Archive for December, 2009

Disaster Recovery Plan for Your SaaS providers

Software business started fluctuating when SaaS (software as a service) start gaining popularity. Established companies who were ruling the globe are now facing solid competition, because customer now wants more economical and easier solutions. By signing up for SaaS solutions customers enjoy a position from where they can cut a lot of cost including that associated with expenditures and applications can be delivered with less risk and lower cost. Problems arises when SaaS providers get over confident and the necessary procedures are neglected. For example system dependability and security levels can’t be compromised. Companies stop paying attention on issues like business continuity and in its place they start depending on their SaaS vendor to solve their problems. We all know SaaS is the easiest way to deliver applications at lower costs and with no or less risk. Some SaaS vendors are promising that they will eradicate the IT completely and thus making the problem of business continuity more serious. SaaS vendors are reliable but they are considered as a single point of commercial control. If you look back in history single point of control often ends in single point of failure.

One best example to understand this point is of search engine king Google. Many Google users find it unable to access their gmail accounts. Many called it a minor fault, but in reality a giant companies failures were shown to the world. You can learn from this incident that what can happen when a company completely depends on a SaaS provider specially for some business critical function. Just think for a second not having mail access for minutes or worse for hours. Now more and more companies are considering to use SaaS solutions, it is important to have a sure disaster recovery plan. Redundant servers and back up for data center are standard features of SaaS vendors. SaaS vendors also concentrate on providing best possible downtime. Even after these services future not guaranteed, so its better to check the disaster recovery plan of the SaaS vendor. Disaster recovery plan will make sure that your business is safe from all manmade or natural disasters.

Automation in Data Centers

In early processor era Data Centers was intended to maintain single centralized platform with only a single operating system. These Data Centers were usually related to only few applications. In those days computer workloads were well understood and well controlled, irrespective of the fact that Data center management was complicated. Those days are history now and today’s data centers are now have countless servers which are running n-number of applications with various operating systems. This complexity in managing data center is dangerous and in many organizations it the level has fallen to a risky level. The ability of data centers managing complex computer workload is falling day by day. Adding to the woes the observing bodies now demands that whole IT operations of an organizations should be monitored more strictly. In a Data Center each supportive function is separate from other supportive function inside an IT department and they do not coordinate with each other. Just for an example the group taking care of data management does not coordinate its activities with network management. This lack of incorporation can be blamed on the uncontrolled growth of the industry. When in the initial stage Data Centers were never expected to grow such huge. Even many Data Centers are experiencing lack of space. If you go through the below given points it will be clear why a volatile solution is needed:

  • Efficiency is required in the use of resources and power consumption.
  • Virtualization techniques should be implemented to facilitate resource productivity.
  • Observation and security of multiple servers.
  • Process management should be changed in contrast to the industry needs.
  • Business process which are circulated in highly distributed environments should be encouraged.

In this age business world has became more and more active then before. Often an organization is seen with a network which is spread amongst its suppliers, partners and customers an that needs a perfect management. With time as companies will depend on these networks it is crucial to manage the networks associated in a active and expected style. Though requirements for different business are different but what every business expects from the Data Center is all the same, like excellent level of the staff  and expected costs. On the first impression it looks simple but it actually gets complicated when new connections are demanded and  new applications are added on the demand for new suppliers, customers and partners. Those business which fails to provide new technologies soon get obsolete from the demand.

Tags: ,

Best Way to Keep Data Centers Cool

Equipments in Data Centers work round the clock let it be day or night, hot or cold, data centers works continuously. Huge power is used in order to keep data center working. Main part of that power is used in keeping that equipments cool, because continuous working can cause overheating and eventually burning down of the equipment. Not few but most of the data centers waste their power in their cooling systems. Many research are being conducted by scientist round the globe to find out the solution. So far the best solution is shown by a team of engineers led by The Lawrence Berkeley National Laboratory. This whole project is backed up by company giants like IBM, Intel Corporation, HP they all are funding the team to bring new system which can cool the data center without using much power. Till now the team has find out a system that delivers the most favorable amount of cooling to the equipment. The system is not a complex machine meshed up in infinite wires and toughest looking programming codes but it works as simple as a pendulum. Simple first take temperatures readings from the servers which are feed into the Data Center Building controls through sensors. As a result air conditioning system keeps the facility cool on its own. The team wrote a software which was used to fill the gap between management systems. Previously these systems were running separately for facilities management and IT. Typically these systems were kept separate from each other in most of the companies. Not the team of scientists connected the two systems and now the companies are reaping huge profits. About 80% of the Data Centers have their temperature sensors on their computer air control systems better known as CRAH units. 11% of the data centers are following a better way then the previous one and placing their temperature sensors in the aisles between server racks. The best way is to link the servers up directly into the cooling system itself. As promised by the team their research will be ready commercially within a year. One strange fact in this research is that nearly 90% of the Data Centers are keeping their temperature down by 5 degrees cooler than required.

Tags: , ,

Trends in Data Center

Companies need to enhance their technology investments, specially because of the economic ambiguity. Every industry was affected because of the economic recession.  Data Centers solutions providers were also affected and they are under pressure to preserve the levels of reliability, security, flexibility and efficiency and that to without increasing any cost even in some cases cost was asked to be reduced. To cover increasing demands from business Data Centers should be modernized if they want to harvest future rewards. Some of the best modernization available in market are automation, blade server and virtualization. In recent times two other data center modernization which are in limelight are utility computing and infrastructure as a service. Ass according to the industry experts in the upcoming year there will be shortage of cost cutting proposals. There is a excitement around that “cloud computing” is the best solution for cost cutting. Cloud is more a new kind of technology which fetches new worth to the enterprise, then being a cost cutting solution for the existing services and enterprise apps.

First step to be followed for the Data Center is to think like a Cloud-based service provider. Start using technologies which are used to boost the performance of Cloud computing. Also put into practice the service which will make the infrastructure more supple, so that resources are used to make the infrastructure to work like self-service instead of build to order. Sharp focus on right technologies and change of mind status are two requirements which are needed to increase value of the existing resources. Its true that virtualization is important, but its not universal remedy. There are many steps which are needed to be followed in a data center if someone wants their business to grow on normal rate in this economical roller coaster ride.

Data Center Trends Trends in Data Center
Automation of Server, Network and Storage Devices

Virtualization helps in enabling technologies for organizations to convey new hierarchal levels in efficiency to the data center. However disadvantage of virtualization is that it blurs the contact between the clients and server, storage and network devices. This happens because virtualization permit multiple resources to be configured on a single CPU. If you attempt to create interoperability and uniformed management it can be challenging to you. For this reason automation is gaining more importance then virtualization. In a Data Center where server management is automated it makes virtualization inappropriate to be used. In the data center where servers are automated, companies can enjoy the advantages of fast provision applications as well as that of virtualization and all this without taking any risk. Virtualization is preferred by most of the users and companies have to manage both virtual as well physical servers. Sometime they use different kind of tools for this process. Standard management is required on the tools which are required to manage virtual and physical servers. Also the changes should be monitored to find out the impact of change. Virtual machines need to be managed just like other technological machines. This is important to in order to avoid any possible performance problems along with service and security issues.

Blades Servers are the Best

We already discussed how Blade Servers are best server types used in contemporary Data Centers. All the important tasks in a data center are easily covered by Blade Servers, including supporting most important computing to workplace application. New delivery models like utility commuting and infrastructure as a service are also easily execute on Blade servers. Mostly customers look for change, energy saving, cost and time benefiting hardware and ultimate solution for them is Blade Servers. The technology behind Blade Servers is still growing and is expected to grow more mature in the year 2010. With time Blade server are expected to gain perfection from best option. They will be able to reduce data center footprints and also increase energy efficiencies. Nowadays server providers are also developing special blades for special applications. Say for an example cloud computing’s special requirement is an ultra dense, which can execute large amount of data swiftly and competently. On the other hand virtual organization requires amplified storage, data storage, network connections and memory. This is not a problem at all for Blade servers to run these applications sufficiently. Still special blade servers with modified capabilities are equipped to optimize the functionality of a server for a given application.

Data Center Trends 1 Trends in Data Center

Tags: , , ,

Rightsizing Servers

Commonly people consider performance with speed. If go deeply and find the meaning of the word performance in dictionary it will be defined as how good something performs the functions or actions for which its is assigned. When you are dealing with servers in Data Center, you should care for more than speed. Traditionally speed is measured according to latency and response time. Cost effective measures are also important. To understand this more fact take an example of automobiles or cars, there will be different customers. One who are interested in mileage will not care for speed of the car. What they want from their car is maintainability and all other criteria related to economy. Going the same way when you look at servers you care more for performance per watt, performance per dollar per watt and performance per dollar. Data center admins discover the definite energy consumption and how much work is done with that energy.

Look at the performance of your CPU systems in real world. Mostly the memory sizes which are produced by used benchmarks is smaller then the amount of limited disk I/O. This is the reason why CPU utilization in the industry is so low and thus the balance of server systems are not well. Solution for the problem includes using balanced servers which use main platform resources like memory, disk, network and CPU in a right manner. Another option to solve the problem is to use fewer memory in order to have fewer disk admissions. Although this process adds a little cost to you but it can help you to improve the performance to a great level. Performance defines cost effectiveness of your business process. Along with performance and cost consider power too. Working of benchmarks may get affected by the working environments of your company. So it is suggested that these companies create their own workload process. These process will be more effective in finding out the activities of the applications used by your servers.

Tags: , , , ,

Challenges in Operating Data Center

Data center are important for every kind of business. Saving your data keeps your companies operating and nay downtime can cost you a lot of time and money. Maintaining a data center, running 24X7 is a big challenge and the companies which provide data center solution face many problems. Early when computer era started dedicated rooms were created to keep computer equipments safe, cool and free from other contaminants. Only these dedicated rooms later evolved into Data Centers. As the computer era was developing very swiftly, various departments congregated these dedicated rooms to form earliest data centers. With dot com bubble client-server networking become significant for IT companies and the dedicated rooms were changed to higher designs. Now these dedicated rooms are grown up into fully managed Data Centers, and maintaining these is not easy and include several challenges like:

Operating cost of Data Center

Cost of operating Data center is very high and companies often got shut down due to this factor. Companying a large data center with all the essential equipments is not a small job. All the necessary equipments and servers are expensive. Companies providing data center solutions have to pay employees who are skilled in maintaining servers in Data center. Also those employees have to work round the day. Getting employees for a data center is very costly business. Solution for this problem is that instead of buying the equipment and hiring all the staff on your own, companies can use the service of some other data center providing their services. This process helped many companies to decrease their cost of operating with providing good service.

Physical Space

Early data centers were small not more than a room or few rooms, but as the industry grown larger and larger seize of Data center grew considerably and nowadays they cover acres of land. It is hard to manage the servers, important equipments and cables in a Data Center. Finding space for the huge data center is a big problem for many companies. You can outsource data center from other companies, instead of building dedicated space for large data center.

Reliability and Security Dilemma

Now even if the companies pass the challenges of staffing, equipment expenses, and physical space, companies still have to face the challenge of providing security and reliability. Reliability of a company is directly related to the reliability of Data Center and if the data center goes down for any possible reason it will affect the companies business. Security and reliability is only possible if the staff is properly trained and they security is at the right place. Another factor for good reliability of a data center is disaster recovery plan. Disaster recovery plan is the reason for the coming back of a company.

Tags: , , , ,

Before Choosing a Colocation Facility

In the jet age communication and data access are two most important factors for business operations. Email, messaging websites and servers have to be up 24×7 and they should be protected for business communication to function. Specially Data Centers which have servers with mission critical information. Most important are company databases and servers, which should be kept up at any cost. Important factors in a colocation facility are:

Redundant Systems
Systems used in colocation hosting should be redundant and at every point of working they should have proper capacity. Many people get confused with multiple units considering them as redundancy. Ideal configuration for systems used in colocation facilities should be n+1, which means that if there are two units they should run at 50% capacity, and in any case if one fails other will handle the load. Same if there are three units working capacity should be 66%.

Colocation facility Before Choosing a Colocation Facility

Best Connectivity

Major points which decides the Internet Connectivity of a colocation facility are as follows:

  • Total Bandwidth available for usage
  • Accessible Internet backbones
  • Uptime guarantee
  • Packet loss, jitter and Latency

Carrier-neutral access and Tier -1 carriers, tier-2 carriers all get access to internet using Colocation data centers. Carrier-neutral can be used by any service from any carrier but customers are requested to configure their own routing. Also customers are should maintain their own connections to Internet backbones. Tier-1 is used by Internet giants like AT&T for direct access. Tier-2 is used for routing traffic amongst the multiple Tier-1 providers. In these kind of tiers speed and reliability depends on how effectively traffic is routed. Tier-2 is preferred by most of the clients since routing is configured by the Colocation facility. Also it is more reliable because of the presence of backbones, these backbones network performance and uptime.

Colocation facility 1 Before Choosing a Colocation Facility

Control on Climate

Design of the colocation facility itself affects the performance of the systems used in the colocation facility. Air flow between rack rows should be controlled by the server rooms. Hot air generated due to the heat produced by the servers should be blown in one direction and cool air to keep the servers cool should exit from another direction. Only dedicated hot and cols air flow can save the servers from overheating and finally burning down. Cold sir and humidity is controlled by correctly sized and redundant climate control units. As we all know climate requirements for the server are strict and they only work properly when they are installed under suitable working environment. Servers need 50% humidity and 70 degrees of temperature. Both CRAC units and chillers for the server rooms should be included in a Colocation facility. Cooling capacity is calculated by total tonnage by the square footage.

Power Required

Power to the colocation is like oxygen to the humans, if it fails to perform then you are in deep trouble. UPS and generators should be used for backup and they are necessary to keep the network online. UPS are used when power is shifted from regular to the generators. Generators should be there 24X7 for immediate backup power. Power system of a colocation facility should be properly working and should have following features for perfection:

  • Redundant and multiple UPS because of the reason UPS fails very oftenly.
  • Generator which can handle at least 1.5 time the total building load

Equal Distribution of Traffic

Both hardware and logic of distribution has an important effect on connectivity. The way traffic is divided plays an important role in the performance of routing. If your distribution covers three below given points than it is perfect:

  • In order to continually adjusting routes to find the most effective one Performance based distribution is should be used. Performance based concentrates on performance rather then price.
  • Switches and routers should be redundant and be working on full working for the customers.
  • Distributions should be done between more than 4-5 clients.

Colocation facility 2 Before Choosing a Colocation Facility

Tags: , ,

Enjoying Virtualization

For most of the people server virtualization sounds something complex, but if you will go through it you will find it much more easier. Virtualization is considered as one of the mainstream technology and it is the popular solution for data center management for every type of company. Understanding the basics server virtualization allows you to run multiple guest  operating systems which can be either same or different on a host operating system on a physical server. Operating systems are divided in order to reside each application on its own virtual machines or virtual server on the hardware platform. Benefits associated with virtualization are endless and it is specially very much good for small business. Most important benefit is lowering the cost for the owner and it also includes many smaller benefits like:

  • Improve physical space in Data center and also reduce server sprawl.
  • Hardware maintenance cost is reduced since number of physical servers are  also reduced.
  • Speed up the network thus making provision for new services.
  • Performance and resource utilization is improved.
  • Applications are quickly developed and organized and downtime is avoided due to maintenance windows.
  • Flawlessly mange and scale the infrastructure based on various industrial demands.

Virtualization 1 Enjoying Virtualization

There are many software which helps in organizing server virtualization. Some of the most famous and most conveniently applied virtualization servers are:

Windows Virtual Server

Windows virtual server is loved by those who only want to have Microsoft type of environment, using Windows XP or Windows Server 2003 operating systems. There are two versions provided by Microsoft for its Windows Virtual Server 2005 Standard and Enterprise Edition. Enterprise Edition supports can host a server with as much as 32 processors. Standard on the other hand support server with only 4 processors. The enterprise edition can be downloaded free of charge.

XenSource

It is launched almost three years ago. XenSource is based on a open source project sponsored by Linux and known as Xen. It is a good option for those who like open source products then proprietary ones. XenSource proved itself in cost effectiveness but it still has to prove itself in quality service.

VMWare

It is the leading software of the virtualization process. Launched in 1998 it was VMWare which bought virtualization to the mainstream through  X86 computing platform. There are option for everyone from small to big business and at the wide range of price points. Features of VMWare are never ending and it benefit’s the host for every server environment.

Virtualization 2 Enjoying Virtualization

Tags: , ,

Types of Cloud Computing

Types of Cloud Computing-3

Whenever IT always need to add capabilities or increase capacity without investing in a new infrastructure, licensing new software, or training new staff first word that came to your mind is Cloud Computing. Any subscription based or pay-per-use service over the Internet is included in Cloud Computing, it helps in extending IT limits. You may have come across various analysts, IT customers and vendors talking cloud hosting and its features. Based on the various discussions here is in rough what really means:

Utility Computing

A comparatively new but well acquired form of Cloud Computing. It is major support is from industry giants like Sun, IBM and Amazon.com, and various others who are offering virtual and storage servers which can be accessed by clients on demand. Newbie’s use Cloud Computing for non-mission critical, supplemental needs but it will be soon be replaced parts of data center. Other providers are offering solutions that will help in creating virtual data centers from commodity servers. These virtualized servers enable users to stitch up storage, I/O computational capacity and memory.

SaaS (Software as a Service)

In this kind of Cloud Computing a single application is delivered through browsers to thousands of customers using a multi tenant architecture. From the customers point iof view it means no extra investment in software of server licensing. From the vendors side only a single application to maintain and costs are way low as compared to conventional hosting.

Types of Cloud Computing-2

Web Services in Cloud Computing

This is closed related to SaaS and Web service providers offers APIs. These APIs in place of delivering full-blown applications allows developers to develop functionality over the Internet. Full range of APIs offered are ADP payroll, Google Maps, Bloomberg, U.S. Postal Service and may other credit card processing companies. Small time API vendors include Xignite and Strike Iron.

Platform as a service

This is another variation of SaaS, it delivers development environments as a service. Application is build by you which will run on the provider’s infrastructure and later delivered to your users through Internet from the server of the provider. These services do not offer you complete freedom as they are controlled from vendors end, but instead they offer pre-integration and inevitability.

Managed Service Providers (MSP)

This is one of the oldest form of cloud computing. Managed service is basically an application which is directly exposed to IT Companies instead of end users. Best example is application monitoring service or virus scanning service for e-mail. Some of the most famous managed security services are delivered by IBM, SecureWorks and Verizon. Other famous offering include CenterBeam’s and Everdream’s desktop management services.

Commerce Platform Services

This is a hybrid of SaaS and MSP. Its basically work is to provide service hub that can be used by users to interact with each other. Most commonly they are used in trading business like expense management system that allows users to order secretarial services or travel from common platform and then organizing the service delivery and pricing inside the specifications set by the user. You can also call it as automated service bureau.

Types of Cloud Computing

Tags: , , , ,

What is Client Server?

Client Server

Client server networking or computing helps in partitioning tasks or distributing work load between service providers and service requester’s, who are known as clients. Client Server is a distributed application architecture. It is seen that frequently servers and clients operate on a computer network on separate hardware. Server machines which share its resources with its clients, are high-performance host running one or many server programs. On the other hand a client does not have to share any of its resources, but on the other hand requests a servers service function. As a result clients begin communication sessions with servers which listen to incoming requests. There are many advantages associated with client servers, like:

  • Data is stored on the servers, thus providing the clients a greater security than client machines. Also Servers control the access and resources in a more proper way only appropriate clients are allowed to access and change data.
  • Client server technologies are available for the use which were initially designed for friendliness and security of the user interface.
  • Can function with multiple clients of each with different capabilities.
  • Client server architecture allow the responsibilities of computing system to be spread amongst many independent computers which are connected to each other with network. It makes it more easy to be maintained, for an instance it is possible to upgrade, relocate or repair servers with clients remain unaffected and unaware of the process.
  • Update in the data is easier to control then the P2P pattern, since the data is centralized .

Everything has brighter and darker sides, client server with lost of advantages got a lot of disadvantages too like:

  • Client server lacks robustness of P2P network. Like under the circumstances of critical server fail clients requested cannot completed.
  • Traffic jamming on the network is one of the most common issue since the beginning of client server pattern. Server can become overload if client requests to a given server increases.

Client Server-1

Tags: , ,