Conquering the Challenges of Managing a Data Center

Over the last few years, we have seen the number of data centers growing at an exceptional rate. The term data center brings to mind a picture of a highly secure room spread over acres of land, with organized cabling infrastructure, extreme cooling and dedicated power house. Well, in reality things are slightly different. A data center, in simple terms is nothing but a place that holds your data, IT infrastructure and applications. In the early days, there was no term called 'data center', as such. There were server rooms where all the servers were kept and managed by an expert IT team. With the 'Dot com' boom, emphasis on datacenter has risen at a phenomenal rate. Initially, these data center facilities were constructed by ISPs for hosting applications, servers etc, for their clients. These days, each and every enterprise has its own data center in place. But on the other hand the complexity of managing these data centers efficiently has also become a challenge for the CIO. Therefore, we decided to find out the key data center management issues faced by CIOs and try to find answers for them. For this, we interacted with 28 key CIOS from across the country.

Understanding the key challenges

52 % of the respondents said that ensuring high availability was the most challenging task for them. Around 24 % respondents said capacity planning was their key concern. While 10 % of the respondents primarily faced issues like keeping costs under control, the remaining 14 % said ensuring optimal utilization of resources was their prime challenge. To ensure high availability, you need to have redundant power backup. Secondly, data center should adopt network load balancing along with DR so that stress on the data center can be minimized. For the critical applications running in your data center, also you should have an automatic fail-over setup. Build redundancy into all the possible elements that can affect high availability, for example, switches and routers. If you think, you don't have enough trained staff to provide high availability then better outsource the management of your critical apps. To combat the issue of capacity planning, one of the options suggested by some of our respondents was server consolidation. One of the requirements for doing a successful server consolidation is monitoring your IT resources and then formulating the strategy. Broadly speaking, server consolidation translates into IT resource management. If you think that your current data canter can't take load of your upcoming projects then only you should revamp your data center (DC) or if you don't have enough time and budget then outsourced DC would be a better option. Virtualization is another solution for capacity planning. It taps the unused processing power of the servers in your data center. Moreover, with virtualization, you can add more apps in the same environment in order to utilize unused server power for efficient resource management. This will also help other concerns like ensuring optimal utilization of resources and keeping costs under control.

Key management concerns

Power concerns top the list, followed by crash and recovery. There are also connectivity related, cooling and data backup issues. Let's take these issues one by one. Yes, power is the basic need for a huge data center, moreover as your data center grows you would require more electricity in order to power your infrastructure. Here also, capacity planning plays a major role. One has to evaluate the present and future power requirements for a data center. Then deploy a power conditioning system for your DC, which includes UPS and indigenous power generation unit. These days, many organizations have their own power generation units for powering their data center grid. Next you may face crash recovery issues. For instance, if any of your mission critical applications fail due to hardware failure then what would be your recovery strategy to bring back the application with minimum down time. In that case, keep your hardware and spares ready in stock, so that you can just replace the hardware and host the application on a new piece of hardware. Connectivity issues are another common area of concern that CIOs face, while managing their data centers. In fact, one interesting aspect that came up from our survey was availability of network equipment. What if one switch fails somewhere in your large data center? How quickly would you be abe to find and rectify it before something disasterous happens? For this, you need real time monitoring of the networking equipment and failover support for the most critical ones.

Data centers have a lot of servers and other equipment that generate huge amount of heat as well. As the temperature rises, it adversely affects the performance of the data center, plus chances of wear and tear of equipment also increase. Therefore, cooling plays a very important part in your DC. Before building a DC, you need to analyze your cooling requirements and design your DC accordingly. For your existing data center, you should put in temperature monitoring and control equipment. One of the respondents said that for additional cooling on demand, you can also deploy emergency chillers.

Future readiness

How scalable is your data center? Has it already reached its limits? Or will it last for a few years? Or it's got sufficient capacity so you don't have to worry about scalability issues for a long time? 41% of the respondents to our survey said that their datacenters would last them for a few years. However, there were 27 % respondents who said that their data centers were only somewhat scalable, i.e. only minor additions to the infrastructure were possible. Another 17% of respondents were worried, saying their data centers were not scalable and so they would have to setup a new one in the near future. Only 10 % of respondents actually said that their data centers were highly scalable and would last for a long time. Now the question is what are the things you should do to make your DC future ready? First and very important, while building a new DC, you need to evaluate your future requirements, which include new technologies and scalability options. If you can't afford costly technology today, make your DC scalable enough to accommodate new technologies, in future. So, that whenever your budget permits, you can add new technologies without any hitch. Some of the new technologies for data centers include virtualization, blade servers, SAN, DR/BCP, server consolidation, biometric security, 10 Gigabit Ethernet backbone and KVM over IP for remote management. According to our survey, 51 % of the respondents have already done server consolid ation and 41 % will be doing it in near future. Likewise, another 51 % have already deployed SANs and another 24 % respondents would be doing so in two to three years. Physical data center security is also a key concern, and 34 % of the respondents had already deployed biometric security solutions, while another 24 % were planning to do so in the near future. While analyzing the survey data, we found a surprising fact in the area of DR and BCP. Only 34 % of respondents have got DR and BCP deployed at their place, whereas 41 % respondents plan to implement it very soon. This is quite surprising given the criticality of having DR in a data center. Only 37% had deployed KVM over IP for remote management, and another 27% were planning to deploy it in the near future. Connectivity is another area that needs to be looked upon, while building a data center. Does your data center require 10 G? As it's still a new technology, only 27% of the respondents had already deployed it, and another 34 % were planning to do so it in near future.

Monitoring and management

Monitoring and management is a key element that is essential for any data center. This helps CIOs evaluate the health of a DC on a real time basis. So what kind of a monitoring setup should you have in place for your data center? According to our survey, this was evenly split between either having a 24x7 monitoring setup manned by in-house staff, or a completely outsourced management model. Very few respondents said that they didn't have a dedicated monitoring setup. So, one thing's pretty clear from this that you need to have a 24x7 monitoring in place for your data center, whether in house or outsourced.

About 44 % of the respondents had their own in-house staff handling the 24x7 monitoring, while another 41 % had outsourced it. Now the key question here is which should you opt for? In house or outsourced? Of those who had outsourced felt that getting the right staff to manage a huge setup is extremely difficult. Even if you manage to get the right people, retaining them can be a challenge. Outsourcing can have its own set of factors that need attention as well. For instance, when outsourcing the monitoring and management, how trustworthy is the agency you're outsourcing to? What kind of SLAs can be accomodated to ensure that it's absolutely safe to outsource needs to be looked into. If you still prefer to have inhouse staff for doing so, then constant motivation is required to retain the staff. They need to see a clear cut growth plan for them.

How do you handle issues like peak utilization of your data center? It's obviously not going to reach peak utilization all the time. So you need to identify the exact period when this happens. For instance, many of the respondents said that month ends were when their data centers reached peak utilization. Once you know this, you need to ensure that sufficient resources are available to handle the additional load. One way of ensuring this according to some of responses is to do proactive monitoring and terminating the lower priority, resource intensive processes and diverting the resources to applications with high priorities.

Inhouse or Outsourced?

By and large, 51 % of our respondents were running their own data centers. Only 20 % had completely outsourced it, while another 27% were using a mix of both. That raises one more interesting question. How many service providers should you outsource your data center services to? And for how long? Interestingly, a majority of the respondnets were not outsourcing the services to a third party. Only 28% had outsourced it to one service provider, and a very small fraction had gone beyond outsourcing to more than one service provider. These statistics bring out many questions to answer, when should one go for outsourcing? Is it safe? And, what kind of SLA you should agree to?

We suggest you to keep your needs in mind and check three major things before deciding outsourced DC. One of course is budget. Do you have sufficient funds to host another new service in your data center or would it be more economical to just outsource it? Second is whether your data center will be able to handle additional workload? If it's not scalable and you're not able to move to a new data center in the near future, then it's better to outsource. Third is availability of manpower. How difficult would it be to get new skilled manpower for any additional service you're going to deploy? If it's going to be difficult, then outsourcing could be a consideration. Some of the responses we got suggested that mission critical applications such as mail and transaction processing should be hosted inhouse, whereas applications like help desk system, hosting of services, server management, DB management and application management can be outsourced.

Another point to note is for how long do you want to outsource? Should it be a bulk deal lasting many years or is it better to go for annual deals? While long term deals might be more cost effective, it bounds you with the service provider for a long time. So you may want to weigh the pros and cons of the duration for which you should outsource.