Synchronize CAD Files over a Wide Area (CAD Manager Column)31 Aug, 2008 By: Robert Green
Planning and proper tools keep data current and users happy.
In last month's "CAD Manager" column, I discussed managing multioffice environments and touched on some of the related challenges. Perhaps the biggest headache CAD managers experience in multioffice environments is keeping the entire company working with the right data and somehow keeping everything organized. Synchronizing all this data is a high-stakes job — if you lose control of your data, you have big problems.
Software Tools for Data Synchronization
In this month's "CAD Manager," I'll define the problem of data synchronization, examine some useful hardware and software tools, and help you analyze your company's needs so you can put the right tools in place.
Users Need Files Fast — or Else
When users arrive at work, they want to open the files they need and work with their data — now. But when users are forced to open files that reside on a server on the other side of the country via a slow data connection, the process is anything but fast. When users perceive how slow the file access is, they take matters into their own hands and make copies of files to store on their local machines. It's faster that way. This illicit file copying leads to a host of problems, which include at minimum the following:
Parallel revisioning. This problem occurs when unauthorized copies of the same file exist in many locations. In this scenario you never really know who has the latest file version, and all sorts of errors occur as various team members work with different versions of a file.
Broken file links. In cases of externally referenced files or parts within assemblies, unauthorized file copying means somebody is referencing an old version as he or she works on the geometry adjacent to the attached file.
Forgotten updates. When a user copies a file to a local computer, he or she usually intends to update the original file later — but occasionally forgets. This oversight means that the supposedly current version of a file is actually sitting on someone's C drive.
All these cases point out the dire need to control your CAD files in a way that is relatively transparent to users but still delivers good speed. I've found that keeping things moving fast would have prevented 95% of data problems.
Quantifying the Problem
Now that you understand why data synchronization is a user-driven, speed-based problem, you need to understand the environment in which you operate, how your workgroups interact, and why real-time data sharing is critical. Understanding these factors will help you select appropriate data-synchronization tools. Following are the key parameters you'll need to consider and the questions to ask yourself:
Network environment. How many servers are in your wide-area CAD network? Many times each remote office will have one CAD server, so you can think of a server as being a geographical location.
Work team sizes. How many CAD users use each server? This parameter is critical because the more users per server, the more data traffic ultimately will be handled by that server.
Connection bandwidth. How fast are the connections between the servers in the CAD network? This parameter is critical because the faster the connections between the servers, the easier it is to move data between those servers.
Work coordination between servers. How often do users located on different servers need to share data? Do they need to be connected to the same file (such as xref files and underlying GIS databases) in real time?
Average file size. What types of CAD files do your users need? Users who typically share only small DWG files have very different requirements than a wide-area team working on a GIS map of an entire county.
Available data-synchronization tools range from hardware to software, automated to manual, and inexpensive ($50) to pricey (more than $10,000). To decide which types of tools are appropriate for your environment, you'll need to understand each type of tool and match its capabilities to your environment. There are no perfect answers for any scenario, but some careful analysis should narrow the choices quickly. I've listed some file-synchronization products I've used in the "Software Tools for Data Synchronization" sidebar for your reference. Below are the types of tools available with some recommendations for analyzing your environment for each:
Hardware synchronization. These types of tools actually are very high-end network devices that buffer, cache, and optimize servers' file-transfer capabilities. For this approach to work, you need a hardware device at each company location (or branch office) so the various servers can communicate with each other using the optimized/cached file transfer.
Hardware synchronization offers the advantages of letting users work as they always have, accessing data in real time, and obtaining the best file-transfer speeds across low-speed networks. The disadvantage is the high price of equipping your entire wide-area network (WAN) with the devices. If speed is the driving concern and coordination between many users in many sites is critical, then hardware synchronization offers the best choice if you can afford it.
Automated software synchronization. As with hardware synchronization, software synchronization runs on each server in your network, but it creates a mirrored file database on each server rather than caching data between servers. Because the solution is software based and requires a good bit of analysis to keep up with file changes, it creates an additional computing load for each server in the network and thus requires more-robust servers.
Software synchronization offers the advantage of letting users work with files over the existing network as they always have, at a substantially lower cost than hardware acceleration. The disadvantages of this type of tool are increased server load, greater configuration and maintenance for you (or the IT department), and somewhat lower speeds than hardware alternatives. If having consistent network access to all users on all servers is critical but you can't afford the hardware alternatives, then software synchronization probably is the right choice.
Manual software synchronization. File-transfer protocol (FTP) tools can perform data synchronization between servers on the cheap ($50 per seat or less), if you're willing to learn how to use them and to be personally responsible for keeping files synchronized. Some of these tools have scripts you can create to automate directory lists, but you'll still need to make sure the synchronization runs. These tools are very slow because they use little or no file compression (as hardware solutions do) and can't perform byte-by-byte file comparisons (as automated software solutions do).
Manual use of FTP software obviously is the least expensive option, but because you must perform synchronizations yourself, you'll never achieve real-time synchronization as you would with hardware or automated software systems.
If you need to keep data synchronized periodically, can afford the longer transfer times, and can keep track of the process, then manual software utilities might be a solution worth investigating.
As you may have guessed, all the data-synchronization methods I've discussed require some close coordination with your IT department. I've found that software synchronization actually requires the most IT configuration work, followed by hardware devices that tend to be hands-off after initial installation.
Gain the cooperation of your IT department by communicating the technical problems you face, including the size of your files and the challenges of your work environments.
Keeping your CAD data synchronized within a WAN requires equal parts detective work, planning, and proper technology. For best results, research the topic carefully and involve your IT department in the process. Anything you can do to stabilize your data and deliver better speed for your users will help you minimize data-related CAD management hassles and keep your users happy at the same time. Whatever you do, don't ignore your data-synchronization problems because they only grow larger with time.