Let me get one thing clear here. SEGA should not release a “Dreamcast 2” - at least not in the traditional sense – it just wouldn't succeed and it would cost SEGA lots and lots of pennies. But, SEGA should take a leaf out of Nintendo's book, and resurrect the Dreamcast in a similar way to Nintendo has with the NES mini. Some of you may know that SEGA actually has already attempted something similar to the NES mini with the SEGA Genesis (Mega Drive). You can go to a shop now, and buy a cheap, third-party developed, badly-emulated box of plastic that looks nothing like a Genesis, and use it to play a number of built in games.
SEGA can avoid this mistake with the Dreamcast. And here's how…
Step 1: Recreate the Hardware
This really isn't a lot of work. It would take a small team of proficient electronics engineers to take the original Dreamcast circuit boards, shrink them down just by using modern components, and add an HDMI output and WiFi (the modem can be easily emulated). Adding these two features isn't outside the capabilities of most electrical engineers, we all know an HDMI output for the Dreamcast exists, and I myself have examined the DC modem, and emulating it using a WiFi adapter is certainly doable. In fact, the man behind the HDMI output has already designed a half-size Dreamcast mainboard. If one very talented hobbyist can do it, I'm pretty sure a small team from SEGA's arcade hardware department can!
They should probably add in an additional modern ARM chip to power the system firmware before handing off to the original SH4 processor to allow more flexibility. They could also double the RAM, which would allow the Dreamcast-related NAOMI Arcade games to be played too. It's worth mentioning that the SH4 processor used in the Dreamcast and the counterpart PowerVR graphics chip will both have their patents expiring over the next couple of years. Re-licensing these old chips will be cheap.
The GD-ROM was just essentially a modded CD-ROM drive, so those components are easy enough to acquire, and then you just need a plastic shell around the whole thing. We're talking a small team of engineers a few months to get a prototype together. The controller can literally be a re-issue of the original one. Maybe they could make one with the cable coming from the top (controversial!). Hell, they could make nice rechargeable VMUs with modern screens.
Step 2: Build a Digital Store
The SEGA Dreamcast has one of the greatest back catalogues in console history. Franchises like Craxy Taxi, Shenmue, Skies of Arcadia, Phantasy Star Online, Soul Calibur… the list goes on for a while. These games exist, they don't need to be developed, or ported, or brought up to date, they just need a new platform to be showcased. They're almost free-money for SEGA, if only SEGA had some vehicle to distribute them.
Step 3: Open Source the KATANA SDK
This is essentially supporting homebrew on the “Dreamcast Mini” - it would be an extra selling point. Along with the outstanding existing back-catalogue of Dreamcast games, SEGA could take a cut of homebrew games purchased from the store.
The KATANA SDK is the original developer kit for the Dreamcast, and the source code for it is surely just pointlessy bit-rotting somewhere at SEGA's offices. It would cost them nothing to post that code on GitHub and start accepting improvements and updates from a community of developers.
That's it. A simple 3-step plan for SEGA to make a bunch of money on a load of stuff it basically already has. A new revision of the original Dreamcast hardware, shrunk down a bit in the same style as the NES mini, with modern components and HDMI and WiFI connections. A digital store for selling the existing back catalogue of games, and for taking a cut on new homebrew releases.
Most importantly though it should focus on quality – not cheap crappy third-party emulated rubbish – but a serious, sturdy little machine with the resilience and design of the original.