While many companies sell cloud services to track and manage customers, Aquent's staffing business is based on providing better services, so the company's software development team decided to port its existing system and looked to host it on an infrastructure-as-a-service cloud.
2. Copy and paste is not so simple
To reduce latency and ensure maximum availability, Aquent maintained databases in colocation facilities in Australia, the United Kingdom and the United States. Yet, the company feared that data maintained in those systems would quickly become fragmented, says Bolick.
"If you were in your office in Hong Kong, and you needed to find a resource for a specialty in Hong Kong, you would have to search the database in Sydney, but also the databases in New York and London," says Bolick.
The company hired technology-services firm Distributed Logic of Woburn, Mass. to create a solution, a multi-master database replication system that could synchronize data between the three operational databases.
3. Latency, consistency matters
With a global operation, latency in the cloud starts to matter tremendously.
"It becomes very difficult or nearly impossible to use a real-time application a continent away," says Mark Parsons, a lead developer with Distributed Logic, who worked on the Aquent project.
While the company has a standard three instances of its ERP system in the Americas, European and Asian markets, the flexibility of the cloud allows the company to add new instances as needed. It will likely add a fourth instance on the U.S. West Coast and may expand with a Japanese instance, says Aquent's Hunter.
Selecting Amazon as its cloud infrastructure provider allowed the company to quickly spin up new instances even in countries where the communications infrastructure and language are different.
4. Be prepared for outages
Moving to the cloud is not an automatic disaster recovery plan. Planning is required. While cloud services can improve availability, they can also result in outages that affect a large swath of businesses. The recent Amazon EC2 outage in North America is proof of the dangers.
In Aquent's case, its technology provider Distributed Logic created a multi-master replication system, so that Aquent's cloud infrastructure could be continuously updated amongst all three of its current zones in Australia, the United Kingdom and the United States.
The benefit of this infrastructure design became evidence during a recent outage. While Aquent was not affected by the North American Amazon outage, its infrastructure allowed it to avoid a 12-hour Amazon outage in Ireland that took the company's application server down in March, 2011. The company was able to point its European workers to the North American server while it quickly rebuilt the European server.
Sign up for CIO Asia eNewsletters.