Data Consolidation Phase 1: Study and Document the Current Environment

With the increased use of high-bandwidth, fiber-optic networks and cloud computing, more and more companies are looking to consolidate their data center operations. In his book Administering Data Centers Servers, Storage, and Voice Over IP, Kailash Jayaswal provides insight regarding the best practices to consolidate data center operations. Phase 1 of data center consolidation consists of studying and documenting the current environment, which is discussed in this section.

[Phase 1: Study and Document the Current Environment] is the first and the most important stage. During this stage, you understand well as collect and document all technical information. What are the servers, storage, networking, and application requirements? How do these intercommunicate and how do they depend on each other? all facets of the IT environment that are targets for consolidations, as

This phase can be very detailed and time-consuming. Therefore, it is vital to decide on the amount of details required and then stick to the requirement. If the evaluation is cursory, then significant details will not be understood, which will later lead to service outages, user frustration, and possible failure of the project. At best, it would lead to expensive rearchitecture and failure to remain within the allocated time and budget. If too much data is collected, portions of the data would be irrelevant. It requires an in-depth technical and business know-how to judge the extent of information required for the entire project.

Evaluate Application Requirements

The purpose of this step is to understand the function of each application, as well as the flow of data from server to server and process to process. There are several information sources. The best source, however, is speaking with system administrators, application managers, and developers. Another way is to delve into existing documents, if available. Whereas third-party applications usually have documentation, home-grown processes and applications have very little. In such a situation, you are left by yourself to use OS and application tools to collect requirements.

Following are the important tasks that help you in understanding applications:

  • Data flow — Document all software components, their inter-dependencies, and how they communicate with other servers or applications. Figure 10-1 shows an example of data flow. The front-end Web servers receive requests from the end-users through a load-balancer. The Web servers communicate with the application servers, which, in turn, talk to the database servers.

    Besides the data flow for providing services, you must understand the communications required for administrative upkeep (such as access from central login or gateway servers, monitoring stations, and backup servers). At some stage, you will wonder if you should dig deeper for more information. The right question to ask yourself is, "Do I have enough to rearchitect the environment?" For simple physical consolidations, details are unnecessary. For logical application consolidations, details about port numbers, network access control lists (ACLs), and intervening routers and load balancers are required. Data Center Consolidation - Data Flow in an Environment

  • Interview architects, developers, and system administrators (SAs) — The objective of speaking with these people is to get details on configuration and server dependencies for the next two phases (Phase 2 (Design) and Phase 3 (Implementation)). Various areas must be covered, such as applications release levels, operations, code development, service level agreements, previously determined downtime windows, user expectations on uptime, testing, planned software and hardware upgrades, and expected increase in user load. Ensure that you understand and document at all points. If, after the meeting, you come out thinking, "I do not know what he or she was trying to explain," the meeting was a waste of time.

Evaluate Hardware Infrastructure

Over the last few years, the demand for more and more storage has far exceeded the demand for more servers or network bandwidth. Adding storage to existing servers was least disruptive and, consequently, it was added whenever needs arose. Because of all the factors listed earlier in this chapter, storage consolidation has become a key project in most data centers.

While assessing storage requirements for the subsequent phases of consolidation, you must document the following:

  • Storage hardware — What is the quantity of direct-attached, SAN, and NAS storage for each server? Are the logical units (LUNs) or devices configured as hardware-RAID volumes? Is there cache on the storage subsystem? Is the data replicated to a remote site for disaster recovery, and, if so, how is it copied over and how frequently?
  • Logical configuration — What are the volumes set up? Are they RAID or logical volume management (LVM)–based volumes? Are the volumes used raw, or are they file systems?
  • Data — What is the type and quantity of data that resides on the storage devices? What percentage of the storage is actually used up by data? Is the data copied from another server, and, if so, how frequently is it refreshed?

Like server data collection, storage data gathering can be automated. There are also many third-party applications for this purpose. They provide the percentage used, usage trends, predicted data usage (based on historical trends), and data categorization by size, type, or access data.

Evaluate Networking Requirements

Network assessment is necessary to make provisions for enough bandwidth in the new architecture. The network is used not only for server communication but also for NAS and for storage over IP (SoIP). There are two important areas within network evaluation: setup and performance.

  • Setup — This refers to the LAN topology, link speeds, and technologies used for the network such as 100baseT Ethernet, gigabit Ethernet, fiber distributed data interface (FDDI) and asynchronous transfer mode (ATM). For WAN topologies, the setup comprises a number of T1, E1, frame relay, and integrated services digital network (ISDN) links, as well as pattern and amount of usage.
  • Performance — Network performance is the amount of network bandwidth that is actually used by the traffic. You must be aware of the maximum throughput of the links, the percentage utilization at different times of the day, and the allowable network latency for the users.

If you are consolidating several servers within a single data center, you do not have to worry about anything more than the client-server LAN configuration. But if you are consolidating servers currently located at different data centers around the world, you must document the current link speeds between servers and users and provide for enough throughput between the new consolidated server(s) and users around the world.

It is also critical that the network reliability and speed from the consolidated server to administrative networks (such as NAS and backup networks) are not negatively impacted after consolidation.

Evaluate Operations Requirements

Staff and the processes they follow are always an important, and yet overlooked, part of an assessment process. Studies have shown 80 percent of data center and service outages are caused by people or process mishaps, and only 20 percent are caused by purely technical problems. It is essential to identify the processes and make suitable updates to make them pertinent to the post-consolidation environment. It is also important to update the staff skills regarding various data center activities (such as server, storage, and network management, security, disaster recovery, asset management, and hardware resource utilization management).

During one DC-consolidation project at a customer site, the client deployed a third-party application to facilitate server auditing and administration. The software reduced the time and effort required by SAs, and they were able to get by with fewer SAs.

Evaluate Risk

Consolidation, like any other IT project, has its fair share of risks. Like all risks, the main risk is looking the other way when it is staring you in the face. You do not have to like it, but you have to understand all the risks and implement ways to alleviate them.

The first step is to identify all the risks. The risk could be financial. It may be difficult to keep expenses within the allotted budget. The initial estimate of the project duration could be at risk if the parties you rely on do not respond or complete their parts in time. The level of expertise within your staff members could be inadequate for the work as it unfolds. Also, the scope of the project must be kept within the predetermined bounds. Your executive support might wither if the champion or main sponsor were to leave the corporation or division.

Once you have looked at the risks in the face, you must make a list of dos and don'ts to reduce their brunt. For example, to get a constant and high degree of executive support for your project, keep management posted on all happenings. Ensure that all parties are aware of deadlines, and give them as much notice as possible. One way to mitigate risk is to use common sense.

Continue to Phase 2: [Design and] Architect the target consolidated environment.

 

From Administering Data Centers Servers, Storage, and Voice Over IP by Kailash Jayaswal. Copyright 2006 John Wiley & Sons, Inc. All Rights Reserved. Used by arrangement with John Wiley & Sons, Inc.