Boost logo

Geometry :

Subject: Re: [geometry] Storage allocator questions
From: Henry Roeland (henry_at_[hidden])
Date: 2014-12-01 15:29:40


Hi Adam,

> On 1 dec. 2014, at 17:20, Adam Wulkiewicz <adam.wulkiewicz_at_[hidden]> wrote:
>
> Hi Henry,
>
> Henry Roeland wrote:
>> It would be greet if I can focus on storage allocator and leave the rtree part to you or somebody else.
>>
>
> Ok, let's do it this way. FYI, I won't have time this week to work on this. But at least I'll answer on this email briefly.
>

No problem.

>> I hope I can still ask you guys questions concerning this storage allocator? I have some global Idea’s and probably need help both on design level and implementation.
>>
>>
>> Questions so far:
>>
>> 1. Is paging the only mechanism to get memory growth under control for an rtree? Its the only one I can think of, but maybe there are other ways to explore?
>>
>
> For the first version I think that our design should be as generic as possible:

One vote for generic! :-)

> - the Storage concept should be an abstraction of a persistent storage, just like an Allocator is an abstraction of "directly-accessible" memory.
> - the rtree would request something or notify the storage about some events but it wouldn't have to know anything about the internal structure, techniques, heuristics, etc. E.g. the rtree wouldn't know anything about paging.
OK. But what about the other way around? Should the Storage/Allocator not know how the container is build up? This in order to choose efficiently how to store its data?

> - the interfaces should probably be as compact as possible, e.g. as I said before, I think that in addition to the Allocator's interface we'd need 2 functions:
> * storage_traits<StorageOrAllocator>::get_pointer(ptr) - notify the Storage that data pointed by a pointer is needed
> * storage_traits<StorageOrAllocator>::release_pointers() - notify the Storage that pointers aren't used anymore
> - the storage probably shouldn't write the data directly to persistent part, instead it should keep the modified data cached and then write them all in one go. It could depend on some heuristic, the Storage could write changes when release_pointers() was called or it could define flush() method which could be called explicitly by the user. Still, the changes of persistent data should be done only if all pointers was released. At all times the storage could choose to keep the data cached. It could choose to unload all of the nodes, or only some less frequently used ones.
>
> Is it reasonable?
>

Yes it sounds reasonable. But are we not creating a generic storage allocator which also can be used on e.g. std::vector or other stl containers? I have no problem with this but I cannot imagine that we are the first…

Other questions/points that popup:
What about relative pointers e.g. node vs. storage/page?
What about memory management issues:
Fragmentation
Alloc ahead
Garbage collection (We are going to do it already using smart pointer(s) but on low level?)
malloc libs like jemalloc and tcmalloc can do this on low level…
Move pages/blocks: Pointers become invalid?
Object/Page/Node pooling???
Does STL have any guidelines restrictions on memory management?
Is rtree thread safe? Should our Storage/Allocator be thread-safe?

>>
>> 2. Is there any way an rtree can be read-only? Or set read-only after (bulk)insert?
>>
>
> It could be done either in run- or in compile-time.
> The most obvious way would be to throw a run-time exception in the get_pointer() after some flag (rtree created) was set.
> Another possibility could be to disable rtree mutable methods like insert() and remove() in compile-time if some specific parameters or Storage/Allocator was passed.
> Do you have some specific use-case in mind?
>
At the company I’m working for :-) They use readonly maps a lot. So in the future(hopefully)rtree with storage/allocator as cache.

>>
>> 3. One problem/challenge is how to get and keep region(box) coordinates “divided” over the page nodes. At least thats my understanding of an (r)tree. You need the region coordinates as fast lookup ID to know if a query should “dive” into the page or its sub pages or not.
>> Crazy idea: Maybe the allocator should manage an rtree of pages with references/pointers to the “real” tree nodes?
>>
>> Questions:
>> Is my understanding of data lookup right? So coordinates determine the location of the data inside the RTree?
>> If this is true does this then mean that all Pages (Loaded or unloaded) must be known including its region coordinates?
>> This in order to get/keep the tree balanced and know which page must be loaded (by coordinate query)?
>>
>
> Each node internal node (in a classic R-tree) is a container of (Box, Ptr) pairs. During any traversal (insert, query, etc.) the rtree probably should load the whole node to have access to all of the Boxes of children regions. AFAIR all traversing algorithms access all Boxes. Then it chooses which children should be traversed and it traverses into the next level using the corresponding pointers.
>
>>
>> Lets discuss the different actions an storage allocator can/must handle:
>> Load in Page data on request by RTree
>> “Load" can mean anything: From File, From DB, From MessageQueue,…
> So the load would potentially be done in the call to get_pointer(). AFAIU the Storage would have to check if the underlying data (identified by Ptr storing some ID or handle) isn't already loaded/cached. If it was it'd just return the Ptr, if not then it'd:
> - load a file,
> - (maybe somewhere cache the raw file),
> - create a node from loaded data and keep it,
> - return the pointer.
>
> So in my scenario (release_pointers()) it'd have to keep some container of returned pointers and when release_pointers(), flush() was called then just iterate over them and save changed nodes or create new pages, etc. It should probably also be possible to search the specific pointer in this container, so it should probably be map or unordered_map, or sorted vector, etc. Or do you have other ideas?
>
> E.g. when construct() was called the Storage probably shouldn't create a Page or other persistent data, it should be created later, e.g. on flush(). The same with destroy().
>
This means lots of housekeeping. Kind of transaction…
At this time I have no other idea’s.

> Btw, the Storage/Allocator would have to support rebinding and creation of various types, not only nodes. It'd have to know if a type is a node or some other type. So some types would be handled differently than other types. I'm not sure if this is a good or a bad thing. I guess it's possible even with std::allocator<> to specialize it for some specific type.
>
>> “Request" means query hits coordinate region that matches the Page coordinates
> This isn't needed, the rtree would know exactly which child node must be accessed becaused the Ptr is bound with the region in the internal node.
>> Bulk Load multiple pages at once???
>>
> Hmm, it could be done and potentially usable. Though it'd complicate the storage_traits (1 additional methog/overload). But do you think that we could gain anything from that? I mean, various pages would have to be loaded and cached anyway. This could be added later if needed.
>> Unload Page data on request by Storage Manager itself(?)
>> “Unload” can mean:
>> To /dev/null :-) Only useful when read-only rtree
>> To File: Memory mapped or otherwise
>> To DB: Save data to (spatial) database
>> To MessageQueue
>> ???
>> “Request” can mean:
>> Smart_pointer(s) say that nodes inside page are not used anymore
>> Smart_pointer(s) say that page is not used anymore
>> Amount of usage(hit count) is small
>> ???
>> Bulk Unload multiple pages at once???
>>
>
> So in my scenario the rtree would notify the Storage that all pointers are released and that the Storage may do whatever it like with them. So it could save changed nodes, unload some of them, etc.
>
> Hmm, the Storage also probably should know that a node was modified, or that pointers are released after mutable operation like insert() or remove(). So e.g. in the first version after an insert() the Storage could just save all nodes.
>
>> NOTE: The storage allocator must, as Adam already pointed out, deliver an interface to make above points possible. But not implement them in anyway.
>>
>
> So as mentioned above I was thinking about storage_traits which could be specialized for some StorageOrAllocator. Some member functions would be empty for Allocators but for Storage would have to be implemented. They would probably call some Storage's methods which could be called whatever the implementor liked.
>
>> Other questions/points that come to my mind are:
>> Must the storage allocator always store? What about transactions between in memory (RTree) and Persistence storage?
>> State-full allocator: Who owns the nodes and the data inside it? I always thought of an allocator like a factory that does not own the data…
>> When data is owned by another STL(like) container then IMHO a storage allocator has no use. This because the storage allocator then does not own the data and has no means to free/unload it.
>>
>
> I see 2 ways how the ownership could be implemented:
> 1. Similar to Boost.Interprocess where there is a memory Manager owning the data and AFAIU an Allocator only wrapping some reference to this Manager. So AFAIU (I didn't check the code) the Allocator is only an interface between a Container and a Manager. This simplifies the copying of an Allocator, rebinding, etc.
> 2. An Allocator which is in the same time a Manager or rather owns a Manager, so also owns the data. This'd probably require that the Allocator would have to store a shared_pointer to some Manager internally. An instance of a Manager would be created under the hood and a pointer to it would be copied automatically.
>
> 1 is more clear but it'd require to design not only the interface for a Storage/Allocator but also for a Manager, like in the case of Interprocess. 2 would only require to design the interface of an Allocator/Storage. If we choose 1 we could divide the logic into 2 Concepts, the Allocator could implement the in-memory-part (pointers, memory allocation, construction, destruction, etc.) and the Manager the persistent part (some file IDs or handles, saving and loading files, nodes serialization, etc.). So as in the case of Interprocess we could have 1 Allocator and could have many Managers.
>
> In general I propose to work iteratively, that is at the beginning support minimal set of required operations in the Storage/Allocator to be able to e.g. insert() values and perform a query() in the most simple, probably inefficient way, but with a simple and elegant interface. And then optimize operations and extend the interface. Or would you prefer to design the whole Storage/Allocator theoreticaly?
>
I prefer both :-) I will try to keep an UML overview with our suggestions to keep an theoretical overview. Next to that its probably already time to dive into code to see what is usable and what not.

Pratical question: Should I fork BoostGeometry on github or do you prefer a branch?

> Regards,
> Adam
> _______________________________________________
> Geometry mailing list
> Geometry_at_[hidden]
> http://lists.boost.org/mailman/listinfo.cgi/geometry

Thanks and Kind regards,

Henry



Geometry list run by mateusz at loskot.net