Local Cloud / Server-side software download

qa

#1

I know that developing the software for every desktop oriented operating system is unreasonable, but I also do not like that cloud based software means that an update comes out and the old version - including rough edges and unpolished pieces - is permanently unavailable.

Sometimes cloud updates also lead to wasted time tracking down what changed in the plan when in fact it was simply that the cloud side updated its handling of a design and the fault did not lie with a locally-made adjustment. Other times it means quite simply that a consistent result is impossible, like maybe I wanted and got a little charred bit here and there until the cloud update to ‘better handle (your obscure material).’

To address this problem and others, will you be releasing the server-side code or vm template for users to host? Obviously this would be for users more knowledgable in computers, who would be able to install whatever operating system, patches, and local dns captures that would be necessary. These users, in turn, could become a valuable archive of prior versions.


#2

I don’t think you have to worry too much about material handling. You will be able to create material profiles (with your own settings) that will be maintained through updates. You also have the ability to set power and speed to obtain your desired effects/results.

As far as them releasing the code to run virtual servers. This is definitely not going to happen. It has been discussed at length here in the forms. And Dan has said himself that they will not be doing that. I can see a hack being developed ones the firmware is released but that voids your warranty. And also the software is a webapp so it should be compatible with whatever os you’re using and on any device (not 100% on this)


#3

I have not had the experience of an update breaking an existing mission critical process but do appreciate the concern that cloud updates and versioning could pose. I would like to play my music that I bought whenever I want to, especially out in the woods where there is no internet. Apple doesn’t make it easy. The cloud feature of the Glowforge is key to the price point so I am willing at this time to forgo a little independence and hope for the best. At the point where someone offers a Glowforge local software control that functions like DD-WRT does for a router, I’d be tempted to try it, after I have purchased a few other Glowforges and can spare one to play with. Everything that I read from the Glowforge team is end-user centered and I trust they really would upgrade only to make better. But, then, I’m not thinking of using the Glowforge as a critical tool for work projects so I can afford to be optimistic.


#4

I do not believe this has been discussed much at all.

I am not asking for:
1: Custom firmware
2: Desktop support
3: Development of offline support
4: Use of unofficially developed software
5: Almost any time from Glowforge employees

At simplest I am asking for a .tar.gz that, from the existing server hosting the live CAM and motion planning, could be made in about 2 minutes. I’ll write a script for free, and they could cron job it to run once a week or whenever it detected a change to have a new .tar.gz created and uploaded to a publicly viewable ftp. In other words, just the backup that glowforge or their cloud provider is probably already making.

If the web app is not doing requests via ajax to that part of the code, then also an advanced setting to point at a different cam planning provider or a copy of the web app might also be needed.

If, and this seems likely, there’s some google or amazon api in there that requires hosting the code at google, then obviously the task and costs of getting it there would lie on the user who wanted to run the old version. If not, I can easily find some unused hardware to handle the processing of a few projects - although it would be a little tougher if a few projects honestly need more than 1Gbps random disk IO, 36x3.7ghz, and 256GB of RAM.


#5

Thank you for the clarification. I should read topics more carefully and original posts.


#6

But essentially you want versions of their software? or you want them to host old versions of their software? As a company its not in their best interest. The reason it is cloud based is so they can push out changes to the masses and know their users are all on the same version. This is really important since all the features aren’t going to be in the initial release. Im confused by the need to"backup that glowforge". The firmware probably won’t change much once its release. Since the motion plan is coming from the cloud all the unit itself has to do is execute it. And if there firmware does change you would have to install it manually. Everything else is handled by them in the cloud.


#7

I think the question was brought up before to have a Local Cloud/Server to run the Glowforge. @dan has basically said its possible but not currently a priority. The quote is below.

On the flip side, if you are willing to do the work for free, maybe @dan will change his mind. :wink: But I wouldn’t be surprised if there are other issues to get this working that they just don’t have the bandwidth to deal with as of yet.


#8

I feel one glaring issue has not been discussed: security/copyright. As a company it is upto glowforge to maintain it’s security, privacy and it is their “right” to protect themselves from breaches of copyright. Most modern hardware/companies do this via updates(but I don’t need to tell you, this is for less savvy guys) as well as adding/adjusting content services. Due to this reason I think their trusty techies/lawyers would advise against it even if 1% of users wanted either glowforge maintained images of old versions or to public ally released images(that may have bugs/exploits)

Sorry that it’s not what you want but I doubt this will happen in the next 3+ years, their long term plan will likely evolve and might enable this sort of functionality.


#9

Software, in many different versions bugs and all, is traditionally desktop side and archived as a cd with an installer. I should not neglect to mention that it is normally a binary blob at that point, obfuscating the operation and to reverse engineer it one must intentionally and knowingly go out of the way therefore the law protects the company’s interests. If the core of the server side software can come in this form, that is fine with me. If it is php or something that is basically open source if anyone else gets their hands on it, I still have hope since the firmware was put under the GPL.

In addition, GlowForge is not a software company. Is their plan to perpetually sell more and more machines to pay for the bandwidth and cloud computing of all machines sold up to the current time? If this long term plan sounds familiar, look up the definition of a ponzi scheme. I think GlowForge could look at the 22 year old, billion dollar enterprise that is built entirely on giving away free software. RedHat’s model is selling support subscriptions and certifications.

The GlowForge cloud will inevitably get costly, just not on a per-user basis. They could provide support only for those using the GlowForge cloud, latest version of the software, etc, and several years of free access for each machine sold. After that, a very reasonable subscription (rough estimate around $15-$20/year/GlowForge) would cover their bandwidth and computing costs for the user to have access to the latest release, the GlowForge cloud, phone support for said software, etc. In this way the machines profit margin can perpetually pay for development of new machines, and the software could perpetually pay for advances to the software. Otherwise growth of the software requirements will inevitably consume the entirety of the machine’s profit margin. This is a fact that was not previously seen because in a traditional model a company would not need to shoulder the cost of electricity, bandwidth, cooling etc for running the software of a long lasting device - once a piece of software was no longer affordable based on the number of machines being sold, they just end support for it without the consumer losing the ability to use the product.


#10

If anyone believes the company is going to make all of it’s profit through selling GF units then I have a bridge in NY to sell you. Obviously I’m guessing here but there will be profit from, selling the H/W, potential H/W uprgrades, selling unlock codes for previously unannounced features, materials sales, catalog sales, probably subscriptions to some other services. Their stated intent is not to charge for server time to the initial backers for the original features. That’s great. You want more, you might have to pay for it. If you get GF 2.0, maybe you’ll have to have a subscription service. I know what I paid for and am perfectly fine with that. Anything additional, I might have to pony up some more cash. Such is the nature of every profitable business.


#11

@dan has already stated they will not charge for “currently announced” features. I took this to mean that there are more features that they haven’t announced yet that will have to be paid for, either by subscription or per feature or something. I also wouldn’t be surprised if there is a small fee for selling designs in the Design Catalog (similar to etsy).

But more importantly, @rpegg, tell me about this bridge. :grin:


#12

A bridge in NY you say? I could do with one of those, a bridge from the UK to NY would be better :stuck_out_tongue:


#13

As much as I would like to say that Glowforge exists to put a 3D laser printer on every desktop, I grudgingly take off my rose-colored glasses and know that the economic model that gives the best return will win the day. What the future of the cloud model will bring in terms of subscriptions, what future iterations of the product require more investment, I don’t know. All will lead to comodification of product and services. That’s ok. For the immediate future I see the Glowforge experience to be end-user driven and that will make for some wonderful stuff and happy people.


#14

This just means we want to be careful not to promise anything other than what we have solid plans to deliver.


#15

Hi Dan,
Thanks for visiting!

Any thoughts on the home archiving or data security and the eventually unsustainable model of hardware-pays-for-software? I know I for one wouldn’t want to have dropped a couple thou just to to sit through an add for every model I want to upload.


#16

I don’t really follow your questions @SolesUrsa, sorry - can you clarify what you’re asking?


#17

I will restate a few points.

First, what will you do long term? Cloud computing costs money, more users means it costs more money, and eventually your cloud costs will not fit within the profit margin of your hardware sales. In other words you may need to split your software and hardware budgets and get a revenue stream on both sides, but you cannot just brick the machine of a user who does not want to pay for their portion of your cloud.

To address this and other concerns listed below, I proposed something.

I am not asking for:
1: Custom firmware
2: Desktop support
3: Development of offline support
4: Use of unofficially developed software
5: Almost any time from Glowforge employees

At simplest I am asking for a file that, from the existing server(s) hosting the live CAM, motion planning, and other pieces, could be made in about 2 minutes given a script (which I would be happy to write). It would be a list of installed libraries, kernel version, and a copy of the necessary code and configs to restore the server in case of an outage. In other words, your backup if you have one.

It would need to allow someone with enough experience setting up servers to create an environment where the backup could be restored thereby achieving:
1: Result consistency via access to archived “same-as-the-first-time-I-printed-this” versions of the GlowForge cloud
2: Data security/enterprise options/local hosting potential
3: Third party solutions for the cloud pricing problem ( if enough users don’t like your cloud pricing they would have options which you obviously don’t have to support and you could focus on the software/hardware not this part. I wouldn’t want ad-supported but I’m sure a third party would be happy to provide it and some users would be ok with that. )

There may be hurdles, some already mentioned and others like the potential use of certificate pinning, but even before the difficulty is fully assessed I am curious about your thoughts on everything above. Any chance on archived versions of the cloud or other approaches to the three points it may achieve?


#18

To your question, we budgeted for the cloud utilization over the lifetime of the device at the start, so we can afford to run the servers for the machines we sell for many, many years to come.

Regarding your request - understood. No plans but we’ll let you know if anything changes. It sounds like this is really important to you, so if our current plans don’t meet your needs, mail support@glowforge.com and we’ll regretfully issue a prompt refund.


#19

I won’t be asking for a refund, the GlowForge is far more important to me than software versioning or local cloud. I simply find irony that cloud apps, app stores, and subscriptions decrease the people’s ability to own their copy of software while at the same time they are gaining the ability to manufacture more and more physical things.

In addition, I have am intrigued by any difficulty more so than deterred by it - I sometimes spin up several custom servers on any given day so for me the option to do a local GlowForge cloud would definitely be utilized.

Anyway, thank you for your answer. I hope you consider the options again in the future, but I would far prefer not causing delays in getting the first shipments out!


#20

I think it would be pretty cool to be able to host the software myself. I don’t know what kind if cloud resources it takes to do the processing for the GF (it would be an interesting statistic to say the least), but I like to look at the worst case.

If something were to happen and GF were to have to close their doors (I don’t suspect that will be happening, but just for the sake of argument), what would the plan be for all of the owners of the GF at that point? If it resides on the cloud and the cloud goes poof, do we all own big paperweights? Or what about the scenarios where people don’t have a strong Internet connection, or are limited by bandwidth. Are there any details about the bandwidth requirement or amount of data transferred per job on average?

Luckily I think most of us don’t have to deal with those situations. I am still a geek, though, with a rack cabinet of servers in my house running VM’s (part for work, part for learning) and think it would be pretty cool from an academic point of view to create a ‘private cloud’ that can handle the processing at some point down the road. Really what you are doing is an interesting use case I think, so I suspect others will want to explore it as well.