Sometimes you need to do some due diligence, just as a check on "are we (or or current vendor) already doing it as well as we could?"
There are *some* cases where it makes sense to run your own CDN. If you already have a significant installed plant (co-lo in the right locations, large server base, adequate bandwidth) and either specialized needs or perhaps you only need to distribute a limited amount of data. $work used to run its own (rather small) CDN a few years ago because we met the above criteria and weren't pushing a lot of data, all the time. As our needs changed, we re-evaluated and went with multiple CDNs over the past few years. For this kind of exercise, you should actually "run the numbers". You need actual costs to figure out if this is a good idea. Back of the envelope (BOTE) costs can tell you rather quickly if this is something you should do. And, if your BOTE numbers are close to a commercial CDN, then you need to re-visit and refine those into actual committed costs, just to be sure. Look at how much data you want to distribute, how often and look especially at your expected peak bandwidth from each location. Then look at your costs to host, costs to purchase/lease/reuse the needed servers and storage. Do you need to buy or use a geolocation service? Will you need to build or buy load balancers? Would you need to upgrade or replace existing network links? Would you exceed your bandwidth commits and hit overage charges? Would you need to expand a datacenter or co-location cage? Then look into the often overlooked costs: people and software to operate and monitor all this. Do you need front end software to allow non-technical people to load content? What software will you need to push data from the master site to all the endpoints? (How much does *that* bandwidth cost?) Even if you aren't actually planning to build you own CDN, this exercise can help you figure out if you're paying too much for an existing service, or help you shop among potential providers. You can use this as an exercise to create a checklist of features and services that you can use to evaluate potential vendors... "How many distribution points do you have? How are they monitored? How long does it take for my data to get to all endpoints and be ready to be distributed?", etc. Good luck, and please share your results! On Wed, Nov 21, 2012 at 8:57 AM, Jeremy Charles <[email protected]> wrote: > I’ve been asked to look in to what it would take for my employer to > build its own content delivery network hosted on our own hardware at > various physical locations around the world (all two of them, soon to be > four). The intent is to host our own content, not anybody else’s.**** > > ** ** > > If you’ve done this or looked in to doing this yourself, I’d be interested > in leads regarding what you used. Google is showing me plenty of vendors > who are willing to sell me the service of hosting my private CDN (say > what?), but I’m not finding leads on what tools I would use to build it > myself.**** > > ** ** > > ** ** > > ** ** > > For completeness: I don’t yet understand why the business wants to do > this in lieu of continuing to use the CDN we’ve been contracting with for > the past few years. I’m asking about that in parallel with being a team > player and figuring out the answer to the question they’ve asked me. :-) > **** > > ** ** > > ===**** > > Jeremy Charles, [email protected]**** > > Epic's Computer and Technology Services Division**** > > ** ** > > Phone: 608-271-9000, Fax 608-271-7237**** > > ** ** > > _______________________________________________ > Discuss mailing list > [email protected] > https://lists.lopsa.org/cgi-bin/mailman/listinfo/discuss > This list provided by the League of Professional System Administrators > http://lopsa.org/ > >
_______________________________________________ Discuss mailing list [email protected] https://lists.lopsa.org/cgi-bin/mailman/listinfo/discuss This list provided by the League of Professional System Administrators http://lopsa.org/
