Server consolidation and data center can move

fournirdes significant benefits – including savings, business continuity renforcéela, optimized améliorationconformité services and regulation.

The impact of this physical displacement should not be overestimated êtresous. If you do not understand and addressi problems occur when you put more distance physiqueentre users and servers, you can sit for gravesla pain and possible failure.

Here are four killer mistakes that you should be especially careful to avoid:

1) Confusing network latency with a latency déplacementvos If application servers does not add 50 milliseconds of network latency, Quena, that your reaction time of neaugmenter 50 milliseconds. Instead applicationsnécessitent back and forth a lot of interaction between the user and the server (often called the application “is considered). Thus, the additional network latency milliseconds ED50 cause an action, the NEA took 3 seconds to stop at the local level take 30 seconds pourcomplète after changing the server.

Network Manager can not the speed of light, or make Tokyo Plus in the vicinity of New York. So it makes no sense to fix it problèmeentièrement. sontsouvent In fact, because the issues of application design responsible for poor response times to change a server, additional investments in the network of small utilitéque will ever be.

2 absent) to see how network latency distribute serveurServeurs impact performance and scalability to resources and customers chaquesession blocked until the local session terminée.clients their sessions end quickly because leurapplication transforms experience minimal delay in connection with distance réseau.sessions take much longer.

For example, to remote user is busy on the conservation of resources of the server for more longuepériode – preventing the release sesressources server for use by other customers and limit the ability to scale. Therefore, it should envisagerla organizations the opportunity to network latency, they are investirdans additional server infrastructure.

3) Set the distance between the servers – even temporarily – to destroy peutperformance It may take weeks or months coupvers a new position. During the stages, some original systèmesopérer of their locations, while others fonctionnentpartir the new position.

The impact of this separation between the servers on requestThe power can still dramatic and unexpected that the introduction of the latency between user and server, carprocessus calculation are almost never designed to accueillirsignificative inter-server latency.

Consider the following example of an application créditauthentifie users on a Active Directory (AD) and serveuraccède a database to check the creditworthiness of customers:

1 Client accesses the server application for credit (10 laps) 2.Request credit client-server authentication using the AD server (50 laps) 3 Application server receives credit information from credit deserveur Database (200 laps) 4 The customer gets an answer laserveur use of loans (5 laps)

Note that it does not apply, the number of rounds between Clientête demand for credit by the end, but there are many viragesentre servers. So if estprésenté the distance between these servers during a movement, it is likely to be a sévèresur impact application performance.

Every IT organization is planning to ask undéplacer data center so that a variety of questions.Qu what happens when servers are disconnected sonttemporairement with critical dependencies? What are the server with other servers must be moved? When are moved to Active Directory server? Quelsles servers must be replicated for the duration of the move?

4) does not match the users’ performance expectations qu’aprèsle movement around Sometimes it does not make sense, unaprès Relocation Service Level Objective (SLO) is the identiqueà fix as before the move. If done originally by a local user trois seconds to perform a task, it is very unlikely that lestâche have the same amount of time after déplacésà servers across the country. Thus, a SLO seven seconds, for example, is perhaps more reasonable.

It is therefore necessarily to the level of service utilisateursattentes front address. If, after the move and direutilisateurs they need with what you can offer live you wait for the floor you are setting up a battle. But if you’re in the process of planning, you can adhésionpréalable avoid telstracas and to ensure that no unrealistic expectations. reach

To this aim acceptance pre-move, it must be able to simulate the performance deprévoir and post-move. Cescapacités forecasting and simulation for development of IT “environment the acceptance, “where users can post profiterperformance resettlement before the move is actually executed, and buy where they can post-ALS pass is installed a seulserveur.

seven steps to successful IT project organization can avoid all these mistakes lesde . But do they have to build laexpertise team of application, system managers and réseauarchitectes you also need to be able models virtuelsà both pre-and post-business environment to provide resettlement -. commeainsi that all stages of transition organizations can avoid these errors. But to do this, they must rely on the expertise of the application, system managers and network architects. You must also être, virtual models of the two pre-and environment of the company after the . to create moving – all lesphases transition

shows the following seven-step plan on how etmodélisation cooperation can ensure a successful data center move: is

1, create a virtual model of the pre-and. Post réinstallationenvironnement-prévuephases business and any transitional arrangements. All participants in the planning, need including includedThe professionals specific information about réseauinfrastructures the performance of applications with data nouveaucentre impact.

2 Establish a baseline by measuring the SLO before demandeperformance the move. pasexister user needs and expectations in a vacuum. fournircontexte pass the decision prior to the transaction response to the determination of appropriate SLA for ladéplacer.

3 Measure the performance of applications after moving into an environment virtuelenvironnement. The only way to accurately predict the impact Dese moves on server application performance in a fully simulated lesapplications run after the move. Cettefournira data on the performance potentielledégradations essential for good planning.

4 applications, the Identify Performance Tuning must spéciales.Plutôt a waste of time, effort and money on all battles jusqu’àéléments business infrastructure and instead focus on application-specific or network components can êtreparticulièrement problematic.

5 Assess the dependencies between the back-end server to appropriate flatness away. analyze addition to meeting réseaulatences between clients and servers, the team must pleinementcomprendre the impact of latency, which can be generated between lesserveurs during the transitional period phases of the movement.

6 problems and to validate fixes pasapplications not possible. Before investing in the provision of a solution, and it estimportant ensure that it really works. surbande assumptions bandwidth or processor power should always be tested before êtreexécutée in the production environment.

7 Managing Users expectations and obligations faveuramélioration to obtain the membership of the investment performance through the hands of acceptance. Users should be able to directementles application performance experience post-removal in advance, as qu’ilspeuvent informed consent provide a view on the relocation.

Through following this seven-step plan peuventsensiblement to reduce IT organizations risk to eliminate inutilesdépenses infrastructure to accelerate delivery times and overcome a wide political éventaildes pitfalls. So if you are planning a serveurconsolidation or other types of data center move, examinerinvestir in simulation technology, with which you expérimenteravec alternative scenarios and determine in advance what and what not vatravail This is an excellent way to ensure that your company fully use the movement -.. without its potentially catastrophic consequences souffrirl’un

For these and other Articles, network performance, white papers, industry and resources, please visit http:// Shunra at www.shunra.com / resource_center.aspx.

About Shunra Shunra solutions enable companies deniveau service and performance concerns Shunra VE déploiement.La solution creates an exact copy of the laenvironnement production network so that all security professionals to develop, test and test applications before deployment etplan infrastructure, effective and etcroissance change. Designed for networks etprofessionnels proficiency tests and software developers who Shunra VEfacilite cooperation between the disciplines of IT organizations to quickly and efficiently discover ITles etrésoudre problems before they impact business. 1.500principales More companies and technology providers in the world award-winning Shunra utilisentLes solutions, including 3M, Boeing, Cisco, Dow Chemical, EMC , FedEx, General Electric, General Motors, JPMorgan Chase, Kelly Services, Merrill Lynch, Motorola, Nestle Pitney Bowes, and Vodafone. Shunra New York headquarters and in Kfar Saba, Israel, with subsidiaries entierle United Kingdom, Sweden and India. Shunra is by uneréseau distributors worldwide support.


Moving a data center