Data center migration is never a happy IT decision for CIOs. But it can be executed in a hassle-free manner if planned meticulously By Mehak Chawla
While we sit back and enjoy our much awaited national holidays every year, there are some IT heads who sit with fingers crossed and await the successful migration of their data centers. Almost all CIOs have been there. At some point in time, the data center they have been relying on is sure to run out of something – space, power, tolerance. From there begins the daunting, and often terrifying, task of migrating a data center.
Migration of a data center and all its IT assets is a complex and large-scale change management effort, involving several stakeholders, and significant investment and risk. As per Gartner, the Indian IT infrastructure market is expected to reach US$ 3 billion by 2016. McKinsey has predicted a 32% CAGR growth for third-party outsourced data center market in India by 2017. Quite a bit of this growth in the business of independent data centers (IDCs) is expected to come from migrations, either from captives or from other service providers.
The reasons for rising data center migrations differ from cost efficiency to regulatory pressures. Though data center migrations have become much simpler than they were, they are not without significant complexities.
As 
Sridhar Pinnapureddy, Founder & CEO, CtrlS Datacenters, observes, “Migrating a data center is much easier today than it was few years ago, but it is still not a decision to be taken lightly or without careful consideration. Over the years, we have learnt many lessons in this area, and this has translated into better processes and a more structured approach to data center relocation.”
However, what a migration certainly does involve is meticulous planning, extensive testing and a strategy involving all partners- right from the networking managers, to the service provider and the end -user. Express Computer looks at factors encouraging the migration drive, and the challenges that CIOs are most likely to face when dealing with a data center migration.
Key drivers
According to Krishna Shankaran, Delivery Manager, Infrastructure Management Services, Infosys, the primary reasons that compel enterprises to undertake data center migration are business drivers, technology drivers and regulatory drivers. Business drivers include things like mergers and acquisitions, or the possibility of lowering TCO by shifting to a new facility. There are quite a few data centers in India that are no longer able to keep up with the power, cooling and efficiency demands of today’s high performance, high-density equipment. Lack of upgraded data center facilities, resulting in business disruptions, is making a strong case for data center migrations.
When it comes to technology drivers, the technologies that are compelling CIOs to evaluate migration are virtualization and cloud. “With the advent of cloud computing and a growth in consolidation requirements, there is a growing trend amongst today’s businesses, with upwards of 53% of companies, planning either data center re-locations or consolidation of some kind within the next two years. Two disruptive technologies impacting the future of data centers are virtualization and cloud computing,” elaborates Deepinder Bedi, Executive Director, Tulip Telecom.
“Moving dedicated applications to cloud or even modernizing applications may drive organizations to go for data center migration,” says Shankaran. Huge underutilized capacity of both servers and storage is also assisting migration decisions. According to Sunanda Das, Managing Director, Pacnet India, “Enterprises nowadays are increasingly realizing the benefits of IT consolidation and virtualization from a cost and management standpoint, prompting them to consider migrating to modern, cloud-enabled data centers. In India, migration to a hosted data center is a growing trend, with larger enterprises demanding dedicated co-location services.”
Mandar Kulkarni, VP, Solutions Engineering and Private Cloud Practice, Netmagic corroborates by saying that currently, “Most of the business that is coming to us is either from greenfield projects, or from migrations.”
Apart from these, regulations, especially in economies that have stringent environmental guidelines, are also triggering migrations from inefficient captives to more efficient, hosted data centers. A few companies/countries’ regulations may demand a disaster recovery strategy, as a result of which organizations may need to go for migrations.
These drivers have also sparked another trend in the data center space. They have compelled many enterprises, that typically prefer to own a captive data center in the interest of security, to seriously consider the hosting option. Also, since third party data center providers are able to bring significant value to clients with varying compute workloads, the proposition of migration has become more attractive than ever before. This level of flexibility can be a major benefit for smaller organizations. Some of the smaller, internet-centric organizations are actively exercising this option. As Kulkarni mentions, “When a captive data center runs out of capacity and becomes inefficient from a cost perspective, enterprises often evaluate migrating to a third party data center.”
Migration options
Data migration, in today’s virtual world, can be of two types – physically moving hardware to a new location or moving data to a different data center. Either ways, it is critically important to ensure aspects like security, business continuity and minimum downtime. So, what are the options available to an IT head, once he has decided to undertake data center migration?
According to Nilesh Goradia, Head, Pre-Sales, India Subcontinent, Citrix, “Most of the migrations are happening from captives to IDCs or to a hybrid model using bridging technologies. In such a scenario, application workload can pan across both data centers.” He goes on to add that a migration scenario might be a good time to upgrade technology as well as evaluate what cloud model is best suited for your organization. “Efficiency and manageability that comes with the cloud infrastructure in terms of workload provisioning on demand, without the need of buying too much new hardware is driving migrations.” However, an enterprise must engage with a strategic partner in a migration scenario, because a data center migration requires meticulous planning. A planning cycle can be as long as 3 months, Goradia cautions.
Kulkarni of Netmagic elucidates the three key approaches to data center migrations. The first is the big bang approach. Here, the enterprise identifies a lean period (like a national holiday) and migrates the entire data center in that limited window. It works for small to medium data centers with 20-30 racks. “It’s a quick method but requires extensive planning”, advices Kulkarni. The second approach is big bang combined with disaster recovery. Here, the enterprise brings their disaster recovery (DR) site in play during the migration to avoid outage and still serve customers while the data center is in transit. Large organizations that cannot sustain downtime and have a functional DR, adopt this method. The third approach is the bubble set up. This involves creating a small network bubble and moving apps in a phased manner. Though bubble set-up is the costliest option out of these three, Kulkarni reveals that enterprises are using it for their mission critical apps. “Sometimes, customers end up using a blended approach . Bubble set up is the costlier option but it is available on a pay per use model. Big bang is the most widely used approach. BFSI is the sector that is doing big bang with DR,” he explains.
Another trend that is catching up, and coming in handy for mid-sized organizations, is data center migration-as-a-service. Since there are quite some complexities involved in moving a data center — labeling, legacy apps, Unix machines that haven’t been restarted in years and security in transit, organizations are now opting for migration services that most of the IDCs provide. Cost of these services is usually on per rack+manageability basis.
Orchestrating a migration strategy
If there is a unanimous view on data center migration that is shared by the vendors and IT heads alike, it is the emphasis on planning, rather than actual execution. The one thing that IT heads must be prepared for is numerous hidden complexities. They are sure to run into legacy apps, a broken hardwired IP address, Unix machines, archaic mainframes and countless networking issues. Data center migration, say experts, is all about delving into detail.
Das of Pacnet illustrates some unforeseen complications that might arise during data center transition. “Migration often also requires the use of cross-platform technologies, some of which are untested and can potentially affect the speed and efficiency of migration. Factors like this should be accounted for in the planning stages itself.”
Arun Gupta, CIO, Cipla, shares his migration experience and the planning it entailed. “We consolidated from five enterprise owned data centers to a co-location, hosted data center with a service provider. Planning meticulously to cover each and every step and its impact, should any step not work, was critical towards successful transition. We went over the plan many times, each time refining it before execution. A team of 20 people worked for over 48 hours to achieve the migration successfully.”
Any data center migration strategy should work in three parts, believes Ashok Shenoy, Regional Manager, Cloud and Data Center Sales, Cisco India & SAARC. The first part is orchestrating the complex move of mission-critical data center assets from one site to another. The second step involves ensuring business continuity, maintaining application performance and availability, and delivering operational service levels throughout the physical move, with little or no of error for costly, unplanned service outages. The final part deals with growing the investment value of the data center by evolving data center strategy to meet business objectives.
Other aspects that need to figure in a migration strategy are availability of expertise, appropriate and timely communication with the end-user, application delivery optimization, and proper testing for things like load balancers and optimizers to see if they are compatible with the new hardware models. Shankaran of Infosys points out another challenge that CIOs might face. “Managing application to application or application to infrastructure inter-dependencies is a key challenge during data center migration. If this is not done properly, there is always the risk of crippling outages. Lead time required to procure new equipment and getting support on older/legacy equipment is another key challenge.”
Security is definitely one of the key concerns that makes enterprises hesitant about migrating from traditional data centers to virtual data centers and cloud. Many a time, organizations are not fully aware of all the applications currently in use. If critical applications are not properly identified and assessed prior to the migration, there is a serious risk of data loss, leading to downtime. It is necessary to ensure proper backups are in place before migrating the data.
SriRama Evani, Principal Director & Head , IT Services, DST India, runs us through his experience of migrating a data center as a result of a shift in office location. “For us, it wasn’t only movement, we also decided to optimize our data center during this migration process. We wanted to do virtualization and bring in cost efficiencies in our captive data center, while also expanding from a scalability perspective. The key to data center migration is planning and designing. We took almost 2 months to plan our pre-migration activities. The migration was done by our own IT team, since we have senior people who have handled data center migrations before.”
According to Evani, factors to be considered in a migration scenario include data security, business continuity and disaster recovery. “Also, retrieving data from backups becomes crucial. Another typical challenge is combining virtualization with migration. A lot of licensing issues come up. We migrated about 20 racks in a phased manner.”
Major challenges from a service provider perspective are to convince customers, since migration is never a happy IT decision. Also, the window is often too small. Hardware failures are another problem, and the service provider has to ensure that all machines are rebooted well in advance.
Post migration blues
So you have just completed your data center migration. But can you claim it to be successful yet? Well, a migration is almost identical to setting up a data center from scratch. As a result, there is much post-migration work to be done. IT heads not only need to justify the migration investment, but also grow it to improve TCO and RoI of the new data center.
Upon completion of the data center migration, it is imperative to take one additional step: the incorporation of knowledge, updated processes, procedures and documentation into the normal support structure of the IT infrastructure. Post the migration, it is also important to ensure that no traces of data are left behind in the old servers. The three most effective ways to do this, according to Pinnapureddy are: physical destruction of the hard disks on which the data was stored, degaussing, a technique where a strong set of magnetic fields are passed over the hard disks to ensure that the old data can never be recovered, and finally sanitizing, which involves overwriting the disks completely with junk data to ensure that none of the old data can be recovered from it. “There are certified tools available in the market that enable complete sanitization of disks,” he explains.
Apart from ensuring data security, there are several tests to be run, applications to be optimized, network loops that need to be corrected and cloud strategies and protocols to be implemented. So the CIO’s work doesn’t really end with infrastructure migration. It actually begins from there.
