Last year, Analytics Week published an overview of the scope and growth of big data. The findings were remarkable:
- By 2020, for every person in the world, there will be 1.7 megabytes of new data created each second.
- The total amount of data in the world will reach 44 zettabytes by 2020, expanding from its current volume of 4.4 zettabytes.
- There are 1.2 trillion searches annually just on Google, averaging out to 40,000 searches per second — reflecting how users are generating much of this data.
The sheer amount of data, as indicated by this report, is becoming almost incomprehensible. Organizations are increasingly moving systems to cloud providers, propelling cloud hosting at a 23.31% compound annual growth rate (CAGR) through 2020, according to Gartner. Sometimes, though, the colocation of your own hardware makes more sense than moving everything over to the as-a-service model. Hence, the market for colocation worldwide is also skyrocketing — it’s projected to rise 15.4% from 2016 to 2020, according to BCC Research.
The overall trend is that firms are moving away from handling infrastructure themselves.
Regardless of exactly how these transitions occur, a data center migration is clarified by considering the standard steps to proceed. Below are simple steps for a successful data center migration, concluding with a close look at the migration of the data itself.
Determine What Elements Of Your System You Want To Migrate
You may want to buy new servers, only migrate a portion of your infrastructure or send all your hardware to the new facility. If so, this is an ideal time to get updated versions of machines and to end any current leases. Making replacements at this time simply limits your risk by minimizing the number of pieces you move. While the migration occurs, it can be a good idea to use infrastructure as a service (i.e., cloud hosting) or leased hardware (even if you are planning to phase them out) through the transition.
When you move to a different data center, you will have a chance to consider what parts of your infrastructure might be underutilized or otherwise inefficient. Take the time to consider refactoring or repurposing your hardware to better suit the new data center environment. Technologies like pods and aisle containment are good illustrations of this. You may also want to use the transitional opportunity to ratchet up your density. Take advantage of this move to improve your hardware setup and to increase reliability and efficiency.
Having determined the pieces of your back end that should be migrated, the next step is to consider whether you should do the move piecemeal or in unison. If you do the migration in phases, the advantage (other than lowering risk as you move) is that you can get pieces of the infrastructure into the new facility and start moving a few applications. Transitioning in segments like that could help you to prevent the need for cloud hosting or hardware leasing for a seamless transition that maintains your uptime.
Check Your Hardware And Space
Look over your inventory paperwork and your system logs prior to disconnecting cords and sending hardware for shipment. Verify that all your systems are present, making a note of any recently purchased hardware. Check the extent to which your servers are in use through analysis of current workloads, programs and scheduled backups.
It is critical that your services remain available during the move. Examples of key services are client-facing programs (e.g., SAP) and infrastructure software (e.g., Active Directory).
Disaster recovery as a service (DRaaS) or other third-party services should be contacted so that they can be attached to the post-migration facility. It is also possible that you will need a service license for a small period, possibly overlapping briefly with an identical service, as you transition.
Consider The Data Itself
Transitioning to a colocation scenario or from one provider to another is primarily about moving hardware. However, moving into a data center often means you have to migrate the data itself onto new equipment as well.
When you get ready to move data, you want to look carefully at both systems, the source and the target. The more thorough your understanding of the source and target, the better you will be able to efficiently migrate your data from point A to point B.
Think in terms of key risks. You do not want your data to become distorted or for unwanted duplication to occur when you move it; these issues often arise because the old and new environments are using different data formats and profiles.
The most critical elements of this process are the following:
- Extract: Get the data out of its current setting so that you can start to work on preparing it for migration.
- Transform: Change data formats as needed while checking that the metadata matches the applicable data fields.
- Clean: Remove any instances of duplication, test and take steps to resolve any issues with data corruption.
- Validate: Repeatedly test to verify that you are able to migrate and get the results that you should.
- Load: Perform migration, entering data into the new environment. Check again for any errors, and adjust as needed.
The Value Of Colocation
There are numerous reasons that companies are migrating their data centers. It allows them to treat their IT infrastructure as a monthly expense rather than an asset, freeing up cash for other priorities. It means they no longer need to worry about maintaining their own data centers. It decreases the emphasis on operating a facility and redirects the energies of the companies to their core businesses.
Migrating Data While Maintaining Privacy, Security and Compliance
As the data sets are becoming exponentially larger, so are the privacy, security and compliance requirements. Most importantly, verticals like health care and e-commerce are particularly sensitive to the added security. So whether your migration strategy includes a secure cloud or a combination of colocation and managed services, making sure regulatory compliance standards and requirements are met will go a long way in ensuring the integrity of the data.