Server management systems

Enterprises receive the services and functions they need (databases, e-mail, website hosting, work applications, etc.) for their corporate IT systems based on servers , or rather, the so-called. "Server farms" (server farm). They are usually located in data centers, private (located in the organization itself) or public (cloud).

In any case, the organization's servers require constant and effective management so that their resources are used in the most optimal way.

What is Server Management?

Server management is the process of monitoring and maintaining servers in order to achieve their best performance. Server management also includes management of storage systems , software , security, and data redundancy .

The main goals of server management: Aruba Certified Design Expert

Minimization (and, if possible - complete elimination) of slowdowns and downtime;

Creation of an environment for the safe operation of servers;

Creating the conditions for evolution to meet the needs of the organization as it grows.

Various server management features

Virtual server

Virtualization is the main trend in information technology. A traditional server runs on a separate physical machine, while a virtual server allows multiple servers to run as multiple virtual machines on a single physical server. This makes it possible to improve the efficiency of the use of physical equipment. For example, if in the case of single physical servers in a data center, the average processor load is 4%, then in the case of virtual machines (servers), the physical processor load can be up to 80-90% or more.

At the same time, of course, the tasks of managing a virtual infrastructure become much more complicated. However, the general principles are the same for both physical and virtual servers.

The main elements of server management systems

Server management includes the following elements:

Equipment management . Monitoring the state of hardware (processor, memory, etc.) to ensure the best possible server performance.

Software management . Provides for regular updates of software, firmware and operating systems.

Security management . Security measures such as antivirus software, firewalls, access control and data encryption.

Reservation management . Regular data backup and data recovery plan in case of disasters.

Equipment management

Ensuring good hardware performance is the foundation of effective server management. It is important to continuously monitor at least five pieces of equipment:

Central Processing Unit (CPU), processor . The processor is the “brain” of the server, performing the computations that enable applications to run. The processor load must be constantly monitored so that the programs work correctly and without delays. If the processor load approaches 100%, then there is a big risk that when a new process is turned on, the processor will stop working normally, and the server, as IT engineers say, “freezes”. If the average processor load approaches a critical limit, then it is necessary either to replace the processor chip with a more powerful one, or to stop the operation of unnecessary programs that take up system resources. More fine tuning may involve adjusting other elements of the system to reduce the load on the processor.

Random access memory (RAM), memory . RAM is the server's working memory. This memory, designed for temporary storage of program data and calculation results, works much faster than permanent memory - hard drives (HDD), solid state drives (SSD), etc. The more RAM a server has, the better its performance. The use of RAM must be constantly monitored, and if the time average value of the occupied memory approaches its maximum size, then it is necessary to take measures to expand the memory.

Hard disk drive (HDD) . This is a persistent memory, a server data storage. This is where programs and data are stored that remain on the hard drive even when the power is turned off, unlike RAM, where information is stored only when the power is turned on. If the hard disk is "crammed" with data "to the eyeballs", this also greatly reduces the performance of the server as a whole. Therefore, you need to constantly monitor the availability of free space on your hard disk, add new disks, delete unnecessary data. Or you should consider connecting to cloud storage, where all of the above will be done automatically.

Processor temperature . Servers usually generate a lot of heat during operation. Most servers come with wired thermometers to keep track of whether the processor temperature is within specified limits. If the server gets very hot, you need to find the reasons for the rise in temperature. Server operation in this mode can lead to application failures and complete processor failure. Cooling fans are one of the essential components of a physical server. If the fans (which are also called "coolers") fail, there is a high risk of overheating of the physical server processor and its failure.

Environment . In addition to monitoring the temperature inside the server, you must constantly monitor the ambient temperature. The server room must be kept at an appropriate temperature and humidity and airflow must be controlled (“hot and cold aisles”). This not only helps prevent server breakdowns, but also allows you to achieve optimal performance.

When choosing a physical server, it is important to proceed from the requirements for its parameters and purchase a server that slightly exceeds these requirements. However, choosing a server configuration that is too powerful "with a margin" can lead to an unreasonably high price, as well as unnecessary waste of energy.

Fiber Optic Cable Buying Guide

What is a fiber optic cable and how does it work?

Fiber optic cables send digital data at the speed of light ... because they work by transmitting light through flexible, optically pure glass or plastic fibers .

Each fiber optic thread is about the width of a human hair. The fibers are arranged in bundles called optical cables - light is transmitted through this core. An outer optical material called plating surrounds the core and reflects light into the core.

Because fiber cables can transmit over long distances, they are ideal for networking, telecommunications, and storage applications in cable cabinets, distribution frames, gateways, central offices, and data centers.

What are the different types of fiber optic cables?

There are two basic types of optical cable used for data and communications, singlemode and multimode. The differences in principle are the size of the core and the distance signals can be transported. The fiber cable is measured by the core and the diameter of its plating in micrometers (µm).

The multimode fiber has a relatively wide core size of either 62.5 μm or 50 μm. It normally transmits infrared light from light-emitting diodes (LEDs). Generally, multimode cable is used for short or medium distance communication, such as inside a building or on a small campus. Typical multimode fiber ties are suitable for distances under 2000 meters.

Multimode fibers are identified by the name OM (optical module) given in the ISO / IEC 11801 standard: Dark Fiber

OM1

Fiber with launch bandwidth (OFL), 200/500 MHz-km, with a bandwidth at 850 / 1300nm (usually 62.5 / 125 µm fiber)

OM2

Fiber with OFL bandwidth of 500/500 MHz-km at 850 / 1300nm (usually 50/125 µm fiber)

OM3

50 um laser optimized fiber with an efficient bandwidth of 2000 MHz-km (EMB, also known as laser bandwidth), designed for 10 Gbps transmission

OM4

50 um laser optimized fiber with an EMB bandwidth of 4700 MHz-km designed for 10 Gbps, 40 Gbps and 100 Gbps transmission

OM5

50 µm fiber optimized with OM5 laser with 4700 MHz-km EMB designed for 10 Gbps, 40 Gbps and 100 Gbps. Designed to support shortwave division multiplexing (SWDM) technology.

Singlemode fiber has a narrower core size of 8.3 µm. It transmits infrared light from lasers and provides twice the bandwidth rate of the multimode cable and can provide a distance of 50 times the multimode. On the flip side, it costs more than the multimode cable. The singlemode duplex cable is commonly used in long-term network connections.

How far can a fiber optic cable carry a signal?

The distance limit depends on the cable style, the wavelength and the network itself. The usual ranges are about 984 ft for the 10 Gbps multimode cable and up to almost 25 miles for the singlemode cable. If a longer duration is required, optical amplifiers or repeaters are used to ensure signal integrity over the entire distance.

the fiber optic cable is a network cable that contains strands of fiberglass inside an insulated housing. They are designed for high performance data networks and telecommunications.

Compared to wired cables, fiber optic cables provide higher bandwidth and can transmit data over longer distances. Fiber optic cables support much of the world's Internet, cable television, and telephone systems.

BIG-IP Local Traffic Manager

BIG-IP Local Traffic Manager  (LTM) turns your network into a dynamic infrastructure for delivering applications. It acts as a complete intermediary between users and the application server, creating an abstraction layer for protecting, optimizing, and load balancing application traffic. This gives you the flexibility and control you need to easily add applications and servers, eliminate downtime, improve application performance, and meet your security needs.

BIG-IP LTM is a complete proxy between users and application servers that provides security, optimization and load balancing of application traffic.

Main advantages: F5 certification

1. Ease of application deployment and guaranteed availability

2. Ability to manage application delivery

3. Acceleration and optimization of applications: intelligent distribution, compression, flexible QoS L7 management, TCP Express, iSession.

4. Securing applications, networks and data: Hiding resources and service errors, selective encryption, encryption of cookies, analysis of protocol anomalies, firewall capabilities and other functions

5. Reducing server load: content conversion, OneConnect, fast caching, SSL acceleration and offloading, TCP connection queuing and other technologies

6. Easy to configure and manage: iApp configuration templates, correlated application statistics, hardware clustering and configuration synchronization, administrative domains.

Complex load balancing:

BIG-IP LTM includes static and dynamic load balancing tools such as Dynamic Ratio, Least Connections and Observed Load Balancing, while dynamically tracking server performance in a group, so that the best resources are always chosen to achieve the highest performance and scalability.

Application state control:

BIG-IP LTM has sophisticated controls that monitor the availability of devices, applications and content. These tools include specialized monitors for various applications (including various application servers, SQL, SIP, LDAP, RADIUS, Diameter, XML / SOAP, RTSP, SASP, SMB, and many more), as well as custom monitors that inspect content and simulate calls. applications. 

High availability and transaction security:

Device Service Clustering provides flexible high availability scaling and synchronization of active and live application traffic configurations between active and standby devices. In this case, the active / standby configuration is not 1: 1, but N: 1, that is, the application load is distributed to as many active devices as they are required depending on the constraints imposed by the resources and the availability of resources, that is, true horizontal scaling is provided. BIG-IP LTM provides system failover in less than a second and full link mirroring. This achieves extremely high availability of the solution in the event of any system, server or application failure.

Available as:

BIG-IP hardware platform

VIPRION blade chassis

* BIG-IP Virtual Edition can run on VMware ESXi, Microsoft Hyper-V, Citrix XenServer, Linux KVM hypervisors, and Amazon Web Services

The block system for managing the functionality of F5 Networks solutions allows you to combine the proposed solutions in an extensive framework. So the manager of local traffic BIG-IP Local Traffic Manager is available as a separate solution, as part of specialized  bundles , or in any combination with the  modules offered by the company

Linux system administrator

Our values: linux jobs

Take care of our customers.

Always speak openly and honestly.

Respect each other and thrive together.

Create and sell a product to be proud of.

Make the most of the benefits of individual differences and everyone's contribution.

Requirements:

Work experience in a similar field of at least 7 years.

Clear thinking and the ability to reasonably defend your point of view.

Understanding of the structure of networks and the TCP / IP network protocol stack and the ability to apply this knowledge in practice.

Experience in configuring linux kernel for large network loads.

Confident knowledge of administering Linux servers / services: nginx, haproxy, openvpn, dns, iptables, raid, fs, etc.

Experience with virtualization systems (KVM), monitoring (Zabbix), configuration management (Ansible).

Using scripting languages ​​to automate tasks (bash, python).

Ability to work in a team, self-learning.

Knowledge of English at the documentation reading level or higher.

Desire to develop and learn new technologies.

Will be a plus:

Experience in using Logstash, ElasticSearch, Kibana.

Experience in administration and monitoring of PostgreSQL and ElasticSearch DBMS, RabbitMQ message broker.

Experience in using a containerization and orchestration system (Docker, Kubernetes).

Skills of working with the git version control system.

Skill in writing simple SQL queries.

Experience in working with a development team, understanding the software development process, knowledge of tools related to software development (git, gitlab, youtrack).

Introduction to ceph, aws, terraform

Tasks:

Ensuring the health of a small fleet of physical and virtual servers.

Providing fault tolerance for opensource applications.

Prompt response in case of equipment failure and / or loss of part of the system's functionality.

Properties of the wireless transmission medium - CCNA Routing and Switching

Wireless transmission media allow transmission of binary data by encoding them into electromagnetic signals of the microwave and radio range.

Wireless provides the greatest mobility of all transmission media, so the number of devices that support wireless connectivity is growing every day. As bandwidth increases, wireless connectivity is gaining popularity in corporate networks cisco wireless certified.

The wireless environment has the following features to consider.

Coverage area . Wireless data transmission technologies work well in open spaces. However, some building materials used in the construction of buildings and structures, as well as local conditions, can limit the coverage area.

Interference . Wireless connections are susceptible to interference and may be degraded by common devices such as cordless phones, certain types of fluorescent lights, microwave ovens, and other wireless communications.

Security . You do not need to connect to physical cables to access the wireless environment. Therefore, this environment can be accessed by unauthorized users and devices. Hence, security is a major aspect of wireless administration.

General transmission medium . WLANs operate in half duplex mode, which means that only one device can transmit or receive at a time. The transmission medium is common to all wireless users. The more users simultaneously connect to a WLAN, the less bandwidth each of them has. Half duplex will be discussed later in this chapter.

While the popularity of wireless networking of desktops to the network is growing, copper and fiber remain the most popular physical media.

Fiber Optic Transmitters

The most important component of a fiber optic transmitter is the light source (usually a semiconductor laser or LED). Both serve the same purpose - the generation of a microscopic light beam that can be introduced into the fiber with high efficiency and modulated (changed in intensity) at a high frequency. Lasers provide higher beam intensities than LEDs and allow higher modulation frequencies; therefore they are often used for long-haul broadband lines such as telecommunications or cable TV. On the other hand, LEDs are cheaper and more durable devices, moreover, they are quite suitable for most systems of short to medium length, and therefore they are widely used in closed TV systems.

In addition to its functional purpose (i.e. what signal it should transmit), a fiber-optic transmitter is characterized by two more important parameters that determine its properties. One is its output power (intensity) of the optical emission, and the other is the wavelength (or color) of the emitted light. Usually these are 850, 1310 or 1550 nm, values ​​selected from the condition of coincidence with the so-called. "Transparency windows" in the transmission characteristic of an optical fiber material.

Fiber optic installation

An optical fiber consists of a high refractive index center (core) surrounded by a low refractive index material cladding, as shown in Fig. 1.2. a fiber is characterized by the diameters of these regions - for example, 50/125 means a fiber with a core diameter of 50 µm and an outer cladding diameter of 125 µm.

Light propagates along the fiber core due to successive total internal reflections at the core-clad interface; its behavior is in many ways similar to that of being caught in a pipe, the walls of which are covered with a mirror layer. However, unlike a conventional mirror, in which reflection is rather inefficient, total internal reflection is essentially close to ideal - this is the fundamental difference between them, which allows light to propagate along the fiber over long distances with minimal loss.

A fiber made in this way (Figure 1.3) is called a stepped refractive index and multimode fiber because there are many possible paths, or modes, for the light to propagate. This set of modes results in dispersion (broadening) of the pulse because each mode travels a different path in the fiber, and therefore different modes have different transmission delays from one end of the fiber to the other. The result of this phenomenon is a limitation of the maximum frequency that can be effectively transmitted for a given fiber length - an increase in either the frequency or the fiber length beyond the limit values ​​essentially leads to the coalescence of successive pulses, making it impossible to distinguish them. For a typical multimode fibers, this limit is approximately 15 MHz  ·  km, which means that the video signal with a bandwidth of, e.g., 5 MHz may be transmitted a maximum distance of 3 km (5 x 3 km MHz = 15 MHz  ·   km). Attempting to transmit the signal over a greater distance will result in progressive loss of high frequencies.

For many applications, this figure is unacceptably high, and a search was made for a fiber design with a wider bandwidth. One way is to reduce the fiber diameter to very small values ​​(8-9 microns), so that only one mode becomes possible. Single-mode, as they are called, fiber (. Figure 1.3, b) is very effective in reducing the variance, and the resulting band - many GHz  ·   km - making it ideal for telephone and telegraph public networks (PTT), and cable television networks. Unfortunately, a fiber of such a small diameter requires the use of a high-power, precision-aligned, and therefore a relatively expensive laser diode emitter, which makes them less attractive for many applications associated with short-range closed-loop TV systems.

Ideally, a fiber with a bandwidth of the same order of magnitude as a single-mode fiber, but with a diameter of the same as a multimode fiber, is required in order to be able to use inexpensive LED transmitters. To some extent, these requirements are satisfied by a multimode fiber with a gradient change in the refractive index (Fig. 1.3, c). It resembles a multimode fiber with a step change in refractive index, which was mentioned above, but the refractive index of its core is inhomogeneous - it smoothly changes from a maximum value in the center to a lower value at the periphery. This has two consequences. First, the light travels along a slightly curving path, and second, and more importantly, the differences in propagation delay of different modes are minimal. This is due to the fact that high fashion, those entering the fiber at a higher angle and traveling a longer path actually begin to propagate at a faster rate as they move away from the center into the region where the refractive index decreases, and generally move faster than the lower-order modes remaining near the axis into fibers, in the region of high refractive index. Increase speedjust compensates for the greater traversable path.

Gradient index multimode fibers are not ideal, but they still exhibit quite good bandwidth. Therefore, in most closed-loop TV surveillance systems of small and medium length, the choice of this type of fibers is preferable. In practice, this means that bandwidth is only rarely a parameter to be considered.

However, this is not the case for fading. The optical signal is attenuated in all fibers at a rate dependent on the wavelength of the transmitter by the light source. As mentioned earlier, there are three wavelengths at which optical fiber attenuation is usually minimal - 850, 1310, and 1550 nm. These are known as transparency windows. For multimode systems, the 850 nm window is the first and most commonly used (lowest cost). At this wavelength, a good quality gradient multimode fiber exhibits an attenuation of about 3 dB / km, which makes it possible to implement communication in a closed-loop TV system at distances over 3 km.

At a wavelength of 1310 nm, the same fiber shows even less attenuation - 0.7 dB / km, thereby allowing a proportional increase in the communication range to about 12 km. 1310 nm is also the first operating window for single-mode fiber optic systems, with an attenuation of about 0.5 dB / km, which in combination with laser diode transmitters allows the creation of communication lines longer than 50 km. The second transparency window - 1550 nm - is used to create even longer communication lines (fiber attenuation is less than 0.2 dB / km).

Fiber Optic Receivers

Fiber optic receivers solve the vital problem of detecting extremely weak optical radiation emitted from the end of the fiber and amplifying the received electrical signal to the required level with minimal distortion and noise. The minimum level of radiation required by the receiver in order to provide acceptable output signal quality is called sensitivity; the difference between the receiver sensitivity and the transmitter output power determines the maximum allowable system loss in dB. For most CCTV surveillance systems with an LED transmitter, the typical figure is 10-15 dB.

Ideally, the receiver should work well when the input signal changes over a wide range, since it is usually impossible to predict in advance exactly what the attenuation will be in the communication line (i.e., line length, number of joints, etc.). Many simple receiver designs use manual gain control during installation to achieve the desired output level. This is undesirable, since changes in the amount of line attenuation caused by aging or temperature changes, etc., are inevitable, which dictates the need to periodically adjust the gain.

All fiber optic receivers use automatic gain control, which monitors the average level of the input optical signal and changes the receiver gain accordingly. No manual adjustment is required either during installation or during operation.

Cisco has built a modern green data center

Cisco has decided to build a state-of-the-art environmental data center, which it has already completed in Allen, Texas. It used its complete portfolio of technologies for data centers, which includes solutions for computer systems, switches ccie data center jobs.

Cisco has decided to build a state-of-the-art environmental data center, which it has already completed in Allen, Texas. It used its complete portfolio of data center technologies, which includes solutions for computer systems, switches and data storage. It is optimized for virtualization technologies and cloud computing.

The batteries charged by solar cells are not used for operation in this 5 megawatt center, but rotating flywheels and solar collectors covering the roof of the building. Solar collectors on the roof are able to generate around 100 kW of energy and in the event of a power failure, the less environmentally friendly diesel generators will start. Cisco also mentions an energy-efficient cooling system that uses uncooled and filtered outside air for 65 percent of its operating time, saving the company about $ 600,000 a year. Cisco also thought about irrigating surrounding plants for which it used a rainwater lagoon.

And in terms of hardware data, the center uses technologies from EMC, NetApp and VMware partners, Nexus 7000, Nexus 5000 series corporate switches, Nexus 1000V virtual switch, MDS network storage switches, Data Center Network Manager and the NX-operating system. OS. Find out more in the virtual tour of the building (see source).