Start-up planned - C/S Strategy - Help & Discussion

Publicly discuss strategies, opportunities and best practices using the Lobby Suite with other professionals in this ever growing industry.

Start-up planned - C/S Strategy - Help & Discussion

Postby Guest » Sun Jan 16, 2005 2:20 am

Hello all you HT/Automation experts

I'm new to this forum and found it to be an excellent place to discuss strategies for
custom home installations. First off, sorry for my poor english. I'm from Switzerland
and my native tongue is kind of "swiss german". That is not Swedish ;)

For you long-time automated guys I must be just another new n00b coming across that tries to raise
business out of this pc based entertainment market. Well, actually I am as I'm in train to seriously start to write
a business plan for custom client/server based info-/entertainment systems. It's a decade now
beeing in the IT/Network business. As a PC enthusiast, accompanied by hobbies like music, hifi, digital lifestyle and the like,
the soon emerging convergence of media and technology was a huge wave I captured pretty early, so I finally came to
conclusion there should be enough experience around to start-up my own local business. Furthermore, I turned out
to not perform very well when working for another boss and I get sick and tired of the IT monkey-business
and I need to do something that really motivates me and this is exactly what brighten my eyes up. Music, quality,
enjoyment, hi-tech. My enthusiasm is almost boundless when thinking about what is possible and what big market opportunity
it is. I really want to hit the market with...

..installing custom, client/server based A/V and infotainment/automation systems.
Using the latest in standard PC technology and pro audio gear I plan to incorporate an own product line using
tailor made desktop cases with 7" (from a-tech fabric.). I try to position in the higher-end field on just every aspect, just coming in at the point where
cheapo-markets cant deliver the studio level pro-audio hardware and off the shelf MCE solutions are limited in flexibility/performance.
Puristic, aesthetic product design, state-of-the-art SQ amd PQ, easy usability, flexible upgrade path along with aimed long-time customer satisfaction and support/contracts are just some terms that come to mind.
Customer=King is the approach. I think you get the picture. For audio, I will probably use RME audio gear exclusively. But generally I'll try to keep independent of manufacturers to always use the best and to be able to react fast on moving market situations, but maybe some strategical partnerships with vendors/installers/manuf. would be wise, in a long-term shot, not only for the lower purchase prices.

As I'm located "overseas" I have absolutely no hesitations to share all my thoughts on how
I plan this buildup. Over here this is all very exciting but still horrible unknown stuff and the mediaserver/automation wave
will probably hit us very soon like a delayed tsunami (don't wan't to offend anybody;) I really have some feeling in
the guts to be able to start-off with some help and advice (and some initial financial injection).

Well I hope to get some thoughts on my plannings and ideas.

I decided to go with separate clients for each main function because audio, video and game apps demand different hardware setups imho. The dedicated server has a complete different setup anyway. Since high-end hardware is rarely to be found on-board it uses additional cards and ext. gear hooked up, which makes a highly integrated htpc vulnerable and potential unstable or poor performing (stuttering playback, audio drop outs) due to hardware issues or unpredictable issues (i.e. IRQs, memory leaks) in the planning phase. I must admit I have never built a custom integrated high-end htpc so far, but as my powerful workhorse is not able anymore for 100% reliable and stutterfree audio playback (to much software and clutter on the wkst now) I guess this is too hard of a goal when simultaneously playing back hdtv while recording a show on the second tuner. Plus having good ui response. Since most peripheral is still on pci the bus and the legacy pc hardware architecture is the bottleneck, not the ever increasing cpu perf. This led me to the conclusion to separate tasks, which also improves space for powering/cooling equipment (dead) silent.

Another goal is, and this may clarify my point of view, I want all the audio processing on the htpc, that connects to pure analog power amps or active speakers.
To complete my personal system (maybe this is the future demo device / showroom) I will switch the receiver for a analog multichannel amp. The whole system should only consist of htpcs and amps, doing all the digital domain inside the case. Possible external gear like Sat/Xbox (which I don't have) can be hooked up to the audio client pre/pro audiowise and to the video client pre/pro videowise.. I just try to go with a minimal set of external A/V gear to keep it simple and pure yet easily upgradeable.

My RME soundcard uses 2 pci slots (for 6 analog outs) and I have no idea how people incorporate dual tuners, soundcards, raidcontrollers, holo3d, vga and other interface cards into one 2U case without having hardware issues or serious heat/noise issues. Maybe geniouses or its just because most people use the onboard SP/DIF to ext. decoders, which is fine too. I always have/had problems with too many PCI devices and IRQs (sharing) when building PCs, its flawed no matter how many GHz P4 installed, you can't even use all slots w/o issues. Maybe this is gone with APIC Bios but I'm not sure if that really helps to bypass the basic problem of having less underlying hardware irq lines. Maybe somebody can point this out? No more A/V glitches with APIC due to resources/IRQs restrictions?

With such an all-in-one htpc I have figured out I would have to build a large tower with watercooling and a massive passive radiator to fit the bill of dead silent and high-end perf. And I'm only aware of some lian-li tower cases that looks how I like it (clean, no usb/fw front, no fans, no blue leds...). Optically and physically harder to integrate in an environment where the server is in sight with the A/V equipment. Put all this into account and saw that it is not even much cheaper than the separate method, since the case/psu and silencer components alone where over 1000$. Maybe 500$ if fans where allowed :-)

That beeing said, I will follow a modular concept using a audio client, a video client, and a server.
I plan to use the same case for every purpose to keep the look consistent and sleek. The audio client and video client with 7" touchscreen, the server with 4x20 VFD for status and health report. I try to operate everything absolutely fanless, with the storage at the server side in external closed no-noise, no-vibration Sata cases.

No optical drives on the clients allowed and I also try to find out if its achievable to build a complete diskless, remote boot GBit connected network client. Is it possible and practicable to have the client diskless remote booting from the raid server?

Some thoughts in favor of this strategy from my side would be centralized disk operation- and failure management, no annoying noises from optical/hd in the audio signal path and for general flexibility and system maintaining. The diskless client is a recent idea for a further radical system approach so I don't know really if that can be done and which technology to use to achive the goals (SAN/NAS, iSCSI, Wake-on-Lan, SATA Raid 5 simple shared). The clients probably need a lot of RAM to build a Ramdisk and other than this I have no idea how that would perform. Somebody feed me input on this? For instance, the video client must be able to operate from the server storage, playing/streaming video files and maybe recording a show to the server storage all at the same time and in real-time.
If the client is in stand by mode, when waking up, does it have to load the complete image first before it starts to boot rendering the fast standby mode useless? With about 100MB/sec I think it can be very fast, but what is Ram requirements? How is performance in such an environment? The overall bottleneck will probably always be the network with a full-featured fast "thin-client".
This would be no problem, as long as it is enough fast with good head room. Questions over questions.

The audio client would act as a pre/pro, accepting the analog inputs from the video client or other ext. gear (if not directly connected to the amps) The server would act as the central remote control transceiver and trigger remote events on audio and video clients over the GBit link. The server would also be the uirt interface for ext. IR gear. No Wifi/IR planned on the clients since Wifi could interfere and potentially harm SQ and PQ.

Each client should have a dedicated 100/1000 interface with direct cross-over RJ45 to the server. Another question would be as ML newbie, how to distribute the software suite across the server and client systems and where to place girder? I plan to use only Wi-Fi input systems, probably full-featured XP tablets (again, to offload some local tasks and flexibility) or ultra-portable laptops with long battery. Maybe I'll try "dumb" airpanels and ppc too, as some of them look cleaner and are easy to operate on the couch than a laptop/pen tablet. Probably want to find out myself and demo them all anyway as some customers may prefer a pronto over a large tablet...not finally decided which to use for my first installation but its gotta be based on Wi-Fi, over a wifi router connected to the hardwired mediaserver (optionally with ext. wifi antenna too).

I have some X10 light here too I want to integrate but this is currently controlled with IR input from the pronto. I'd like to exchange the pronto for some wi-fi gear so the server/uirt should takeover the light control, or better get rid of the IR and implement a computer/network based X-10 interface. The idea is, the server should do all the IR sending using an ext. IR transmitter on a cable to shoot at the IR equipment.

I hope you see my intentions. Just dedicated (diskless) full-featured clients doing audio or video only and offloading everything else to the separate server. (system maintaining, remote control/interface, server tasks, firewall/antivir, "middleware", storage, whatever).

I'm not sure either how I would implement multi-room (new to this) but if that means to have "hundred" feets of unbalanced audio cables throughout the house I would prefer a separate networked client and amp/active spkr. in that particular room, unless the house isn't built already and cabling is planned at early stages. Maybe it's just me and I miss some major point.

If a fast responding, reliable diskless client is not possible within human price ratios, what would be the best alternative? Maybe 2 disks mirrored in each client, with the PVR capturing to the local drive? Systems/Apps local on clients and media/data off the server? I would not like the idea of a single Sata drive in a 24/7 client and would rather mirror it using a simple onboard or software raid.

Some useful input on diskless operation is very appreciated as I have very little experience with WoL and diskless clients in networking thin client environments and no clue how to setup in high demanding full-featured clients. As I plan to work with SATA storage instead of SCSI (2TB data is not funny with scsi) I find it compelling to have the less reliable short-life disks centralized in a large storage subsystem for all related computers, to minimize risk and data loss. Just one virtual logical huge disk with dedicated partitions for each client on one large Raid 5 with some hot spare, that would be nice. With email reporting to me so I can inform the customer that his audioclient experienced a failure he probably didn't get notice of and that I have a new disk already and looking for an appointment to fix the issue. ;)

Anyone? Some input on my concrete business plans are welcome too. I've seen some other guys here started similar business a year ago. How have you done in the meantime? Thanks for those that have found the time to make it through this post.


PS: from the "old europe", esp. Switzerland still dreaming of HDTV (its a myth) LOL.

edit, addition: Other than some important strategic/financial issues stated above I feel so ready to fire it all off like never before :)
It gradually all ties together from custom machined case to the hard/soft used and a matching concept for a business plan that could be well extended (Maybe shopping/boutique/mall a/v installs).

Looking forward to lurk around often here on the forum as the software suites get more professional/functional and visually appealing. It will show for me as I learn if ML/Girder can do it all for me... I almost hope so, it looks great at first sight and as I've seen there are special license deals/branding options for custom installers. Looks very promising. Hehe, I didn't even manage yet to download and actually try it out, mainly because I have no demo environment and have to install on the already overloaded workstation which would not reflect a c/s situation at all. Also I have to review the J River thing as I sticked with winamp since ever and perfectly happy with, but prefer a external database/archiving within a neat front-end and using the player just as a raw player (with perfect gapless and source in plug-ins)
Maybe the ML thing will change my view and I should d/l it right now, even if my unemployed a** has made through another sleepless night.

Thanks to all dedicated developers and engaged users that share the thoughts!

User avatar
Posts: 20
Joined: Tue Feb 14, 2006 12:41 pm
Location: United Kingdom

Did you get this off the ground

Postby JJPhillipson » Tue Mar 07, 2006 8:45 am

I am thinking of starting something similar for the UK market, I really believe a successful business can be created, ML gives so many opportunities.

I hope you have been successful, please let me know how you are getting on.



Return to “Professional Installers & Dealers”

Who is online

Users browsing this forum: No registered users and 1 guest