Datacenters At Sea
Making IT infrastructure work in one of the world's most challenging environments.
- By Dan Kusnetzky
Recently I was introduced to a rather interesting use case for DataCore's software-defined storage (SDS), deploying a highly available, virtual environment at sea. While at first, a shipboard network that supports daily operations of both the ship's travel management, supply and other internal systems as well as connections to the company's worldwide network appears fairly simple, deeper consideration of what this would take highlights many layers of complexity.
The fleet includes six ships. The actual datacenter configurations are different from ship to ship, depending on when they were commissioned and the best technology available at that time. Systems from HP and Dell have been utilized. They're set up as four VMware ESXi hosts and 50 virtual machines. Veeam provides VM backup, while DataCore SANSymphony 10 provides storage replication and optimization. Dual 10Gb Fibre channel connections were deployed to provide highly available and reliable communications between the onboard datacenters.
Dan's Take: How Ships Are Like Branch Offices
Having developed configurations for mobile applications long ago while I was part of the network services group at Digital Equipment Corporation (now part of HPE,) I understand the complexity of this environment. Building a reliable, compact, agile and flexible shipboard IT infrastructure would have multiple, complex challenges, including:
- Ships have very little space available for equipment, so the systems must be small. Rack-mounted hyperconverged systems would be good candidates for this implementation.
- Every function should be implemented in at least two places that are on separate power supplies, on separate decks and in separate fire zones. Multiple network links should also be supported, using whatever LAN technology is required and allow updates to newer technology when necessary. WAN links using satellite connections at sea and other technology should also be supported while the ships are in port.
- Systems, storage and networking equipment are evolving rapidly. So the architecture of the onboard configurations should allow for replacement of systems, storage, networking and power equipment as technology changes. Furthermore, products from many different suppliers should be supported, allowing the company to select the best equipment available at the time and update as conditions change.
- Applications and application functions must be made highly available. This implies that they should be encapsulated using some form of virtual processing software, such as virtual machine (VM) or container technology, and that VMs or containers must be monitored and be able to be migrated to another system should their original host show signs of failure or performance problems. Storage systems must support replication from one datacenter to another onboard the ship, or from ship to corporate headquarters as necessary.
- Applications and data must be made secure and the environment must resist network attacks.
- The onboard configuration should provide an environment hospitable to Windows, Linux and any other industry-standard operating environment.
- The entire onboard configuration should be self-managing as much as possible and allow shipboard and remote IT administrators to monitor and manage operations.
I could go on and on, but it's easy to see that this type of complex design will have many moving parts. It's probable that many different suppliers will provide equipment and provisions must be made in the architecture to support them.
DataCore's virtual storage environment and Veeam's VM backup data availability tools, I'm sure, were a big help to the designers because of their ability to work with systems, storage and networking equipment from many different suppliers.
What struck me as I was reviewing this shipboard configuration was that similar thinking is required when developing remote office/branch office configurations.
If your company has to support widely distributed offices and make sure that the IT infrastructure is cost-effective and reliable, learning more about this use case would be valuable.
About the Author
Daniel Kusnetzky, a reformed software engineer and product manager, founded Kusnetzky Group LLC in 2006. He's literally written the book on virtualization and often comments on cloud computing, mobility and systems software. He has been a business unit manager at a hardware company and head of corporate marketing and strategy at a software company.