| MJesus | mfkr ? |
|---|
| mfkr_ | any attempt to develop a completly free open platform needs to have also an open hardware platfowm |
|---|
| mfkr_ | however the cost involve on to development a hardware platform |
|---|
| mfkr_ | make very tricky to implement a conventional process |
|---|
| mfkr_ | that is completly free |
|---|
| mfkr_ | even if the architecture became available , |
|---|
| mfkr_ | only a few people will have the resources |
|---|
| mfkr_ | to made improvements to that platform |
|---|
| mfkr_ | There are solutions of core of free processor at OpenCores |
|---|
| mfkr_ | The use of VHDL allow as somthing |
|---|
| mfkr_ | something similra to a free hardware |
|---|
| mfkr_ | however the computing power of these cores is to low , |
|---|
| mfkr_ | so it doesn't allow us to build an acceptable desktop |
|---|
| mfkr_ | The Cluster, and the next step , |
|---|
| mfkr_ | the grid systems will allow us |
|---|
| mfkr_ | to obtain supercomputer resources |
|---|
| mfkr_ | with common machines. |
|---|
| mfkr_ | however only few people really needs supercomputers |
|---|
| mfkr_ | Even in that case the ideas are feasible |
|---|
| mfkr_ | in other scale |
|---|
| mfkr_ | making grid of very low performance processors we may obtain a normal computer |
|---|
| mfkr_ | with a very low cost |
|---|
| mfkr_ | at the same time allowing to the ens user to modify the hardware |
|---|
| mfkr_ | The Free Software , as a matter of fact, have a big problem with the innovation |
|---|
| mfkr_ | being this known even by the OS gurus- the told that no one really innovate on that field |
|---|
| mfkr_ | during the last years , the Open Software have been criticize for cloning the features of copyrighted software |
|---|
| mfkr_ | Even if this is false in most cases |
|---|
| mfkr_ | due to the fact that tcp-ip, dns, anf the email systems have been develop on the open software |
|---|
| mfkr_ | however many free software applications are clones of applications build on other operating systems |
|---|
| mfkr_ | Overall , the development of cientific open software, |
|---|
| mfkr_ | that years ago seems to be about to die |
|---|
| mfkr_ | we could say that today is completly dead |
|---|
| mfkr_ | Some one remember the article that I wrote on hispalinux |
|---|
| mfkr_ | many years ago, and was reposted ? |
|---|
| mfkr_ | Well.. the thing is even worst that in those days |
|---|
| mfkr_ | even if some projects are still developed under Open Software |
|---|
| mfkr_ | for cientific use |
|---|
| mfkr_ | nowadays if any scientific software is made , it is safely kept under three keys to avoid being used by the competitor |
|---|
| mfkr_ | Universitys are most intrested playing with the last functional languages |
|---|
| mfkr_ | or enjoying the advantages of Microsoft World |
|---|
| mfkr_ | We have to choices , to copy the existing Processor Architecture or to develop a completly new one |
|---|
| mfkr_ | the first past won't be creative , and will violate patents |
|---|
| mfkr_ | and won't be easy to develop by commo people |
|---|
| mfkr_ | the other chance is to start a new platform from zero, |
|---|
| mfkr_ | including the new ideas and trend on the systems world |
|---|
| mfkr_ | from the free SW point of view , we could digress about where the IT will be 10 or 15 years from now |
|---|
| mfkr_ | and to develop systems that be ready for that time |
|---|
| mfkr_ | This is the vision and design parmeter of Apila |
|---|
| mfkr_ | in this scenario , the advantages of free software are evident |
|---|
| mfkr_ | the software industry needs to profit in a short time frame |
|---|
| _libra_ | it is probaly that in the future, computer resources will be a comodity |
|---|
| _libra_ | and will have to load the data and with only one procesor unit of low weight |
|---|
| _libra_ | throught suscription, publicity, or just because we are in a public institution |
|---|
| _libra_ | (like a campus or a library) we can spend procesing time to execute the programs |
|---|
| _libra_ | that will let us change the data |
|---|
| _libra_ | (here is where comes the proces migration |
|---|
| _libra_ | and the idea that a launched task is a wearable apila |
|---|
| _libra_ | can be executed in the university grid apila |
|---|
| _libra_ | this is a basic idea of the wearables computers. One have the data |
|---|
| _libra_ | with a minium compute amount |
|---|
| _libra_ | In Apila, using the proces migration algorith, |
|---|
| _libra_ | the fact of carying a minium compute power isnt problemathic, since |
|---|
| _libra_ | in case that we need more power than we have for |
|---|
| _libra_ | some specific application, we can take it power arround us if they want us to take it |
|---|
| _libra_ | About P2P file system, there are some things to say about. |
|---|
| _libra_ | nowadays P2P file systems are a fact. they are used |
|---|
| _libra_ | to exchange all kind of files, of every size. |
|---|
| _libra_ | one of the more famouse implementations of those file systems areused to |
|---|
| _libra_ | exchange unlicensed data |
|---|
| _libra_ | like films |
|---|
| _libra_ | but, in P2P, there is more than that |
|---|
| _libra_ | and the P2P mpre interesting characteristhics havent been used |
|---|
| _libra_ | the more important is that if one of the owners of a file become down, or if this has a low speed conection, we wont have any kind of dificult getting a file |
|---|
| _libra_ | this is esential to favour the knowlegment difusion |
|---|
| _libra_ | other important characteristhic is the censure |
|---|
| _libra_ | it is almost imposible in a P2P network |
|---|
| _libra_ | i e, think how much eforts is RIAA spending agains the edonkey network |
|---|
| _libra_ | but, my ideal isnt edonkey |
|---|
| _libra_ | freenet is a interesting project designed to |
|---|
| _libra_ | free censure information, no corporation could see what we read or share |
|---|
| _libra_ | it is the future |
|---|
| _libra_ | this has beeen a short introduction of the concepts Apila is based on |
|---|
| _libra_ | now lets see the Apila architecture |
|---|
| _libra_ | Apila is a free architecture, that include soft and hardware |
|---|
| _libra_ | of distributed and multiple computation, that follous a grid model |
|---|
| _libra_ | with proces migration using the OpenMosix algorith |
|---|
| _libra_ | and that uses as native file system freenet |
|---|
| _libra_ | the Apila architecture let define a friends group |
|---|
| _libra_ | that share procesor and data, and compoun from small "empotrados" procesors |
|---|
| _libra_ | to personal computers joining lots Apila technology procesors |
|---|
| _libra_ | Apila programs arent compiled in the tradicional meaning |
|---|
| _libra_ | Apila machine code isnt a lineal code structure that can be know for a human |
|---|
| _libra_ | but an execution tree generated from the high level code |
|---|
| _libra_ | Stop making it, no one use assembly in the user area |
|---|
| _libra_ | for wich we can put in the code all the paralelism asociated |
|---|
| _libra_ | to a program that a compiler can calculate, and that is wasted when assembly is used |
|---|
| _libra_ | (ooooooooook, we´ll use it when we calculate the suboptium instruction interlaced |
|---|
| _libra_ | when we reaorder the instructions in compilation fase |
|---|
| _libra_ | but it is vile |
|---|
| _libra_ | and that later will made that the major part of the pipelines will be wasted |
|---|
| _libra_ | in our modern procesors) |
|---|
| _libra_ | Apila procesor can migrate throwght the load equallity algortihm |
|---|
| _libra_ | developed by Ammon Barak profesorand that use SSI cluster system like OpenSSI, |
|---|
| _libra_ | OpenMosix or MOSIX |
|---|
| _libra_ | The Apila procesor can execute Apila aplications in native mode |
|---|
| _libra_ | now the Apila tree proces in native mode, without having a intermediate interpretation layer |
|---|
| _libra_ | It can be in its simple way a FPGA compiled from a VHDL open core. this minium Apila |
|---|
| _libra_ | execute a operation tree following a first depth strategy; but hardware manufacturers |
|---|
| _libra_ | with monetary sesources can made complex Apila procesors |
|---|
| _libra_ | that can evaluate the strategy operations tree of cut of branchs |
|---|
| _libra_ | evaluating subbranchs in paralel, and with all the optimizations that can develop. |
|---|
| _libra_ | But the architecture is free, and also procesors are |
|---|
| _libra_ | it is a good way to introduce Intel and AMD in the free software world |
|---|
| _libra_ | there, althought a propietary company can made a quickly Apila procesor, an there will be |
|---|
| _libra_ | its busines, the platform will be free. Everybody can amde a Apila procesor without paying |
|---|
| _libra_ | royalities from 0 with FPGA´s, and if the big companys use Apila, they should contribute to the comunity, and all the history we know. |
|---|
| _libra_ | Operating system developers, can made Apila compatibility modules to it OS, that emulates Apila |
|---|
| _libra_ | throwght a virtual machine |
|---|
| _libra_ | Fantasy? |
|---|
| _libra_ | Sirs, that is what microsoft want made with .NET, but without FPGA´s, and without free cores |
|---|
| _libra_ | and lot of people follow them |
|---|
| _libra_ | .NEt in swatchs and moviles phones is not fruit of the M$ innocence |
|---|
| _libra_ | They wont have freenet, nor proces migration or the advantages of free software |
|---|
| _libra_ | Lets talk a while about the ApilaOS operating system |
|---|
| _libra_ | ApilaOS implements a minimal operating system, and the ApilaNet protocol |
|---|
| _libra_ | it can be considered as a nanokernel because we have a minium control over procesor |
|---|
| _libra_ | the protocol AnilaNEt and the IPv6 heap. |
|---|
| _libra_ | a design objective is that it can run with the compute power of a PIC, even if we have to use the scisors |
|---|
| _libra_ | if later something is needed, it could be put in the user area |
|---|
| _libra_ | dont worry, it isnt a ApilaOs objective to have a webserver in the kernel ;-) |
|---|
| _libra_ | ApilaOS has "empotrado" characteristhics |
|---|
| _libra_ | it has a two procesor planificator, with fixed priority. one of the two procesors can be |
|---|
| _libra_ | a loadable module that implement a multilevel orund-robin, but in the Apila machines of less |
|---|
| _libra_ | capacity it isnt possible |
|---|
| _libra_ | Is for that that Apila can serve two proces in real time, or be like a clasic Unix, depending on the procesor power and the needs |
|---|
| _libra_ | the OS has iimplemented the comunication protocol with other grid nodes, of functionally branches migration of a node and node proces migration. |
|---|
| _libra_ | the comunication protocol is critical to ensure the migation between platforms |
|---|
| _libra_ | nowadays im working on it |
|---|
| _libra_ | ApilaNET protocol is the Apila basic piece. |
|---|
| _libra_ | this protocol permits the migration of Apila compatible applications,independent of the |
|---|
| _libra_ | procesor or the architecture executing the protocol |
|---|
| _libra_ | Every Apila architecture compatible application can migrate to a node using the ApilaNET protocol |
|---|
| _libra_ | the ApilaNET protocol is the agent who distribute the information load, and send and return |
|---|
| _libra_ | mesages from migrated app´s |
|---|
| _libra_ | Lets see ApilaFS |
|---|
| _libra_ | we have seen in PDA´s that they are joining the memory and filesystem concepts to the user: |
|---|
| _libra_ | wherever the data is, user has a common acces mechanism to them |
|---|
| _libra_ | Apila platform will organish its store in three layers |
|---|
| _libra_ | the first will store in primary/secondary memory the Apila module |
|---|
| _libra_ | the diferent memory levels are invisible to the user |
|---|
| _libra_ | like happens with PDA. the user have not interest in what is in memory and what in disc |
|---|
| _libra_ | the usersç knows he has the data in the machine, and that he modifies it, they are modified |
|---|
| _libra_ | data is marked as private, what means that they can only be readed locally |
|---|
| _libra_ | or commons, in wich case the acces is made throwght a ACL |
|---|
| _libra_ | (listing wich users can ahve acces to data) |
|---|
| _libra_ | or publics, in that case, everybody have acces to them |
|---|
| _libra_ | and we put it at the disposal of all the P2P network |
|---|
| _libra_ | the second layer of memory/file system correspond to the common or public data |
|---|
| _libra_ | of the platforms Apila directly visible |
|---|
| _libra_ | s |
|---|
| _libra_ | if the user needs a data, and this isnt in the primary lvl, the system automatically requieres it to the second lvl |
|---|
| _libra_ | in third lvl we find the data that can be accesed throwght a direct transfer repositorium who they dont have direct acces |
|---|
| _libra_ | it is used if required data isnt in the 2º lvl |
|---|
| _libra_ | we canput here internet while it continues being free and public |
|---|
| _libra_ | and also the intranet and the systems pay per view |
|---|
| _libra_ | that can be accesed by our pc |
|---|
| _libra_ | the fourth lvl is a P2P network |
|---|
| _libra_ | with like freenet architecture, where we canrequire our public data or the public data of other people |
|---|
| _libra_ | free Apila memory/disc space can be used by ApilaOS to store parcial or totally P2P data system |
|---|
| _libra_ | like in freenet |
|---|
| _libra_ | The call to the system, if it is necessary to acces to this lvl, warn that the data is being searched |
|---|
| _libra_ | and the time it will spend in it |
|---|
| _libra_ | The Apila userinterface: |
|---|
| _libra_ | complicated to fanathic command console |
|---|
| _libra_ | Apila interface could be any interface |
|---|
| _libra_ | but it is necessary to used the new platform to propose a new system: |
|---|
| _libra_ | thre are someideas about its manage |
|---|
| _libra_ | but, there are only ideas |
|---|
| _libra_ | lets see some examples of Apila scenarios of use |
|---|
| _libra_ | Using Apila in large Stores |
|---|
| _libra_ | we get propaganda of a furniture store situed in the industrial poligon |
|---|
| _libra_ | they transmite us an offer |
|---|
| _libra_ | with the route to the store |
|---|
| _libra_ | to ur free souting system |
|---|
| _libra_ | with our free routing system, that accese to GPS satellites we arrive to the store |
|---|
| _libra_ | where we ask for the bed priece |
|---|
| dani_mt | other example: |
|---|
| dani_mt | we arrive to our campus. we carry the data we ae gonna use in the practice. |
|---|
| dani_mt | we sit downin the bar, and we decide to read the document, another time, and made it better. |
|---|
| dani_mt | we start oppening an Apila terminal int the table of the bar. |
|---|
| dani_mt | this Apila terminal has a grafic interface, with an Apila procesor that contact |
|---|
| dani_mt | with our wearable Apila to exchange data. |
|---|
| _libra_ | but we havent the necesary data |
|---|
| _libra_ | see at the public network, and the avaiblity of the data is slow |
|---|
| _libra_ | for wich we should acces to the data of the library |
|---|
| _libra_ | were data is in primary avaibility |
|---|
| _libra_ | because of that you can acces to the library from the bar |
|---|
| _libra_ | (somethings never change) |
|---|
| _libra_ | we shutdown the terminal and go to the library |
|---|
| dani_mt | but the data we are looking for isnt in the P"P memory, now it is in the direct acces memory |
|---|
| dani_mt | we can improve our work and give it to the techaer. |
|---|
| dani_mt | the last example: |
|---|
| dani_mt | there is a expanded reality new game,that needs more resources than we have |
|---|
| dani_mt | but our child want it. |
|---|
| dani_mt | for that, and having in mind that our child have good marks, |
|---|
| dani_mt | we give him a Apila procesor head to its console. |
|---|
| dani_mt | well, it isnt a bad idea, the procesor time dont used by his console -that will be a lot if he get bad marks :-) - |
|---|
| dani_mt | benefits all the family, |
|---|
| dani_mt | because it is integrated in the Apila home cluster |
|---|
| dani_mt | and will help tothe domotic house system |
|---|
| dani_mt | this scenarios could be real in a near future |
|---|
| dani_mt | most of the technologies we have talk about here are developed in free software |
|---|
| dani_mt | and the implementation of the process mean save cost |
|---|
| dani_mt | and for the moment, thats all guys |
|---|
