real-time status information on any issues that may be occurring. Any delay in the identification of a problem can result in manufacturing mistakes costing tens of thousands of dollars each second. With QoS, the manufacturing status data stream can take priority in the network to ensure information flows in a timely manner.

Architectural decisions are design decisions that address architecturally significant requirements; they are perceived as hard to make and/or costly to change. The architectural decisions are always based on the target Quality-of-Service (QoS). Three things have to consideration before choosing an architecture, i.e. the traffic, the workflow, and designing requirements like higher throughput, lower latency, lower cost etc.  The QoS is not a static item. The QoS is extremely dynamic based on the user-types  , data patterns (video, audio, files)  and application set(For example, Organizations use QoS to meet the traffic requirements of sensitive applications, such as real-time voice and video, and to prevent the degradation of quality caused by packet loss or delay). The arbitration, number of Internet interface, request data sizes, response data, mode of operation (For example, distributed coordination function (DCF) and point coordination function (PCF); the DCF mode provides best-effort service, while the PCF mode has been designed to provide real-time traffic support in infrastructure-based wireless network configurations) and the clock-speed(For example, a CPU with higher clock speed from five years ago might be outperformed by a new CPU with lower clock speed, as the new architecture deals with instructions more efficiently) will affect the desired Quality-of-Service.  QoS can also be seen as a form of traffic control. Every day, a company’s network is bombarded by an onslaught of traffic. Some of this traffic is critical to the success of business operations, and some, while important, isn’t as critical or doesn’t require time-sensitive delivery. For example, many companies rely on File Transfer Protocol (FTP) as well as video-conferencing applications like Zoom or GoToMeeting. While both are paramount to employee productivity, FTP packets are not nearly as latency-sensitive as Voice over Internet Protocol (VoIP) packets. If delayed, FTP packets will still arrive intact. A delayed VoIP packet, on the other hand, runs the risk of arriving fragmented, ultimately resulting in disjointed video calls and ineffective business meetings.

There could be different QoS mechanisms (For example, Classification and Marking mechanism comprises two tactics working together to manage and prioritize network traffic whereas Congestion management evaluates the markings on each packet and then queues them appropriately based on a set of elaborate algorithms. ) for different modes of operation (For example BW Regulation mode vs. latency regulation mode). According to the mode, the QoS criteria also changes.

The Quality-of-Service metric is a combination of latency, throughput and energy consumption.  The QoS trade-off is between latency and energy which, if evaluated with many use-cases would lead to a lower cost and accurate processing capacity.

Capacity planning can improve the overall performance and availability of the server environment. While experienced IT staffers may be able to make an educated guess as to their capacity needs at any given time, such guesses are fallible and may not take all contingencies into account. As a result, many IT organizations that do not use capacity planning tools occasionally find their servers oversubscribed, delivering slow response times or timeouts that may slow business processes or cause customer dissatisfaction. With capacity planning tools, however, the enterprise can track trends—or, in some cases, model new applications or services—to anticipate potential performance and availability problems and eliminate them before they occur.

Capacity planning tools can also help deliver another key benefit in the current economy—budget predictability. During the boom years, enterprises could afford to provision their servers on the fly, over-provisioning new services at the outset and then purchasing new capacity when existing servers were oversubscribed. Under today’s tight budgets, however, the ability to predict future server needs is crucial to making intelligent buying decisions. Capacity planning enables the enterprise to develop an educated server purchasing strategy, and helps to reduce the need for “panic purchasing” situations that a server vendor might exploit to charge premium prices for its equipment.

With the right tools, you can locate underutilized capacity and move applications and subsystems to take advantage of that capacity, oftentimes delaying the purchase of additional servers. 

This type of in-depth analysis will eliminate large-scale over-provisioning and will aid in reducing both capital- and operational-budget. Management of the servers, software updates and the highly expensive energy bills will be reduced by this approach.

Advantage

The advantage of using full-scale modeling and simulation is that no hardware and software purchase or system administration is required. An architectural model of the entire data center can be visualized and tested for average, low and high-bandwidth situations.  This can help size the current system and provide space for future capacity expansion.  A dynamic simulation can replace the traditional network simulations that are slow and lacks configurability and visibility to analyze performance. Risk include over or under design, cost increases or schedule delays. The role of architectural model is for experimentation on capacity and power requirements of your data center or your server. The integrated graphical environment supports all flows for architecture design that includes workload and hardware modeling, architecture mapping, parallel sweeping, and analysis.

Using of a simulation model that can prototype the data center before its real-time deployment gives a clear scenario like the power requirements, bottlenecks of design, capacity requirements, fault analysis etc. of your application. The model so simulated produces statistics and graphs like the end to end latency, power, delay etc. which could help in making architectural decisions. It is capable of giving the designer a unified view of activity, performance and power for Root-Cause Analysis. The architectural model uses the graphical environment to capture, configure, simulate, and analyze the system-level performance and power of multicore system and next generation SoC architectures. It explores and optimizes the hardware-software partitioning and the configuration of the SoC infrastructure, specifically the global interconnect and memory subsystem, to achieve the right system performance, power, and cost. All of these can help in improving the profitability and reduce ongoing expenses.  For the data center management, server rental agencies, VM vendors and application providers, these models can be used to assist their customers with a reference and also demonstrate their unique value proposition.

Web reference: https://en.wikipedia.org/wiki/Quality_of_service