Archive for November, 2009

Remote Graphics Unit Explained

A System which permit computer system to be separated from output/input devices like mouse, keyboard, display monitors, audio speakers is called a Remote Graphics Unit (RGU). Graphics sub-system of the computer is the main part being remoted. Typically there are three main causes behind using separate user interface of a computer from the actual computer are:

  • Decrease noise and heat from the rooms where many computer operators are working.
  • Smooth the progress of  computer maintenance by placing all computers in very close closeness to each other.
  • Locking computers away from users for government or commercial security.

RGU or Remote Graphic Unit originated with researches on graphics controller on mainframe computers in the 70s era. Earlier Remote graphic units are associated with high end workstations running Unix-like operating systems and later with Windows. Normally RGUs are used for some unique applications like Computer Aided Drafting, Financial Services Commodity Trading Desks, Remote Sensing e.t.c. Dedicated X terminals are also included, depending on the choice of the user how he wants to use RGU. A Remote Graphics Unit effectively split computers PCI-Express bus or PCI and then transmit only bus command over to the user side. RGUs process of working is different from that of DVI Extension and KVM Extension. Traditional graphics unit or GPU perform processing in DVI and KVM Extensions. Bus data however is way too small then rendered graphics data, theory behind the working of RGU is like it is possible to achieve better graphics and resolutions performance if only the distance of separation is large between the computer side and user side input/output devices.

Remote Graphics Unit

Tags: ,

19-inch rack – most famous rack type

19-inch rack is the most famous standardized enclosure used for mounting multiple equipment modules. In a 19-inch rack there is a front panel 19 inches wide with ears that stick out on each side allowing modules to be fastened to the rack frame with screws. Rack-mount instrument, rack-mount, rack mountable, subrack, a rack mount chassis or just shelf these are some of the terms coined for the equipment which is planned to be placed in a rack. Electronic modules are also fixed at the height of  “U” one rack unit which measures 1.75 inches. Most commonly used to house professional video and audio equipment, this type of rack is most preferred one by Data Centers to place their servers. 19-inch racks allow dense hardware configurations devoid of excessive requiring shelving. One of the other uncommon used of 19-inch rack is industrial control, automation and power hardware.

There are four important steps inn Equipment Mounting:

Fastening – mounting equipments starts from the process of fastening. In order to receive particular type of threaded bolt mounting holes are tapped. However this hole tapped kind of racks are only used for the hardware that really changes like relay racks and network or telephone cabling panels. Hardware which is frequently changed is mounted on racks which are provided with slide rails, since there is a possibility that the frequent movement can break of a bolt or damaging a thread making mounting hole unusable.

2 u rack

Structural Support – mounting of the equipment is confirmed by clipping equipment’s front panel to the rack. One of the disadvantage of the process is that all the structural support is centered only to one side of the edge, so heavier equipment is designed to mounted on a second pair of mounting strips located at the back of the equipment. A variety of spacings are used between front and back strips, with 31.5 inches being the most common one. Equipment mounted is deigned to manage a range of depths and 39.4 is the most common used space because of the reason is more depth permit more space for to route cables at the back of the rack. Racks are commonly secured to the adjacent wall or floor so that it will not fall over. This process is done specially with 2 post racks.

Slides -  heavy equipment which is continuously used for servicing an in the process all the four corners are detached concurrently can cause a threat of damage. To avoid damage heavy equipment is either mounted via rails or mounted directly. A rail is mounted directly onto the rack equipment is then slides into the rack along with the rails which supports it. When the equipment reaches the place it is then bolted back to the rack. Rails also support the equipment in a position where it is clear of the rack and maintenance and inspection can be done easily.

19 inch Rack

Tags: , ,

KVM switch review

A hardware device which allows users to control multiple computers from a single mouse, video monitor or keyboard is called a KVM switch. KVM is an abbreviation for Keyboard, Visual Display Unit, Mouse or Keyboard, Video, Mouse. Multiple Computers can be connected to KVM, but usually only few computers are controlled at any certain point of time. There are also few KVM switches which work in reverse and i.e. s single computer can be connected to multiple keyboards, monitors, and mice. This kind of configuration is useful when operator wants to access a single computer from two or more locations. Latest KVM switches can be connected to USB devices, multiple computers and speakers. There are many kind of KVM switches available in market like USB KVM, PS/2 KVM and IP KVM, all offered in 16 Port KVM, 8 Port KVM and 4 Port KVM. You can also integrate KVM switch into LCD Drawers whilst occupying only IU of space. Both PS/2 and USB devices are supported by KVM on same switch. KVM switches can provide remote BIOS Level access to as much as 128 servers. KVM are quite popular and organizations from hospitals to banks use them, specially because their networks are widely spread out. One more feature or advantage with KVM is Climate monitoring, i.e. they provide resolutions to guard hardware at the Data Center.

KVM over IP

For hardware KVM switch there are many software available for use. Input Director, VNC (Virtual Network Computing), MaxiVista, PC Anywhere, Synergy and Synergy + are some of them. They do switching in software and forward input over standard network connections. Advantage of these software alternatives of KVM switch is that they reduce number of wires required plus it makes users forget that they are using two computers because of the screen edge switching. Now looking for disadvantages first of all these software alternatives usually require additional software to be installed in each of the target computers so that clients can remotely attach to it. Software alternatives are impossible to use is the computer has stopped responding to system network connection or it is very busy also, if the Operating system is not installed on the host side these software can’t be accessed.

KVM over IP-1

Local Remote and KVM over IP are the two remote KVM devices available. Out of these two KVM over IP is more widely used. A KVM over IP uses a dedicated microcontroller and potentially specialized video capture hardware in order to capture mouse and keyboard signals, video, they are compressed and converted into packets and then sent over an Ethernet link to a remote console application which convert them back into dynamic graphical image. To make it available throughout the BIOS boot process KVM over IP switch is connected with a system’s standby power plane. KVM over IP allow multiple computers to be controlled remotely across a wide area of network, telephone line, local area network using TCP/IP protocols. KVM over IP devices have been applied in many different ways. PCI based KVM over IP is used for the graphic capture portion. It uses variation of a technique called screen scrapping where the PCI bus master KVM over IP card access data directly from the graphics memory buffer. However in these cased PCI card must know which graphics chip is working with and what graphics mode this chip is currently in. This is because contents of the buffer can be interpreted properly as picture data.

KVM over IP-2

Tags: , ,

What is Data Center Bridging

Data center bridging

Augmentation of Ethernet local areas networks for the use in Data Center this whole process is known as Data Center Bridging. The Task Group of IEEE 802.1’ Work Group carries out the efforts to improve the quality. Ethernet is traditionally a primary network protocol in Data Centers for computer to computer communications. Ethernet is designed in such a way that in order to best effort network, it may deliver packets out of order or drop packets when devices or network are busy. In internet protocols responsibility of the transport protocols is traditionally judged by transport reliability. Best advantage of Ethernet is to add extensions to the existing protocol suite in order to provide reliability without inviting penalties of TCP. There are good chances of higher granularity in bandwidth allocation control. Also to use it more competently. Ethernet is also a good transport for server cluster traffic and storage.

New Ethernet protocols are developed or enhance the connectivity offered by Ethernet protocols by IETF or IEEE. Enhanced version of Ethernet are:

Data Center Ethernet or DCE – enhanced Ethernet based on the Data Center Bridging on the Data Center Bridging standards. Layer 2 multipathing implementation is also included. It is based on the IETF’s Transparent Interconnection of Lots of Links (TRILL0 proposal.

Converged Enhanced Ethernet or CEE – this generic term is used by many vendors like IBM, HP, Dell e.t.c. other vendors also refer it simply as enhanced Ethernet.

Priority-based Flow Control

Some of the famous Data Center Bridging Standards Task Groups are:

ETS orEnhanced Transmission Selection – common management framework for assignment of bandwidth to 802. 1p CoS based traffic classes.

Congestion Notifications – congestion management for protocols that are capable of transmission rate to avoid frame loss are Congestion Notification. Many protocols like TCP are benefited because it respond to congestion in more periodic manner.

DCBX or Data Center Bridging Capabilities Exchange Protocol – basically a exchange protocol it is used for transmission of configurations between neighbors in order to ensure reliable configuration across the network.

PFC or Priority-based Flow Control – main motive of this mechanism is to make sure under congestion in DCB networks loss is zero. It provides a link level flow control mechanism that can be controlled separately for each Class of Service CoS.

DataCenter Ethernet

Tags: , ,

Network Operations Center

A single place or many places from where computers, servers, telecommunication network or television broadcast are controlled is called NOC (Network Operation Center). It is pronounced as “knock”. Huge companies usually operates one or more NOC in order to provide geographic redundancy or Managing different networks in the case if one of the site is offline or unavailable. Certain dangerous or alarming conditions which can affect the networks performance on a large scale are monitored by NOC. Like NOC monitors communication line alarms because of various errors like bit errors, line coding errors, circuit downs and farming errors. NOC also monitor any possible power failures. Not just monitoring problems NOCs also analyze problems, carry out troubleshooting and communicate with other technicians. If the problems is too heavy, like optical fiber cable cut and power failure NCOs have to contact the technicians to solve the problem.

NOC Network Operations Center

NOCs normally solve issues in increasing order, if any case if the issue is not solved in a given span of time, next level is informed to accelerate and work faster. In order to rank the staff working in NOCs they commonly have multiple “tiers”. These tiers often describe how skilled a technicians is. “Tier 1” includes newly hired technicians and technicians working for a long time of period are placed in “tier 3” or “tire 4”. Like this many problems are solved without any help from the site technicians. Often it is seen that most of the servers of few of the servers are placed within the NOC which are necessary in order to running network. However it is impossible for a single NOC to control and monitor a number of geographically dispersed sites. So just like the other important parts of a data center like cooling devices, power sector, server room e.t.c. Network Operation Center is also important to the Data Center. In simple words Data Center is managed from Network Operation Center.

NOC 2

Tags: , ,

Data Centers and Colocation

Since a Data Center works as a heart of network service delivery it requires best of best technologies available in the market. All the hardware and software technologies used in the Data Center should ensure the customers 100% availability and scalability at the best possible lowest cost. As we all know in colocation all the servers are placed off-site mostly in the premises of service provider. One benefit your company can take pleasure is that colocation permits your network to stay up and running without disruption. Mostly colocation providers are judged by the availability of the services at their Data Center. A good colocation provider is the one who offers multiple active power and cooling delivery options, constant maintainability, system redundancies, enhanced physical security and fault leniency.

Different types of charges included by colocation providers are:

  • Bandwidth required for network connectivity
  • Physical space required to house all the servers
  • Category and quantity of power required
  • Quantity of cooling required
  • Amount if UPS required
  • Level of  physical security

Other than the above given services many colocation providers also offer managed services with benefits of patch management services, monitoring, configuration and data backup services, for each of these services however they charge extra bucks. Hosting providers is the most related business to colocation. Hosting providers use all the above given services as well as the servers on which they are hosted on. However this different from colocation where you have full control over the server, since you own the server. Hosting is less expensive, but colocation is more flexible because you can expand what is provided. Colocation plays an important part in today’s modern IT infrastructure. Colocation is a good option in terms of cost, but if your company needs larger pace requirements then its not a wise option.

Tags: , ,

Why you need a Data Center?

In today’s jet age, data is saved on the servers instead paper documentation. All companies save their data on web servers save their precious data on web servers, including customer records and other valuable information. Server on which your data is saved needs intense care and continuous monitoring, which is only possible if they are placed in a Data Center. Data center will help you to enjoy all the advantages of the servers. Multiple servers in a data center won’t act as a cluster but definitely help you by working in a group and in any worst conditions if one of your server fails other will work as a backup. So the first advantage is perfect availability.

Data Center-1 (2)

Now secondly recoveries are fast is you are working with a Data Center, hence your business will  not suffer any loss if server fails or any other unexpected problems occurs. Business continuity is another benefit of data center. Data Centers are extremely scalable and easily accommodating all the growing needs of your business. Chances of any possible glitch are rare with Data Center. It is not necessary that all the organizations and business a Data Center, organization with only one or two servers don’t need a full data center. Only growing and large organizations need data centers. Managing a Data Center is not an easy job, there are many kind of servers to look after and to be managed. One good solution is to outsource the management of the Data Center to some professional group. This will help to concentrate on your core business without any disturbance. You can even keep the data center in the premises of the service provider. This will help you in reducing the over all cost of the server

Data Center

Tags: , ,

Ideal Floor Plan for Data Center

Ideal Floor for Data Center-2

Data Centers Floor plans are as much important as other important factors of a Data Centers like cooling and power supply. Floor plans requires planned placement of all the computer equipments in order to optimize cooling and air flow without any ill effects on the working environment, still providing utmost scalability and accessibility. The process of designing floor plans for a data center needs important planning and  evaluation of elevation and top down methodologies. First important step is pre-design and planning for creating an competent floor data center floor plan. Investing time and money at this stage can save you a lot of bucks and time on fixing your mistakes. First real step is shaping a project budget for support infrastructure equipment including electrical rooms, mechanical rooms and computer rooms. Next if the plan is for an existing data center then an on-site survey of the facility with examination of the future growth prediction. Real construction work begins only after confirming these steps. Two important Data Center Floor Plan Methodologies are

Elevation View

There are three levels in Data Center floor plan they are: The layout above raised floor, layout below raise floor and layout above dropped ceiling. To set up Data Center floor designers begin with dividing the floor. An important step at this level is to include a sub-floor with efficient drainage system to remove some possible leaks from sensible equipments or water. Cooling pipes and Power Cables ought to come through the sub-floor. Data cables are installed at the top of the raised floor. There should be some free space in the middle of the raised floor to permit unhindered air flow. Equipment should be placed in such a way that it should improve support quality and performance. Most common way to do so is applying the Hot aisle/cold aisle layer. Ceiling height of most of the Data Centers is 9 feet and all the equipment should have at least 18 inches of clearance.

Ideal Floor for Data Center

Next layout is above the dropped ceiling. Plumping and water should be strictly banned from passing over the room. Most suggestible is to avoid these pipes in the building and place them somewhere in the perimeter. This will reduce the possibility of water leak which can done serious damage to the computer equipments. Dropped ceilings presents an optional cooling by allowing the air to flow through ceiling plenum based on the requirements of the cooling.

Top Down View

Before starting work on top down view of a Data Center, designers have to examine the equipment location, deciding equipment size, essential space area and work flow process. It also involves analysis using CFD (Computational Fluid Dynamic) modeling. This process is performed for the area above the floor as well under-floor area. CFD allows designers to take decisions about where to place IT equipment, perforated tiles, rack location units, CRAC equipment e.t.c. There are many software available in the market which helps in concluding both under and overflow airflow obstacles. CFD analysis also allows you to find which cooling zones should be used for high density, medium density and low density equipment. Once all the above steps are complete you can head towards the hot aisle/ cold aisle step. In the hot and cold aisle layout cold air is delivered using pierced tiles in the cold aisle. When heat is generated in the hot aisle it uses the ceiling height as a way to back to the CRAC units. If the heat generated is high it can be cooled down by using a new kind of CRAC units commonly known as “in row’ system. Design teams also need to find out specific cooling and powers zones. There may be some zones which are in the low-density RLU zones which only want raised floor cooling, and there may be some other zones which needs medium-density cooling options. For ultra high-density server, there must be a cabinet cooling solution. Before drafting a plan for Data Center, there are few technical and strategic issues to be worried about. For best scalability and availability hire an expert.

Ideal floor plan for Data center

Tags: , ,

Blade Servers

Blade Server

There are many kind of servers in Data Centers, they are differentiated according to their arrangement and their size. One of the most famous and known servers are Blade Servers. Blade Servers are design in such a way that they occupy minimum space. Unlike other rack-mount server which have many heavy parts including network cord and power cord, Blade server are stripped of anything unwanted leaving only the most useful functional parts. In a blade server enclosure it can hold cooling, network, power and many other interconnects. Blade server also differ from each other on the question of what should be included in the server enclosure. Blade server are good for normal hosting as well as Cloud Computing. Blade servers are hot swappable, i.e. if you put more pressure on them like adding more users more processing power it will work with more diverse workload. 1RU (one rack unit)  is the minimum possible size of any equipment in a standard server-rack configuration. When relating Cloud Computing to Blade Servers biggest advantage is lifting the restriction of the least size requirements. 42U is the most common size of computer rack and in restricts computer devices in a rack with only 42 components. Blade servers on the other hand don’t have any such limitation and can hold up to 128 components per server rack. Only two things are required by computer server that are:

  • Memory for reading data and input commands
  • Processor in order to control the data
  • Data storage program

All the above requirements are necessary if a computer is working real hard. However in Blade Servers most of these are removed, they are either virtualized, with the help of remote console over IP or iSCSI storage, or they are included in the Blade enclosure and lastly they are completely discarded. As a result Blade Servers becomes infinitely easy and cheaper and smaller for the manufacturer. Many of the non-core computing services which are found in most of the computers, are performed by the enclosure of the Blade servers. Requirements of the non-blade servers is vast and looks useless, like spacious, bulky and inefficient components. Power supply to the blade enclosure is provided by a single power source to all the blades in the enclosure. Single power supply is mostly a  dedicated separated PSU supplying DC to the multiple enclosures. A number of UPS can be reduced by applying the dedicated UPSsystem, which are otherwise required for the flexible power supply. Because of the popularity of the Blade servers demand for rack mounted UPS has gone considerably up.

Blade Server-2

Tags: , , ,

Pros and Cons of Colocation

Colocation anew type of hosting service allows companies to buy rack space in Data Centers. There they store their own server equipment. Facilities provided by these Data Centers is purposely built for Colocation hosting services. By following this way companies can have their own servers without placing them on their won campus. There are many customers placed in these Data Centers, but each one of them have their own server and space. Colocation have many advantages but there are few backdrops also when compared to regular or basic web hosting. Going against normal hosting in Colocation hosting you owes the server and have full access and control over the server. You can also upgrade it if you want it to be. Mechanical and power systems are much more powerful and so speed of the speed can be much faster. Since your server are placed somewhere far from office site, internal networks are freed from the extra load and can be used for some other purpose. Best advantage however for Colocation hosting is Bandwidth. For Bandwidth is cheaper as it shared between many clients. Backup protection of Colocation centers, even in the situation of some natural disaster or other threat your website can continue to run. Colocation is best suited for small business as they can have the features of large IT department without spending much. Some Colocation centers even charge more for maintenance and management of the servers.

Now everything has darker side, with Colocation as there are advantages there are disadvantages too. Although Colocation is cheaper then having your own server, but it is too costly when compared to normal hosting. Therefore many people consider basic web hosting as best is they are provided with all the requirements. Also initially you have to a lot of money to the Data Center for server. And even after paying for the server you still have to pay for the Bandwidth, mostly on the monthly basis. Cost of the bandwidth fluctuates as per the usage. Of large amount of Data is transferred during the month then it will cost you more. As the rate fluctuate month by month you will find it difficult to draw an average scale. Whenever your equipment needs up- gradation you will have to pay more. So there are both plus and negative points in colocation servers and it depends on you what to choose according to the needs of your business.

Tags: , ,