Prashant Baweja is working as a senior associate consultant with Infosys Ltd. in infrastructure domain and has a total experience of about six years.
For any business, delivering services (at right time, right place and at right cost) to its end customers is one of the most important objectives. This will drive a company’s growth and profitability. Information Technology plays a major role in achieving the same and IT infrastructure (or more generally, the data center) is a critical asset within it.
A large number of external and internal factors force a company to look at its IT capacity and data centers to be revamped or build afresh. Some of the factors like growing businesses want data center capacity to be expanded to handle more data and transactions, New or re-vamped backup or DR/BCP strategy, modernization of legacy hardware, increasing data center utilization and performance by consolidation/virtualization/transformation, etc.
Data center design focuses on many aspects like Site selection, facility design, IT hardware, cooling, network, power distribution, etc. For a successful data center built project, few parameters needs to be kept at high priority and planned for effectively. This includes characteristics like reliability, availability, serviceability, capacity, scalability/growth plans, manageability and criticality. These factors will help design team and management to design a data center (facility and IT) effectively with required business needs.
Below I outline the parameters and will help in data center planning phase:
Reliability means ability of the system to perform its task and function under stated conditions for a specified time. It is calculated by measuring Mean Time Between Failures (MTBF) and Mean Time To Repair (MTTR).
While designing a data center, reliable components should be used like power supplies, cooling units, servers, storage units, network, etc.
Availability refers to the time or proportion of time for which a system is available and operable. It is represented by ratio of system uptime to total time:
Availability = System Uptime / Total Time (Uptime + Downtime)
Data center tiers and availability is interlinked and are as follows:
- Tier 1: Guaranteeing 99.671% availability.
- Tier 2: Guaranteeing 99.741% availability.
- Tier 3: Guaranteeing 99.982% availability.
- Tier 4: Guaranteeing 99.995% availability.
While designing a data center, high availability of servers, storage, network, application, power and cooling units, etc. should be considered. Based on requirements, clustering, load balancing, redundant systems needs to be designed.
Another parameter which is of importance while designing a data center is criticality of application/service provided to end customers. If service rendered is highly critical and its downtime is costly for the company, data center and IT infrastructure has to be designed accordingly. In contrast, if application is less critical and its downtime is acceptable to certain extent, data center has to be designed accordingly.
Criticality can be represented on a scale of 1-5 with 5 being most critical and 1 being least. For e.g. Banking and financial applications will be categorized as scale 5. While designing data centers, this factor needs to be considered and redundancy of power, network, etc. should be considered.
Capacity parameter is crucial to know the size of data center and other facilities to be designed. It represents maximum load a data center can handle. Goal of capacity parameter is to make sure that enough power and cooling is available at data center.
Growth plan is another major factor which will provide us information on what can be expected over a period of time in terms of load and requirement. data center should be designed to handle the load as per the growth plans over the years.
It is important to factor-in growth plan as it will help to add hardware and facilities at required timelines, thus saving money by not keeping them idle.
Scalability is closely related to capacity and growth plans. Based on these two parameters, infrastructure and facility is designed which is scalable over years and can be adjusted smoothly. Servers, Storage, Network, power, cooling etc. should have provision for scalability. Scalability can be Horizontal or Vertical Scalability. Horizontal Scalability implies adding more and more separate independent systems to service the increased load. Vertical Scalability is adding resources within the box to cater to increased workload.
Due to more and more development in green technology, companies are trying to achieve a target for greening and efficiency. For a better and robust design, it is recommended to freeze efficiency targets during design phase itself.
It refers to information on floor space available and number of racks in the same. In addition to that average and peak power consumption of each rack will be useful. This parameter will be tightly coupled with availability, efficiency and scalability factor.
Serviceability refers to how easy it is to find a root cause for an incident or problem and solve the same. Ideal situation will be when issues will be tracked automatically and infrastructure in itself solves the issues on the go. During data center design, facilities, hardware and architecture should be finalized keeping serviceability characteristic in mind so that failed components can be replaced without bringing down the entire system.
It refers to how easy, efficient, less time consuming, detailed and effective ways are present to manage facilities and IT infrastructure. This factor will help to make data center operations, management and monitoring easier and secure.
With above characteristics, data center design will be stable, secure and secure enough to provide services as required. Each data center setup/transformation project is unique with different set of expectations from stakeholders, but if above mentioned is taken into consideration during design phase, chances of project success will be increased manifold.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission processfor information on participating. View previously published Industry Perspectives in ourKnowledge Library.