As the demands placed on data centres intensify, more quality cloud-based solutions emerge. This means Data Centre Migration to the cloud, is becoming an increasingly unavoidable consideration for IT managers. Although a complex undertaking, the efficiency benefits further down the line can often justify the time, effort and capital required.
[easy-tweet tweet=”Migration to the #cloud can take several forms: #IaaS, #SaaS and #PaaS” user=”comparethecloud”]
Migration to the cloud can take several forms: Infrastructure as a Service (IaaS), Software as a Service (SaaS) and Platform as a Service (PaaS). IaaS involves substituting servers in your data centre for servers in the Cloud. The leading providers of these types of services are Amazon Web Services and Microsoft Azure, though a plethora of local data centres provide similar, albeit somewhat less scalable services. SaaS, of which Salesforce is probably the most well-known, is at the other end of the scale to IaaS, whereas PaaS, such as Amazon Elastic Cloud, Force.com (by Salesforce) and Azure web platform, sits in-between these two approaches.
Managing your own data centres allows you to remain in complete control
Managing your own data centres allows you to remain in complete control, but can be very costly. For this reason, many have decided to look at migrating some or all of the hosting from their own data centres to the cloud. However, this does mean entrusting your data and some or all of your operations to a third party. What happens if there is a security breach, data loss or an outage? These issues boil down to two questions; do you trust the third party, and if you do, will your regulatory environment allow you to entrust your data and operations to them?
Once a decision has been taken to migrate to the Cloud, it is important that you plan thoroughly and stick to a clear and thought-out structure. Here are some tips to help ensure your Data Centre Migration to the Cloud flows well:
1) Consider the risks and how you are going to manage them
Generally, the main risks in a Data Centre Migration are loss of data and unplanned periods of downtime. Risks can generally be mitigated by performing trial migrations and testing the relevant applications after the trial migration. More testing incurs greater costs, however, and a balance may have to be struck between the risks avoided by thorough testing and the cost of the testing. Classifying applications by business criticality and the technical difficulty of migrating them can save time and ensure consistency in this decision making process.
[easy-tweet tweet=”A balance may have to be struck between the risks avoided by thorough testing and the cost of the testing”]
2) Identify potential incompatibility issues
Legacy apps often do not translate into new environments due to not having the right operating systems and hardware – IaaS does not normally include the older operating systems which have legacy hardware. There are two ways of dealing with this issue: either finding a way to run the legacy application on modern hardware, which either requires emulating legacy hardware, or replacing the legacy application with a newer one with similar functionality. The new application, may even be SaaS, which would eliminate infrastructure concerns after migration. Replacing the application is a project in its own right and may come with extensive licensing, integration and training costs, but such an exercise may be necessary anyway, and doing it as part of the migration can be cheaper than performing the migration and then replacing the application a short time later.
3) Check the network you are migrating to can be configured to run your apps
During a Data Centre Migration you will be migrating multiple applications which currently exist within a network configured to a certain spec. For IaaS migrations, if the new network does not work in a similar way to the current one, the migration will have to take this into account or the applications will not function correctly after migration. To help avoid issues like this, make sure that you fully understand your current data centre network infrastructure, considering elements such as firewalls, domains and trusts and ensure that you are given a good understanding of the one you are moving to. If the IaaS does not permit the desired settings, adjustments must be made to the applications to allow them to work within the constraints of the IaaS.
make sure that you fully understand your current data centre network infrastructure
4) Consider network latency
It is likely that Cloud based infrastructure will be physically further away from your users than your data centre was. Very often, a traditional data centre is on the same site as the users of the applications it hosts. Cloud infrastructure may be thousands of miles away. For many applications, especially those with a browser based architecture, this will not be a problem. For some applications, especially older ones, an increase in network latency can cause a major drop in performance. To counter this, the application can either be replaced by a more modern one, be hosted on a thin client system such as Citrix XenApp, or be redesigned to work better with higher latency.
5) Talk about discovery
Depending on how long the original Data Centre has been up and running, there may be existing applications which have outlived the presence of the employees that set them up in the first place. It is very feasible that the knowledge of how exactly these applications work and even the very fact that they exist left your company with those staff members.
[easy-tweet tweet=”The main risks during #datacentre migration remain unplanned downtime and loss of #data”]
If you are considering the migration of an entire data centre, it is crucial to know in full what is hosted in the data centre and how each application interacts with others. Lack of awareness about this can result in lost functionality and potential loss of business-critical data. Network tracing and discovery tools can be used to analyse systems over time and re-discover any forgotten components and relationships.
The main risks during migration remain unplanned downtime and loss of data. Those technical risks are complemented by the risks of unforeseen delay and cost, caused by such factors as migrated applications failing tests before going live. To mitigate these risks, a strictly structured and planned approach is essential.
Francis Miers, Director, Automation Consultants
Francis Miers joined Automation Consultants in 2002 as a director after establishing a career in IT and the City, starting out in a French software and services company and has since been engaged in corporate finance in the technology sector. He has a BSc in physics from Imperial College London and an MBC from Erasmus Business School in Rotterdam.