I'm trying to understand total US or worldwide spending for the servers in data centers used to deliver digital media, especially video and audio. This can include data centers used by companies like Comcast to deliver commercial movies AND those used by companies like You Tube or MySpace to deliver media created by customers. I'm willing to use estimates along they way. As a high-level example of what I mean by estimates, if one could estimate both industry revenue for media delivery and the percentage of revenue spent on storage and compute servers in the data centers supporting this revenue, the result would be spending on the servers. IDC estimated the market for content delivery at $500 million, but that doesn't include companies like Google with their own data centers...and doesn't try to convert revenue to storage and compute server spending.
Another approach would be to know what one or more top sites spend (Yahoo, MSN, Google) and build a model from that.
Yet another approach is to determine what the top server vendors (IBM, HP, Sun, Dell) sell exclusively in this segment and extend from there.
The best approach is to follow 2 or more paths to see whether multiple approaches deliver a similar result.
The foundations of another way to do determine worldwide market potential for the storage side of video ingest, storage, and delivery follows. This may be easier:
• Start with established suppliers of broadband service by country and/or by region
• Determine existing subscribers
• Determine growth of subscribers
• Do not double count subscribers (i.e. if four vendors have 10% market share and are growing at 50%/year, do not assume that 5 years from now, all have grown so much that the total penetration is over 100%!
• Define pragmatic options for storage requirements per subscriber for an emerging web-based storage service option (nPVR or dynamic storage used by individuals)
• Estimate rollout rate of the storage service (with multiple scenarios for rollout rate)
• Estimate sign-up rate for storage service (with multiple scenarios for sign-up rate)
• Determine implied storage requirements
• Use $US prices for $/GB to forecast market potential
The best approach will attempt to use at least two approaches to test whether the magnitude of the forecasts is reasonable.
The deliverable must be clear, with open, adjustable assumptions.