🔊 Take a moment and listen

Neeta Verma
Senior Technical Director, National Informatics Centre

Penetration of Internet, use of Web technologies, globalisation of operations have moved most of the critical  IT Infrastructure of organisations  be it Intranet or Internet applications to the Data Centres, resulting in the need for much larger capacities in Data Centres. Requirements for reliability and availability of services has also risen as it now has  a direct impact on the functioning of the organisation. While on the other side, increased energy consumption, booming operations cost, increasing threats of natural disasters are making Data Centre (DC) an expensive asset of the organisation, not just the means to delivery of  services. Thus, there is a pressing demand on cost effective operations and management of DCs with high level of quality and reliability of services being delivered.

Setting up Data Centre

Setting up Data Centre needs a lot of understanding of current and future requirements, planning and provisioning for future technologies and is generally considered to be a capital intensive activity. However, what often one does not realise is that the Operational Cost of Data Centre over the years  also  becomes very high.  While a lot of emphasis is being laid on energy efficiency and green computing to reduce, rather optimise the over all power consumption in Data Centre. There are other aspects of Data Centre operations which are also of high significance. There is in fact a need for a  comprehensive view of Data Centre services and operations. Forrester in one of its recent report has highlighted some simple measures one can take to increase the Data Centre efficiency. (refer box)

Six Steps to Greater Data Centre Efficiency Rationalise your application portfolio Consolidate & eliminate under utilised servers Increase server Virtualisation ratio Enforce “virtual first” policy for new Apps Increase storage utilisation and reclaim storage capacity Optimise Data Centre temperature Forrester Research

Layered Infrastructure

Data Centre could be seen as a set of seven layers from basic site,  power and cooling infrastructure to core ICT infrastructure to Services layer. Different approaches are being taken to manage different layers. Basic Site with raised floor, control centre, fire protection, staging infrastructure to power and cooling infrastructure i.e., layer 1 to layer 2, comprise of Physical domain/Infrastructure. These are generally managed either through respective OEMs or through an outsourced agency. Large Data Centres also have their own civil and electrical engineers for overall control of the physical infrastructure in the Data Centre.  Lot of initiatives are being taken to optimise the cost and increase energy efficiency at this layer. These initiatives include proper airflow, hot/cold aisle containment, cable management, alternate cooling technologies, maintaining an optimal temperature in Data Centre etc.

ICT Infrastructure in Data Centre generally comprises of Layers 3, 4 and 5. Layer 3 and 4 are managed by IT Operations Team of the Data Centre. Layer 4 has compute infrastructure housed in Data Centre under different service models such as shared, dedicated, collocated and managed. Accordingly, different management approaches are adopted to manage compute infrastructure setup under different service models. Virtualisation, cloud computing, infrastructure as a service are latest technologies, being deployed in ICT infrastructure layer. Wide range of management softwares are available for this layer. Nowadays most of the hardware manufacturers provide browser enabled software, along with their equipments to facilitate convenient management of the equipment be it provisioning of resources, configuring security or monitoring the performance and trouble shooting.

Generally, large Data Centres are a complex environment with heterogeneous set of equipment from  number of different companies.

Product driven management software firstly becomes specific to only that component; secondly they also create silos of management windows and therefore, getting a comprehensive view of Data Centre becomes a challenge.  Customising proprietary software supplied with different products to get an integrated view is generally not feasible. A number of vendor neutral solutions are also available for ICT infrastructure. They are generally quite rich in terms of features such as topology creation, threshold management, proactive alerts, incident reporting, trend analysis, management reporting, dash boards etc.  These tools are however pretty complex and quite expensive too. Most of such tools are priced on license per node being monitored, whereby one is expected to procure a license for every node/ port being monitored and managed using such software. Alternatively, one can also consider using open source software, customised to their specific requirement. Data Centre managers are therefore, required to assess the criticality of applications and services being delivered and understand the service levels they are bound with to take an appropriate decision.

Layer 6 & 7 comprise of Service layers, dealing with delivery of services to organisation as well as their clients. This is the layer where service levels with respect to reliability, availability, performance and security are generally measured.  A wide range of softwares are also available in this domain to facilitate monitoring and meeting  the agreed service levels.

Integrated Infrastructure Management (IIM) shall enable the Data Centre Operations and Management team to get a comprehensive view of Data Centres across all its layers. It shall present the dependencies of resources in one layer over another layer. With the adoption of Integrated approach in management, Data Centre team shall be able to understand the correlation between adjacent layers.

Security Infrastructure cuts across all the layers, be it Physical Infrastructure, ICT Infrastructure or Service Layer and is a combination of hardware appliances, devices and software tools. Thus, each of the seven layers have their own set of management softwares, helping in  provisioning and monitoring of resources in respective layers, with respect to the parameters defined.  However, most of these softwares have been generally working independent of each other in past. We have seen some integration efforts in the recent past particularly in the ICT layer. They are also driven by the need for virtualisation, cloud computing, infrastructure as service offerings.  Initiatives are also being taken to provide integrated offerings spanning across major components of ICT infrastructure i.e., servers, storage and network primarily for better monitoring and efficient utilisation of resources.

Need for Integrated Infrastructure

Quality of services being delivered with respect to availability, reliability, performance and security, depends on the health of almost all the layers of Data Centre. Any weak link in the chain can have adverse impact on service quality. Thus, there is a need to get a holistic view of service delivery infrastructure deployed across all layers. Failure of one port on a SAN switch can bring the entire application down. Failure of one power distribution unit can create havocs.

Addition of one more application to data centre can impact the performance of existing applications. Need for redundancy configurations driven on service levels need to be not only provisioned but maintained in an ongoing manner, is pertinent. Data Centres are often challenged by varied requests of a wide range of complexities with a high order of urgency. Time to go live is often much less than need for proper planning, conducting an impact analysis etc. Keeping track of the service level requirements of different services, their impact on adjacent layers is quite a daunting task particularly in large data centres.

One also comes across situations where each layer of the data centre is individually performing very well against all defined parameters, still application is not able to meet desired service levels. Different layers being monitored and managed independently restricts correlation between the layers and troubleshooting such incidents is quite time intensive, needs engagement of highly experienced professionals and also impacts the reputation of the service under scanner.

IIM implementation shall also help in effective monitoring of Data Centre resources, application  and services. troubleshooting, fault determination involving a set of resources across layers, such as dip in performance of service with no apparent symptom, can be made far more effective with IIM implementation.

Lack of coordination between the physical infrastructure and ICT infrastructure, between ICT infrastructure and service layer is of great concern. Thus, there is a need to adopt integrated Infrastructure management  approach which cuts across all the layers of Data Centre.

Benefits of Integrated Infrastructure Management

Integrated Infrastructure Management (IIM) shall enable the Data Centre Operations and Management team to get a comprehensive view of Data Centres across all its layers. It shall present the dependencies of resources in one layer over another layer. With the adoption of Integrated approach in management, Data Centre team shall be able to understand the correlation between adjacent layers. It shall help in carrying out impact analysis of addition, removal, of any resource at any layer in data centre to the set of services being delivered from data centre. Upgrading of firmware, software versions, installation of patches can also cause unexpected disruption of services in Data Centre. Comprehensive view of Data Centre with cross dependencies of components shall help data centre team in avoiding such situations.

IIM approach shall also enable service administrator to view the dependencies of a service on various components of data centre across all its layers not only raising his confidence but shall also help in avoiding, deployment of unnecessary resources.

IIM implementation shall also help in effective monitoring of Data Centre resources, application  and services. troubleshooting, fault determination involving a set of resources across layers, such as dip in performance of service with no apparent symptom, can be made far more effective with IIM implementation. Lot of routine activities in terms of reporting, performance tracking, configuration management could be handled far more efficiently with much less human resource.

Integrated Infrastructure Management shall also help in carrying out a lot of proactive monitoring and avert possible failures. History data of performance, reliability and availability can be used in carrying out trend analysis leading to provision of right resources before the service actually has an adverse impact.  Comprehensive view of Data Centre resources shall also avoid situations of mismatch of resources leading to under utilisation of resources on one layer and overconsumption of resources on the other layer such as a large server with high power consumption running at very very low scale of utilisation; while on the other side there is application waiting for deployment in the want of power/ cooling resource. Worse could be a case, when an application is not able to deliver desired service level despite availability of resources on the other end.

Thus, adoption of Integrated Infrastructure Management approach in Data Centre shall help in achieving High order of efficiency in Data Centre operations, effective monitoring and optimal deployment of resources, leading to a significant reduction in Data Centre operations cost with much higher quality of service delivery.

Neeta Verma Neeta Verma is Senior Technical Director, National Informatics Centre, Ministry of Communications and IT, Government of India. She is looking after Data Centre and Web Services. She may be reached at neeta@nic.in

Be a part of Elets Collaborative Initiatives. Join Us for Upcoming Events and explore business opportunities. Like us on Facebook , connect with us on LinkedIn and follow us on Twitter, Instagram.