Meanwhile, Behind the Server Scenes…

With the launch of our latest publication a couple weeks ago, we’ve now released a total of eight projects to the scholarly community, ranging in topic from Black history to archaeology to Middle East studies to ecology. And as different as the subjects each project covers are the technologies on which they are built. That’s why, with each new publication we release comes the need to look again at how we’re serving our content. This post is a return to some of the more technical topics we’ve delved into along our now six-year digital publishing journey and looks at one of the very first technical considerations we had to undertake as a publisher, and thus host and steward, of web-based digital scholarly monographs.

By the time our second publication went into production in 2017, we realized we needed the responsibility of hosting to reside with the Press. By hosting projects ourselves, we were able to establish a set of standard recommendations that would keep our various projects centrally maintainable. And directly managing the server and all the publication content would also allow us the access we needed to be able to make production edits as well as export components for archiving and preservation. We could also provide access to partners along the way who have been willing to work with us over the years on testing an array of approaches to archiving, including emulation and virtualization, none of which we could do if the content wasn’t firmly in our hands.

When we decided back in 2017, for these reasons and more, to do the hosting ourselves, we went with Reclaim. The then small team, led by Tim Owens, Jim Groom, and Lauren Hanks, was extremely approachable and willing to work with us on our often highly experimental use cases. They offered levels of support, tailored to an academic organization, that far surpassed what we could expect from larger corporate hosting providers. Lauren even attended our 2018 preservation summit to hear what exactly it was that SUP was doing and how we were anticipating and attempting to intercept the problem of longevity in web-based digital scholarly content. She sat with us and members of the preservation world, including experts in web archiving, emulation, the Stanford Digital Repository, and our own contracted authors to brainstorm how our worlds could merge for the sake of an enduring scholarly record. Over the following months and years, they graciously continued conversations with us and with each other. Tim Owens patiently and sympathetically heard my monthly-ish requests for support of Docker technology, which in turn I was hearing from potential authors who wanted to deliver projects to us in containers ready-made to load onto a server.

It’s this last part that has inspired the return to this train of thought. Since we’ve been hosting with Reclaim, there have been some exciting developments that we’ve been able to try out and ultimately now adopt and convert to. A couple years ago, one of our projects was in danger of early deprecation due to a limitation on our shared server. A highly complex Ruby-on-Rails-based publication would no longer run in the environment shared by all of our other publications whose dependencies were far lighter than that one that was taking up far more resources. But while Reclaim was sunsetting support for the infrastructure we needed for that one project on the shared server, they were also introducing their new Cloud service, which could, as it turns out, host this one finnicky project within its own container environment for a fraction of the cost. Not only has this new system rescued one publication from an early demise, it also offers a much greater landscape for future projects that use technologies we’ve had to remediate in the past. With the potential now to support Docker containers for the transfer and possible deployment of publications, we open the door to wider possibilities in platform choice for authors. This could in turn mean shorter development times as projects don’t need to be migrated to simpler formats, and easier conversion to archive formats such as emulation, which could in theory now be served right from the same Reclaim Cloud server.

All of this is a bit of background to say I’m currently taking a flex course offered by Reclaim Hosting on how to set up, run, and maintain Docker in Reclaim Cloud. The month-long course led by Taylor Jadin, one of Reclaim’s growing team, familiarizes attendees with the tools and processes involved in using and supporting Docker in the specific environment SUP is using, and I’m looking forward to being able to bring this knowledge to our authors who have been waiting for the opportunity to take advantage of the portability offered by containers, and often Docker in particular. It makes transfer of content much cleaner, and it is our hope that it can also make the hosting and delivery of the content to readers smoother and longer lived as well. If all this still seems like a lot to parse, Reclaim has made a great video overviewing the concept of containers and how they work in their Cloud system. And for any scholars, centers, or publishers out there using Reclaim, be sure to check out their catalog of flex courses. This one fit my needs at SUP to a tee!