Comparison

Although you are probably familiar with many components on Meshy Space, their support by a single infrastructure does not exist yet.

Be aware that Meshy Space is on a higher level of abstraction than ftp or cloud: it will use various existing protocols to transport the data between collections. At the same time, it takes over some tasks from those lower level protocols.

Disclaimer: I do not know everything. Correct me when you think I am wrong.

Comparison to other data movers

For brevety, the term fact is used for a sequence of bytes with its related meta-data; like a file with its attributes.

	FTP-server	Object Storage	NFS, DB, or local disk	SOAP/ Micro- services	Meshy Space
Small facts	inefficient use zip/tar	inefficient	fast	inefficient	auto-packaged
Bulk facts	inefficient	inefficient	fast	inefficient	stream-lined auto-compression
Authentication space	single instance	single instance	local network	single instance	full network
Removal notifications	no	no	no	no	yes
Built-in redundancy	(mirrors)	yes	no	possible	yes
Local caching	no	no	yes	no	yes
Life-time management: release, deprecate, expire	no	no	no	no	yes
Democratic name- and rights rules	no	no	no	no	yes
Sub- and super-setting data-sets	no	no	no	no	yes
Data-structure evolution	no	no	no	no!	yes

Above table should make it very clear that Meshy Space is on a totally different level of data-management as commonly used data sharing infrastructure.

Disclaimer: of course, some of the "no"s are actually "possible in some way". However, those are usually client-side work-arounds, not server-side supported features.

Comparison to data managers

Meshy Space takes data management to new levels, without understanding of what the data is. Most applications do manage complex facts and have complex configuration. The more evolved an application gets, the more it will support some kind of permission and storage management.

Meshy Space introduces Collections with Units. Any unit is a fact: some bytes to be interpreted as file or data with related meta-data. The Collections are logical groups of facts, like a directory or object. These collections implement a super-set of all features which applications need.

Some examples of possible implementations:

a "wallet", which collects public and private user identities in separate units;
a collection with VirtualHosts, each as a separate unit, as part of the configuration for a web-service;
family holiday pictures, organized as collection of collections, one per event;
Linux software distribution, as collection of packages;
software module archives, like CPAN for the Perl programming language;
publications, like news, email blacklists, security advisories, documentation, or logs;
definitions of interfaces, each available operation as a separate unit;
Web-crawler results to be shared; ...

Which collections of data do you maintain in a directory, or distribute in a zip-file. Could it be managed better?

mark@overmeer.net Web-pages generated on 2023-12-19