Here is an update from the Devs, sounds like things are progressing nicely.
-------------
ethoFS Development Update:
We have successfully deployed 35 ethoFS nodes with a simple installation script between alpha users and the dev team to stress test and benchmark various aspects of the ethoFS system and the new MN consensus mechanisms that will be used to move to a completely decentralized version of the node dashboard. We are in the process of updating ethoFS for a second release to the alpha group with many updates to continue testing, and then begin testing the content upload front-end. Once we are comfortable with stability we will start to open it up to more community members to join the testing group.
What does the ethoFS system do?
ethoFS is our implementation of the ipfs protocol to manage/store/deliver decentralized content on-demand. By combining the Ether-1 blockchain and ipfs into ethoFS, we get a decentralized, immutable method of hosting and delivering all of this content.
How will we be decentralizing the node dashboard?
We have developed a brand new consensus mechanism using ethoFS that allows us to manage and verify node health and information in a decentralized fashion. In the future, you will be able to save your node information to a smart contract and ethoFS will take care of the rest!
How does ethoFS manage all of this information?
This is a complex question as storing/hosting information decentrally forces us to use other information storing methods in the place of a classical database. Heres a quick example; if I am storing usernames and email addresses in a database for a user login system, it becomes very difficult to update and maintain this database of users when it is stored on a vast number of nodes across the ethoFS network as updating it locally would cause all of the information to be out of sync (this is a classic problem when dealing with immutable, decentralized data).
What do we replace databases with then?
There is a fairly new data type that is very cutting edge as it is only really natively applies to decentralized systems. CRDTs or Conflict-Free Replicated Data Types are the answer. Each node is responsible for making sure any updates to the data are checked/propagated to each peer to allow us to have a data type very similar to the classic database but is distributed across the entire node network and constantly kept in sync. We are developing custom ethoFS CRDTs in order to maintain all of the information we need about the nodes. When we start to allow more complex websites to be deployed to ethoFS these CRDTs are going to play a very important role in how information is stored and maintained.
How do we make sure any hosted content is on a live and available node?
This is another very important question as determining where and how the data will be made available directly affects the ability and time it takes to deliver this data to the end-user. We have built a custom content replication/propagation system in ethoFS using the ipfs content pinning method. Natively, ipfs allows for pinning of data on a local node to assure that it will always be available to any other node on the system that requests it.
What happens if the node with the pinned content goes offline?
If the content is only pinned to one or two nodes and those nodes goes offline or become unavailable, then the data would theoretically not be accessible across the network until either of the nodes are available again. The problem is a slightly more complex than this as there may be cached versions of the data on other nodes but this is the basic idea.(edited)
How does ethoFS fix this problem?
Using another one of our custom CRDTs we are able to make sure that any uploaded/hosted content is replicated and pinned to a custom defined range of nodes. Using the data we have gathered during alpha testing we have very strategically refined and set this replication factor to allow for content to always be available very quickly to the end-user. If one node with pinned content goes offline, the consensus system recognizes this and quickly replicates and pins this content to another available node.
Please feel free to ask any questions about the system and we will try to address.