I am looking for a way to replicate a small and simple relational database (like SQLite) across peers. This should work in an environment with unstable network connections, hence the need for each peer to have a full copy of the database. This should allow a peer to continue working off-line in the event of network failure.
To keep things simple, replication should only have to support the replication of addition of data, i.e. only INSERTs, not DELETEs or UPDATEs.
Does anyone know of a good - and ideally cross-platform - technology or method of creating such a system? I am currently looking at JXTA and JXSE, but I am put off by its complexity and apparant lack of life in its community after the takeover of Sun by Oracle.
Thanks!
Frans
rqlite uses the raft consensus algorithm, so it should be fairly resilient to unstable network connection.
Also, it seems to be possible to configure rqlite to accept reads even in the case of a network failure.
A similar project, dqlite, exists as a library, available in various languages, but it seems less explicit about the event of a network failure.
You may want to explore JGroups for the communication layer if you don't like JXTA. For the replication, I think you will have to implement your own code.
I am working on something similar (though the code is far from ready). I'll describe a little about my intended approach, but whether that is suitable for you depends on some key design points you'd need to consider. I am not aware of any ready-built projects that will do this, unfortunately.
In particular we'd need to know what language you wish to use, or which languages you'd rather avoid.
Also, consider how you intend to do peer dicovery - can you set up trust between node pairs manually, or do you want them to auto-discover?
Presumably all peers may insert data?
If you are able to use PHP, and are happy manually peering node pairs, then my approach may be of interest. Set up an ORM such as Doctrine, Propel or NotORM, and get each node to regularly sync with an internet time source. For each new row in a db, grab the data (either in an array or ORM object), serialise it, and push it out to all nodes that you have a trust relationship with. Where a push fails, keep a note of this and retry at periodic intervals (potentially giving up after a remote node fails to answer a large number of retries).
Pushes can either be kicked off by your application that creates the row, or can be called by whatever scheduler is available on each machine. A push message can be XML, or for simplicity can be just a POST message containing the new row and whatever metadata (e.g. timestamp of save, so as to resolve INSERT order from several nodes).
If your nodes do not have static IP addresses, they could be registered with a dynamic DNS addressing service so as to allow each node to stay in touch with peers even if their IP changes. You might also consider adding a message signing system, to ensure that messages between nodes are genuine.
Related
I find it difficult to convince myself the advantage of using complex design like DynamoDB over simple duplication strategy.
Let's say we want to build a distributed key/value data store over 5 servers. (each server has exactly the same duplica).
Eventual consistency system, like DynamoDB, typically uses complicated conflicts reconcile, vector timestamp, etc. to achieve eventually consistency.
But instead, why couldn't we simply do the following:
For write, client will issue the write command to all the servers. So all servers will execute the clients' write command in the same order. It will reply to clients before servers commit the write.
For read, client will just do a round robin, only one server at a time will take care of read command. (Other servers won't see the read command)
Yes, client may experience temporary stale data, but eventually all replica will have the same dataset, which is the same semantic as DynamoDB.
What's the disadvantage of this simple design vs Complicated DynamoDB?
Your strategy has a few disadvantages, but their exact nature depends on details you haven't covered.
One obvious example is dealing with network segmentation. That is, when one part of your network becomes segmented (disconnected) from another part.
In this case, you have a couple of choices about how to react when you try to write some data to the server, and that fails. You might just assume that it worked, and continue as if everything was fine. If you do that, and the server later comes back on line, a read may return stale data.
To prevent that, you might treat a failed write as a true failure, and refuse to accept the write until/unless all servers confirm the write. This, unfortunately, makes the system as a whole quite fragile--in fact, much more fragile (at least with respect to writing) than if you didn't replicate at all (because if any of the servers go off-line, you can't write any more). It also has one more problem: it limits write throughput to the (current) speed of the slowest server, so even if they're all working, unless they're perfectly balanced (unlikely to happen) you're wasting capacity.
To prevent those problems, many systems (including Paxos, if memory serves) use some sort of "voting" based system. That is, you attempt to write to all the servers. You consider the write complete if and only if the majority of servers confirm that they've received the write. Likewise on a read, you attempt to read from all the servers, and you consider a value properly read if and only if the majority of servers agree on the value.
This way, up to one fewer than half the servers can be off-line at any given time, and you can still read and write data. Likewise, if you have a few servers that react a little more slowly than the rest, that doesn't slow down operations overall.
Of course, you need to fill in quite a few details to create a working system--but the fact remains that the basic concept is pretty simple, as outlined above.
I am comparatively new to ZeroMQ and would like some suggestions regarding its it's internal architecture.
I am planning to use ZeroMQ as a messaging framework for my work. The basic idea what I want to achieve is to be able to dynamically scale the infrastructure based on the load and computational capacity required to achieve a particular workflow deadlines.
So,if if there is a necessity to add more nodes, then the application spawns new nodes and the messaging framework should be able to incorporate the changes as well. I should also be able to be point where the additional computations should occur or how the framework dynamically adds the new nodes (if any). The event on a particular node decides subsequent actions to be performed on other nodes. Here is my scenario or my stack that I am thinking off, but wanted to know if it makes sense:
User applications
ZeroMQ messaging
Squid-Content based routing
Overlay
Physical Substrate
I am bit skeptical about the above stack as I believe ZeroMQ helps one to achieve most of the functionality and therby thereby making it simpler.
Few points about my stack:
Physical substrate are the total number of nodes that are available for the computations or as data sources.
Overlay is a logical network that is built dynamically upon the physical network based upon the closest nodes available for a particular workflow. i.e. if two nodes exchange data frequently, then those two nodes are placed logically close to one another. Is a separate overlay like CHORD etc required when we use ZeroMQ?
Squid is basically used for content based routing. Is Squid required when we use ZeroMQ?
ZeroMQ messaging is for the communication between different nodes for an application.
Basically, what I wanted to know is whether above stack can be made simpler given that ZeroMQ has richer functionalities. If so, can someone point or share the thoughts. I am however going through the documentations of ZeroMQ, I am finding it a bit difficult to understand the intrinsic design of ZeroMQ. Please help.
Thanks
There's so much specific to your use-case here that it's almost impossible to give any definite answers. ZeroMQ is not a direct replacement for the concepts you've built into your architecture, however it may meet the goals you're trying to meet depending on how you're using them.
My suggestion would be to put your current architecture aside and start trying to build up a new one with ZMQ as its core, and see where you run into limitations that are solved by the other parts of your stack.
As for the "intrinsic design" of ZMQ, here's the basics that you need to understand as a starting point:
A ZMQ socket handles connection details for you, including managing network hiccups - but this has limits that you'll need to know
There are different kinds of ZMQ sockets, and they have opinions about how you use them. Some of them communicate asynchronously, some of them are strictly synchronous, some are one way, some are bi-directional.
If a connection between two sockets is severed (e.g. one node goes down, there is a network failure - something more than a momentary hiccup), it's your job to recognize that and re-establish that connection
There is no built in brokering or topology, you have to design and build that all yourself.
... ultimately, ZMQ provides a toolset for you to build a messaging framework, it does not provide a fully realized messaging framework out of the box. So, yes, it has the power to replace some of the other tools you're currently using, but you'll have to build it.
What is the right approach to use to configure OpenSplice DDS to support 100,000 or more nodes?
Can I use a hierarchical naming scheme for partition names, so "headquarters.city.location_guid_xxx" would prevent packets from leaving a location, and "company.city*" would allow samples to align across a city, and so on? Or would all the nodes know about all these partitions just in case they wanted to publish to them?
The durability services will choose a master when it comes up. If one durability service is running on a Raspberry Pi in a remote location running over a 3G link what is to prevent it from trying becoming the master for "headquarters" and crashing?
I am experimenting with durability settings in a remote node such that I use location_guid_xxx but for the "headquarters" cloud server I use a Headquarters
On the remote client I might to do this:
<Merge scope="Headquarters" type="Ignore"/>
<Merge scope="location_guid_xxx" type="Merge"/>
so a location won't be master for the universe, but can a durability service within a location still be master for that location?
If I have 100,000 locations does this mean I have to have all of them listed in the "Merge scope" in the ospl.xml file located at headquarters? I would think this alone might limit the size of the network I can handle.
I am assuming that this product will handle this sort of Internet of Things scenario. Has anyone else tried it?
Considering the scale of your system I think you should seriously consider the use of Vortex-Cloud (see these slides http://slidesha.re/1qMVPrq). Vortex Cloud will allow you to better scale your system as well as deal with NAT/Firewall. Beside that, you'll be able to use TCP/IP to communicate from your Raspberry Pi to the cloud instance thus avoiding any problem related to NATs/Firewalls.
Before getting to your durability question, there is something else I'd like to point out. If you try to build a flat system with 100K nodes you'll generate quite a bit of discovery information. Beside generating some traffic, this will be taking memory on your end applications. If you use Vortex-Cloud, instead, we play tricks to limit the discovery information. To give you an example, if you have a data-write matching 100K data reader, when using Vortex-Cloud the data-writer would only match on end-point and thus reducing the discovery information by 100K times!!!
Finally, concerning your durability question, you could configure some durability service as alignee only. In that case they will never become master.
HTH.
A+
I want to create a P2P application on the internet. What is the best or if none exist a good enough way to do auto-discovery of other nodes in a decentralized network?
Grothoff and GauthierDickey from the GNUnet project (an anonymous censorship-resistant file-sharing network) researched on the question of bootstrapping a p2p network without any central hostlist.
They found that for the Gnutella (Limewire) network a random ip search needed on average 2500 connection attempts to find a peer.
In the paper they proposed a method which reduced the required connection attempts to 817 for Gnutella and 51 for the E2DK network.
Achieved was this through creating a statistical profile of p2p users for every DNS organization, this small (around 100kb) discovery database has to be created in advance and shipped with the p2p client.
This is the holy grail of P2P. There isn't a magic solution really - there's no way a node can discover other nodes without a good known point to act as a reference (well, you can do so on a LAN by using broadcasting, but not on the internet). P2P filesharing tends to work by having known websites distributing 'start points' for discovery, and then further discovery (I would expect) can come from asking nodes what other nodes they know about.
A good place to start on research would be Distributed Hash Tables.
As for security, that topic will be in the literature somewhere, I should think - again I would recommend Wikipedia. Non-existent ones are trivially dealt with: if you can't contact an IP/port, don't keep it on your list, and if a node regularly provides non-existent pointers, consider de-prioritising it or removing it from your list entirely.
For evil nodes, it depends on your use case, but let's say you are doing file sharing. If you request a section of a file, check with several nodes what the file section's hash should be, and then request by hash. If the evil node gives you a chunk that has a different hash, then you can again de-prioritise or forget that node.
Distributed processing systems work a little differently: they tend to ask several unrelated nodes to perform the same work, and then they use a voting system (probably using hashing again) to determine whether evilness is at hand. If a node provides consistently bad results, the administrator is contacted or the IP is removed from the known nodes list.
ok, for two peers to find each other they both have to know a common, lets say, mediator to exchange IPs once. You can use anything for this kind of the first handshake whilst being able to WRITE and READ from that "channel". i.e: DNS (your well known domains), e-Mail, IRC, Twitter, Facebook, dropbox, etc.
My C++ turn-based game server (which uses database) does not stand against current average amount of clients (players), so I want to expand it to multiple (more then one) amount of computers and databases where all clients still will remain within single game world (servers will must communicate with each other and use multiple databases).
Is there some tutorials/books/common standards which explain how to do it in a best way?
The way you put the database into the picture might be misleading: clustering solutions exist for all of the mostly used RDBMS, so that if you need to support your DB activities with more than one DB node you will just have to check the documentation from your DB vendor.
More complex scenarios are there when it comes to synchronize your non-DB application state that needs to be shared among several servers. There are already a number of questions here that tackle the same problem, like here or here
You might also be interested into some messaging system, I heard good things about ZeroMQ
Hope this helps.