1 /*! @page platform Describing the virtual platform
4 @section pf_res Resource description
6 @subsection pf_res_computing Computing Resources
9 @subsubsection pf_tag_cluster <cluster>
11 ``<cluster />`` represents a machine-cluster. It is most commonly used
12 when one wants to define many hosts and a network quickly. Technically,
13 ``cluster`` is a meta-tag: <b>from the inner SimGrid point of
14 view, a cluster is a network zone where some optimized routing is defined</b>.
15 The default inner organization of the cluster is as follow:
21 ____________|__________|_____________ backbone
23 l0| l1| l2| l97| l96 | | l99
29 Here, a set of <b>host</b>s is defined. Each of them has a <b>link</b>
30 to a central backbone (backbone is a link itself, as a link can
31 be used to represent a switch, see the switch / link section
32 below for more details about it). A <b>router</b> allows to connect a
33 <b>cluster</b> to the outside world. Internally,
34 SimGrid treats a cluster as a network zone containing all hosts: the router is the default
35 gateway for the cluster.
37 There is an alternative organization, which is as follows:
51 The principle is the same, except that there is no backbone. This representation
52 can be obtained easily: just do not set the bb_* attributes.
55 Attribute name | Mandatory | Values | Description
56 --------------- | --------- | ------ | -----------
57 id | yes | string | The identifier of the cluster. Facilitates referring to this cluster.
58 prefix | yes | string | Each node of the cluster has to have a name. This name will be prefixed with this prefix.
59 suffix | yes | string | Each node of the cluster will be suffixed with this suffix
60 radical | yes | string | Regexp used to generate cluster nodes name. Syntax: "10-20" will give you 11 machines numbered from 10 to 20, "10-20;2" will give you 12 machines, one with the number 2, others numbered as before. The produced number is concatenated between prefix and suffix to form machine names.
61 speed | yes | int | Same as the ``speed`` attribute of the ``@<host@>`` tag.
62 core | no | int (default: 1) | Same as the ``core`` attribute of the ``@<host@>`` tag.
63 bw | yes | int | Bandwidth for the links between nodes and backbone (if any). See the @ref pf_tag_link "link section" for syntax/details.
64 lat | yes | int | Latency for the links between nodes and backbone (if any). See <b>link</b> section for syntax/details.
65 sharing_policy | no | string | Sharing policy for the links between nodes and backbone (if any). See <b>link</b> section for syntax/details.
66 bb_bw | no | int | Bandwidth for backbone (if any). See <b>link</b> section for syntax/details. If bb_bw and bb_lat (see below) attributes are omitted, no backbone is created (alternative cluster architecture <b>described before</b>).
67 bb_lat | no | int | Latency for backbone (if any). See <b>link</b> section for syntax/details. If bb_lat and bb_bw (see above) attributes are omitted, no backbone is created (alternative cluster architecture <b>described before</b>).
68 bb_sharing_policy | no | string | Sharing policy for the backbone (if any). See <b>link</b> section for syntax/details.
69 limiter_link | no | int | Bandwidth for limiter link (if any). This adds a specific link for each node, to set the maximum bandwidth reached when communicating in both directions at the same time. In theory this value should be 2*bw for splitduplex links, but in reality this might be less. This value will depend heavily on the communication model, and on the cluster's hardware, so no default value can be set, this has to be measured. More details can be obtained in <a href="https://hal.inria.fr/hal-00919507/"> "Toward Better Simulation of MPI Applications on Ethernet/TCP Networks"</a>
70 loopback_bw | no | int | Bandwidth for loopback (if any). See <b>link</b> section for syntax/details. If loopback_bw and loopback_lat (see below) attributes are omitted, no loopback link is created and all intra-node communication will use the main network link of the node. Loopback link is a @ref pf_sharing_policy_fatpipe "@b FATPIPE".
71 loopback_lat | no | int | Latency for loopback (if any). See <b>link</b> section for syntax/details. See loopback_bw for more info.
72 topology | no | FLAT@|TORUS@|FAT_TREE@|DRAGONFLY (default: FLAT) | Network topology to use. SimGrid currently supports FLAT (with or without backbone, as described before), <a href="http://en.wikipedia.org/wiki/Torus_interconnect">TORUS </a>, FAT_TREE, and DRAGONFLY attributes for this tag.
73 topo_parameters | no | string | Specific parameters to pass for the topology defined in the topology tag. For torus networks, comma-separated list of the number of nodes in each dimension of the torus. Please refer to the specific documentation for @ref simgrid::kernel::routing::FatTreeZone "FatTree NetZone", @ref simgrid::kernel::routing::DragonflyZone "Dragonfly NetZone".
76 the router name is defined as the resulting String in the following
80 router_name = prefix + clusterId + "_router" + suffix;
84 #### Cluster example ####
86 Consider the following two (and independent) uses of the ``cluster`` tag:
89 <cluster id="my_cluster_1" prefix="" suffix="" radical="0-262144"
90 speed="1e9" bw="125e6" lat="5E-5"/>
92 <cluster id="my_cluster_2" prefix="c-" suffix=".me" radical="0-99"
93 speed="1e9" bw="125e6" lat="5E-5"
94 bb_bw="2.25e9" bb_lat="5E-4"/>
97 The second example creates one router and 100 machines with the following names:
99 c-my_cluster_2_router.me
107 @subsubsection pf_cabinet <cabinet>
110 This tag is only available when the routing mode of the network zone
111 is set to ``Cluster``.
113 The ``<cabinet />`` tag is, like the @ref pf_tag_cluster "<cluster>" tag,
114 a meta-tag. This means that it is simply a shortcut for creating a set of (homogenous) hosts and links quickly;
115 unsurprisingly, this tag was introduced to setup cabinets in data centers quickly. Unlike
116 <cluster>, however, the <cabinet> assumes that you create the backbone
117 and routers yourself; see our examples below.
121 Attribute name | Mandatory | Values | Description
122 --------------- | --------- | ------ | -----------
123 id | yes | string | The identifier of the cabinet. Facilitates referring to this cluster.
124 prefix | yes | string | Each node of the cabinet has to have a name. This name will be prefixed with this prefix.
125 suffix | yes | string | Each node of the cabinet will be suffixed with this suffix
126 radical | yes | string | Regexp used to generate cabinet nodes name. Syntax: "10-20" will give you 11 machines numbered from 10 to 20, "10-20;2" will give you 12 machines, one with the number 2, others numbered as before. The produced number is concatenated between prefix and suffix to form machine names.
127 speed | yes | int | Same as the ``speed`` attribute of the @ref pf_tag_host "<host>" tag.
128 bw | yes | int | Bandwidth for the links between nodes and backbone (if any). See the @ref pf_tag_link "link section" for syntax/details.
129 lat | yes | int | Latency for the links between nodes and backbone (if any). See the @ref pf_tag_link "link section" for syntax/details.
132 Please note that as of now, it is impossible to change attributes such as,
133 amount of cores (always set to 1), the sharing policy of the links (always set to @ref pf_sharing_policy_splitduplex "SPLITDUPLEX").
137 The following example was taken from ``examples/platforms/meta_cluster.xml`` and
138 shows how to use the cabinet tag.
141 <zone id="my_cluster1" routing="Cluster">
142 <cabinet id="cabinet1" prefix="host-" suffix=".cluster1"
143 speed="1Gf" bw="125MBps" lat="100us" radical="1-10"/>
144 <cabinet id="cabinet2" prefix="host-" suffix=".cluster1"
145 speed="1Gf" bw="125MBps" lat="100us" radical="11-20"/>
146 <cabinet id="cabinet3" prefix="host-" suffix=".cluster1"
147 speed="1Gf" bw="125MBps" lat="100us" radical="21-30"/>
149 <backbone id="backbone1" bandwidth="2.25GBps" latency="500us"/>
154 Please note that you must specify the @ref pf_backbone "<backbone>"
155 tag by yourself; this is not done automatically and there are no checks
156 that ensure this backbone was defined.
158 The hosts generated in the above example are named host-1.cluster, host-2.cluster1
162 @subsection pf_ne Network equipments
164 There are two tags at all times available to represent network entities and
165 several other tags that are available only in certain contexts.
166 1. ``<link>``: Represents a entity that has a limited bandwidth, a
167 latency, and that can be shared according to TCP way to share this
170 The concept of links in SimGrid may not be intuitive, as links are not
171 limited to connecting (exactly) two entities; in fact, you can have more than
172 two equipments connected to it. (In graph theoretical terms: A link in
173 SimGrid is not an edge, but a hyperedge)
175 2. ``<router/>``: Represents an entity that a message can be routed
176 to, but that is unable to execute any code. In SimGrid, routers have also
177 no impact on the performance: Routers do not limit any bandwidth nor
178 do they increase latency. As a matter of fact, routers are (almost) ignored
179 by the simulator when the simulation has begun.
181 3. ``<backbone/>``: This tag is only available when the containing network zone is
182 used as a cluster (i.e., mode="Cluster")
185 If you want to represent an entity like a switch, you must use ``<link>`` (see section). Routers are used
186 to run some routing algorithm and determine routes (see Section @ref pf_routing for details).
188 @subsubsection pf_tag_link <link>
190 Network links can represent one-hop network connections. They are
191 characterized by their id and their bandwidth; links can (but may not) be subject
196 Attribute name | Mandatory | Values | Description
197 --------------- | --------- | ------ | -----------
198 id | yes | string | The identifier of the link to be used when referring to it.
199 bandwidth | yes | string | Maximum bandwidth for this link, along with its unit.
200 latency | no | double (default: 0.0) | Latency for this link.
201 sharing_policy | no | @ref sharing_policy_shared "SHARED"@|@ref pf_sharing_policy_fatpipe "FATPIPE"@|@ref pf_sharing_policy_splitduplex "SPLITDUPLEX" (default: SHARED) | Sharing policy for the link.
202 bandwidth_file | no | string | Allows you to use a file as input for bandwidth.
203 latency_file | no | string | Allows you to use a file as input for latency.
204 state_file | no | string | Allows you to use a file as input for states.
207 #### Possible shortcuts for ``latency`` ####
209 When using the latency attribute, you can specify the latency by using the scientific
210 notation or by using common abbreviations. For instance, the following three tags
214 <link id="LINK1" bandwidth="125000000" latency="5E-6"/>
215 <link id="LINK1" bandwidth="125000000" latency="5us"/>
216 <link id="LINK1" bandwidth="125000000" latency="0.000005"/>
219 Here, the second tag uses "us", meaning "microseconds". Other shortcuts are:
221 Name | Abbreviation | Time (in seconds)
222 ---- | ------------ | -----------------
223 Week | w | 7 * 24 * 60 * 60
224 Day | d | 24 * 60 * 60
228 Millisecond | ms | 0.001 = 10^(-3)
229 Microsecond | us | 0.000001 = 10^(-6)
230 Nanosecond | ns | 0.000000001 = 10^(-9)
231 Picosecond | ps | 0.000000000001 = 10^(-12)
233 #### Sharing policy ####
235 @anchor sharing_policy_shared
236 By default a network link is @b SHARED, i.e., if two or more data flows go
237 through a link, the bandwidth is shared fairly among all data flows. This
238 is similar to the sharing policy TCP uses.
240 @anchor pf_sharing_policy_fatpipe
241 On the other hand, if a link is defined as a @b FATPIPE,
242 each flow going through this link will be provided with the complete bandwidth,
243 i.e., no sharing occurs and the bandwidth is only limiting each flow individually.
244 Please note that this is really on a per-flow basis, not only on a per-host basis!
245 The complete bandwidth provided by this link in this mode
246 is ``number_of_flows*bandwidth``, with at most ``bandwidth`` being available per flow.
248 Using the FATPIPE mode allows to model backbones that won't affect performance
251 @anchor pf_sharing_policy_splitduplex
252 The last mode available is @b SPLITDUPLEX. This means that SimGrid will
253 automatically generate two links (one carrying the suffix _UP and the other the
254 suffix _DOWN) for each ``<link>`` tag. This models situations when the direction
255 of traffic is important.
258 Transfers from one side to the other will interact similarly as
259 TCP when ACK returning packets circulate on the other direction. More
260 discussion about it is available in the description of link_ctn description.
262 In other words: The SHARED policy defines a physical limit for the bandwidth.
263 The FATPIPE mode defines a limit for each application,
264 with no upper total limit.
267 Tip: By using the FATPIPE mode, you can model big backbones that
268 won't affect performance (except latency).
273 <link id="SWITCH" bandwidth="125000000" latency="5E-5" sharing_policy="FATPIPE" />
276 #### Expressing dynamism and failures ####
278 Similar to hosts, it is possible to declare links whose state, bandwidth
279 or latency changes over time (see Section @ref pf_host_dynamism for details).
281 In the case of network links, the ``bandwidth`` and ``latency`` attributes are
282 replaced by the ``bandwidth_file`` and ``latency_file`` attributes.
283 The following XML snippet demonstrates how to use this feature in the platform
284 file. The structure of the files "link1.bw" and "link1.lat" is shown below.
287 <link id="LINK1" state_file="link1.fail" bandwidth="80000000" latency=".0001" bandwidth_file="link1.bw" latency_file="link1.lat" />
291 Even if the syntax is the same, the semantic of bandwidth and latency
292 trace files differs from that of host availability files. For bandwidth and
293 latency, the corresponding files do not
294 express availability as a fraction of the available capacity but directly in
295 bytes per seconds for the bandwidth and in seconds for the latency. This is
296 because most tools allowing to capture traces on real platforms (such as NWS)
297 express their results this way.
299 ##### Example of "link1.bw" file #####
307 In this example, the bandwidth changes repeatedly, with all changes
308 being repeated every 12 seconds.
310 At the beginning of the the simulation, the link's bandwidth is 80,000,000
311 B/s (i.e., 80 Mb/s); this value was defined in the XML snippet above.
312 After four seconds, it drops to 40 Mb/s (line 2), and climbs
313 back to 60 Mb/s after another 4 seconds (line 3). The value does not change any
314 more until the end of the period, that is, after 12 seconds have been simulated).
315 At this point, periodicity kicks in and this behavior is repeated: Seconds
316 12-16 will experience 80 Mb/s, 16-20 40 Mb/s etc.).
318 ##### Example of "link1.lat" file #####
327 In this example, the latency varies with a period of 5 seconds.
328 In the xml snippet above, the latency is initialized to be 0.0001s (100µs). This
329 value will be kept during the first second, since the latency_file contains
330 changes to this value at second one, two and three.
331 At second one, the value will be 0.001, i.e., 1ms. One second later it will
332 be adjusted to 0.01 (or 10ms) and one second later it will be set again to 1ms. The
333 value will not change until second 5, when the periodicity defined in line 1
334 kicks in. It then loops back, starting at 100µs (the initial value) for one second.
336 #### The ``<prop/>`` tag ####
338 Similar to the ``<host>`` tag, a link may also contain the ``<prop/>`` tag; see the host
339 documentation (Section @ref pf_tag_host) for an example.
342 @subsubsection pf_backbone <backbone/>
345 This tag is <b>only available</b> when the containing network zone uses the "Cluster" routing mode!
347 Using this tag, you can designate an already existing link to be a backbone.
349 Attribute name | Mandatory | Values | Description
350 --------------- | --------- | ------ | -----------
351 id | yes | string | Name of the link that is supposed to act as a backbone.
353 @subsection pf_storage Storage
356 This is a prototype version that should evolve quickly, hence this
357 is just some doc valuable only at the time of writing.
358 This section describes the storage management under SimGrid ; nowadays
359 it's only usable with MSG. It relies basically on linux-like concepts.
360 You also may want to have a look to its corresponding section in
361 @ref msg_file ; access functions are organized as a POSIX-like
364 @subsubsection pf_sto_conc Storage - Main Concepts
366 The storage facilities implemented in SimGrid help to model (and account for)
367 storage devices, such as tapes, hard-drives, CD or DVD devices etc.
368 A typical situation is depicted in the figure below:
370 @image html ./webcruft/storage_sample_scenario.png
371 @image latex ./webcruft/storage_sample_scenario.png "storage_sample_scenario" width=@textwidth
373 In this figure, two hosts called Bob and Alice are interconnected via a network
374 and each host is physically attached to a disk; it is not only possible for each host to
375 mount the disk they are attached to directly, but they can also mount disks
376 that are in a remote location. In this example, Bob mounts Alice's disk remotely
377 and accesses the storage via the network.
379 SimGrid provides 3 different entities that can be used to model setups
380 that include storage facilities:
382 Entity name | Description
383 --------------- | -----------
384 @ref pf_storage_entity_storage_type "storage_type" | Defines a template for a particular kind of storage (such as a hard-drive) and specifies important features of the storage, such as capacity, performance (read/write), contents, ... Different models of hard-drives use different storage_types (because the difference between an SSD and an HDD does matter), as they differ in some specifications (e.g., different sizes or read/write performance).
385 @ref pf_tag_storage "storage" | Defines an actual instance of a storage type (disk, RAM, ...); uses a ``storage_type`` template (see line above) so that you don't need to re-specify the same details over and over again.
386 @ref pf_tag_mount "mount" | Must be wrapped by a @ref pf_tag_host tag; declares which storage(s) this host has mounted and where (i.e., the mountpoint).
389 @anchor pf_storage_content_file
390 ### %Storage Content File ###
392 In order to assess exactly how much time is spent reading from the storage,
393 SimGrid needs to know what is stored on the storage device (identified by distinct (file-)name, like in a file system)
394 and what size this content has.
397 The content file is never changed by the simulation; it is parsed once
398 per simulation and kept in memory afterwards. When the content of the
399 storage changes, only the internal SimGrid data structures change.
401 @anchor pf_storage_content_file_structure
402 #### Structure of a %Storage Content File ####
404 Here is an excerpt from two storage content file; if you want to see the whole file, check
405 the file ``examples/platforms/content/storage_content.txt`` that comes with the
408 SimGrid essentially supports two different formats: UNIX-style filepaths should
409 follow the well known format:
412 /lib/libsimgrid.so.3.6.2 12710497
416 /bin/simgrid_update_xml 5018
417 /bin/graphicator 66986
418 /bin/simgrid-colorizer 2993
423 Windows filepaths, unsurprisingly, use the windows style:
426 @Windows@avastSS.scr 41664
427 @Windows@bfsvc.exe 75264
428 @Windows@bootstat.dat 67584
429 @Windows@CoreSingleLanguage.xml 31497
431 @Windows@dchcfg64.exe 335464
432 @Windows@dcmdev64.exe 93288
436 The different file formats come at a cost; in version 3.12 (and most likely
437 in later versions, too), copying files from windows-style storages to unix-style
438 storages (and vice versa) is not supported.
440 @anchor pf_storage_content_file_create
441 #### Generate a %Storage Content File ####
443 If you want to generate a storage content file based on your own filesystem (or at least a filesystem you have access to),
444 try running this command (works only on unix systems):
447 find . -type f -exec ls -1s --block=1 {} @; 2>/dev/null | awk '{ print $2 " " $1}' > ./content.txt
450 @subsubsection pf_storage_entities The Storage Entities
452 These are the entities that you can use in your platform files to include
453 storage in your model. See also the list of our @ref pf_storage_example_files "example files";
454 these might also help you to get started.
456 @anchor pf_storage_entity_storage_type
457 #### @<storage_type@> ####
459 Attribute name | Mandatory | Values | Description
460 --------------- | --------- | ------ | -----------
461 id | yes | string | Identifier of this storage_type; used when referring to it
462 model | no | string | In the future, this will allow to change the performance model to use
463 size | yes | string | Specifies the amount of available storage space; you can specify storage like "500GiB" or "500GB" if you want. (TODO add a link to all the available abbreviations)
464 content | yes | string | Path to a @ref pf_storage_content_file "Storage Content File" on your system. This file must exist.
466 This tag must contain some predefined model properties, specified via the <model_prop> tag. Here is a list,
467 see below for an example:
469 Property id | Mandatory | Values | Description
470 --------------- | --------- | ------ | -----------
471 Bwrite | yes | string | Bandwidth for write access; in B/s (but you can also specify e.g. "30MBps")
472 Bread | yes | string | Bandwidth for read access; in B/s (but you can also specify e.g. "30MBps")
475 A storage_type can also contain the <b><prop></b> tag. The <prop> tag allows you
476 to associate additional information to this <storage_type> and follows the
477 attribute/value schema; see the example below. You may want to use it to give information to
478 the tool you use for rendering your simulation, for example.
480 Here is a complete example for the ``storage_type`` tag:
482 <storage_type id="single_HDD" size="4000">
483 <model_prop id="Bwrite" value="30MBps" />
484 <model_prop id="Bread" value="100MBps" />
485 <prop id="Brand" value="Western Digital" />
489 @subsubsection pf_tag_storage <storage>
491 Attributes | Mandatory | Values | Description
492 -------------- | --------- | ------ | -----------
493 id | yes | string | Identifier of this ``storage``; used when referring to it
494 typeId | yes | string | Here you need to refer to an already existing @ref pf_storage_entity_storage_type "@<storage_type@>"; the storage entity defined by this tag will then inherit the properties defined there.
495 attach | yes | string | Name of a host (see Section @ref pf_tag_host) to which this storage is <i>physically</i> attached to (e.g., a hard drive in a computer)
496 content | no | string | When specified, overwrites the content attribute of @ref pf_storage_entity_storage_type "@<storage_type@>"
498 Here are two examples:
501 <storage id="Disk1" typeId="single_HDD" attach="bob" />
503 <storage id="Disk2" typeId="single_SSD"
504 content="content/win_storage_content.txt" />
507 The first example is straightforward: A disk is defined and called "Disk1"; it is
508 of type "single_HDD" (shown as an example of @ref pf_storage_entity_storage_type "@<storage_type@>" above) and attached
509 to a host called "bob" (the definition of this host is omitted here).
511 The second storage is called "Disk2", is still of the same type as Disk1 but
512 now specifies a new content file (so the contents will be different from Disk1)
513 and the filesystem uses the windows style; finally, it is attached to a second host,
514 called alice (which is again not defined here).
516 @subsubsection pf_tag_mount <mount>
518 | Attribute | Mandatory | Values | Description |
519 | ----------- | ----------- | -------- | ------------- |
520 | id | yes | string | Refers to a @ref pf_tag_storage "<storage>" entity that will be mounted on that computer |
521 | name | yes | string | Path/location to/of the logical reference (mount point) of this disk
523 This tag must be enclosed by a @ref pf_tag_host tag. It then specifies where the mountpoint of a given storage device (defined by the ``id`` attribute)
524 is; this location is specified by the ``name`` attribute.
526 Here is a simple example, taken from the file ``examples/platform/storage.xml``:
529 <storage_type id="single_SSD" size="500GiB">
530 <model_prop id="Bwrite" value="60MBps" />
531 <model_prop id="Bread" value="200MBps" />
534 <storage id="Disk2" typeId="single_SSD"
535 content="content/win_storage_content.txt"
537 <storage id="Disk4" typeId="single_SSD"
538 content="content/small_content.txt"
541 <host id="alice" speed="1Gf">
542 <mount storageId="Disk2" name="c:"/>
545 <host id="denise" speed="1Gf">
546 <mount storageId="Disk2" name="c:"/>
547 <mount storageId="Disk4" name="/home"/>
551 This example is quite interesting, as the same device, called "Disk2", is mounted by
552 two hosts at the same time! Note, however, that the host called ``alice`` is actually
553 attached to this storage, as can be seen in the @ref pf_tag_storage "<storage>"
554 tag. This means that ``denise`` must access this storage through the network, but SimGrid automatically takes
555 care of that for you.
557 Furthermore, this example shows that ``denise`` has mounted two storages with different
558 filesystem types (unix and windows). In general, a host can mount as many storage devices as
562 Again, the difference between ``attach`` and ``mount`` is simply that
563 an attached storage is always physically inside (or connected to) that machine;
564 for instance, a USB stick is attached to one and only one machine (where it's plugged-in)
565 but it can only be mounted on others, as mounted storage can also be a remote location.
567 ###### Example files #####
569 @verbinclude example_filelist_xmltag_mount
571 @subsubsection pf_storage_example_files Example files
573 Several examples were already discussed above; if you're interested in full examples,
574 check the the following platforms:
576 1. ``examples/platforms/storage.xml``
577 2. ``examples/platforms/remote_io.xml``
579 If you're looking for some examplary C code, you may find the source code
580 available in the directory ``examples/msg/io/`` useful.
582 @subsubsection pf_storage_examples_modelling Modelling different situations
584 The storage functionality of SimGrid is type-agnostic, that is, the implementation
585 does not presume any type of storage, such as HDDs/SSDs, RAM,
586 CD/DVD devices, USB sticks etc.
588 This allows the user to apply the simulator for a wide variety of scenarios; one
589 common scenario would be the access of remote RAM.
591 #### Modelling the access of remote RAM ####
593 How can this be achieved in SimGrid? Let's assume we have a setup where three hosts
594 (HostA, HostB, HostC) need to access remote RAM:
604 An easy way to model this scenario is to setup and define the RAM via the
605 @ref pf_tag_storage "storage" and @ref pf_storage_entity_storage_type "storage type"
606 entities and attach it to a remote dummy host; then, every host can have their own links
607 to this host (modelling for instance certain scenarios, such as PCIe ...)
612 RAM - Dummy -- Host B
617 Now, if read from this storage, the host that mounts this storage
618 communicates to the dummy host which reads from RAM and
619 sends the information back.
622 @section pf_routing Routing
624 To achieve high performance, the routing tables used within SimGrid are
625 static. This means that routing between two nodes is calculated once
626 and will not change during execution. The SimGrid team chose to use this
627 approach as it is rare to have a real deficiency of a resource;
628 most of the time, a communication fails because the links experience too much
629 congestion and hence, your connection stops before the timeout or
630 because the computer designated to be the destination of that message
633 We also chose to use shortest paths algorithms in order to emulate
634 routing. Doing so is consistent with the reality: [RIP](https://en.wikipedia.org/wiki/Routing_Information_Protocol),
635 [OSPF](https://en.wikipedia.org/wiki/Open_Shortest_Path_First), [BGP](https://en.wikipedia.org/wiki/Border_Gateway_Protocol)
636 are all calculating shortest paths. They do require some time to converge, but
637 eventually, when the routing tables have stabilized, your packets will follow
640 @subsection pf_tag_zone <zone>
642 @subsection pf_rm Routing models
644 For each network zone, you must define explicitly which routing model will
645 be used. There are 3 different categories for routing models:
647 1. @ref pf_routing_model_shortest_path "Shortest-path" based models: SimGrid calculates shortest
648 paths and manages them. Behaves more or less like most real life
650 2. @ref pf_routing_model_manual "Manually-entered" route models: you have to define all routes
651 manually in the platform description file; this can become
652 tedious very quickly, as it is very verbose.
653 Consistent with some manually managed real life routing.
654 3. @ref pf_routing_model_simple "Simple/fast models": those models offer fast, low memory routing
655 algorithms. You should consider to use this type of model if
656 you can make some assumptions about your network zone.
657 Routing in this case is more or less ignored.
659 @subsubsection pf_raf The router affair
661 Using routers becomes mandatory when using shortest-path based
662 models or when using the bindings to the ns-3 packet-level
663 simulator instead of the native analytical network model implemented
666 For graph-based shortest path algorithms, routers are mandatory, because these
667 algorithms require a graph as input and so we need to have source and
668 destination for each edge.
670 Routers are naturally an important concept ns-3 since the
671 way routers run the packet routing algorithms is actually simulated.
672 SimGrid's analytical models however simply aggregate the routing time
673 with the transfer time.
675 So why did we incorporate routers in SimGrid? Rebuilding a graph representation
676 only from the route information turns out to be a very difficult task, because
677 of the missing information about how routes intersect. That is why we
678 introduced routers, which are simply used to express these intersection points.
679 It is important to understand that routers are only used to provide topological
682 To express this topological information, a <b>route</b> has to be
683 defined in order to declare which link is connected to a router.
686 @subsubsection pf_routing_model_shortest_path Shortest-path based models
688 The following table shows all the models that compute routes using
689 shortest-paths algorithms are currently available in SimGrid. More detail on how
690 to choose the best routing model is given in the Section called @"@ref pf_routing_howto_choose_wisely@".
692 | Name | Description |
693 | --------------------------------------------------- | -------------------------------------------------------------------------- |
694 | @ref pf_routing_model_floyd "Floyd" | Floyd routing data. Pre-calculates all routes once |
695 | @ref pf_routing_model_dijkstra "Dijkstra" | Dijkstra routing data. Calculates routes only when needed |
696 | @ref pf_routing_model_dijkstracache "DijkstraCache" | Dijkstra routing data. Handles some cache for already calculated routes. |
698 All those shortest-path models are instanciated in the same way and are
699 completely interchangeable. Here are some examples:
701 @anchor pf_routing_model_floyd
706 <zone id="zone0" routing="Floyd">
708 <cluster id="my_cluster_1" prefix="c-" suffix=""
709 radical="0-1" speed="1000000000" bw="125000000" lat="5E-5"
710 router_id="router1"/>
712 <zone id="zone1" routing="None">
713 <host id="host1" speed="1000000000"/>
716 <link id="link1" bandwidth="100000" latency="0.01"/>
718 <zoneroute src="my_cluster_1" dst="zone1"
721 <link_ctn id="link1"/>
727 zoneroute given at the end gives a topological information: link1 is
728 between router1 and host1.
730 #### Example platform files ####
732 This is an automatically generated list of example files that use the Floyd
733 routing model (the path is given relative to SimGrid's source directory)
735 @verbinclude example_filelist_routing_floyd
737 @anchor pf_routing_model_dijkstra
740 #### Example platform files ####
742 This is an automatically generated list of example files that use the Dijkstra
743 routing model (the path is given relative to SimGrid's source directory)
745 @verbinclude example_filelist_routing_dijkstra
749 <zone id="zone_2" routing="Dijkstra">
750 <host id="zone_2_host1" speed="1000000000"/>
751 <host id="zone_2_host2" speed="1000000000"/>
752 <host id="zone_2_host3" speed="1000000000"/>
753 <link id="zone_2_link1" bandwidth="1250000000" latency="5E-4"/>
754 <link id="zone_2_link2" bandwidth="1250000000" latency="5E-4"/>
755 <link id="zone_2_link3" bandwidth="1250000000" latency="5E-4"/>
756 <link id="zone_2_link4" bandwidth="1250000000" latency="5E-4"/>
757 <router id="central_router"/>
758 <router id="zone_2_gateway"/>
759 <!-- routes providing topological information -->
760 <route src="central_router" dst="zone_2_host1"><link_ctn id="zone_2_link1"/></route>
761 <route src="central_router" dst="zone_2_host2"><link_ctn id="zone_2_link2"/></route>
762 <route src="central_router" dst="zone_2_host3"><link_ctn id="zone_2_link3"/></route>
763 <route src="central_router" dst="zone_2_gateway"><link_ctn id="zone_2_link4"/></route>
767 @anchor pf_routing_model_dijkstracache
768 ### DijkstraCache ###
770 DijkstraCache example:
772 <zone id="zone_2" routing="DijkstraCache">
773 <host id="zone_2_host1" speed="1000000000"/>
775 (platform unchanged compared to upper example)
778 #### Example platform files ####
780 This is an automatically generated list of example files that use the DijkstraCache
781 routing model (the path is given relative to SimGrid's source directory):
783 Editor's note: At the time of writing, no platform file used this routing model - so
784 if there are no example files listed here, this is likely to be correct.
786 @verbinclude example_filelist_routing_dijkstra_cache
788 @subsubsection pf_routing_model_manual Manually-entered route models
790 | Name | Description |
791 | ---------------------------------- | ------------------------------------------------------------------------------ |
792 | @ref pf_routing_model_full "Full" | You have to enter all necessary routers manually; that is, every single route. This may consume a lot of memory when the XML is parsed and might be tedious to write; i.e., this is only recommended (if at all) for small platforms. |
794 @anchor pf_routing_model_full
799 <zone id="zone0" routing="Full">
800 <host id="host1" speed="1000000000"/>
801 <host id="host2" speed="1000000000"/>
802 <link id="link1" bandwidth="125000000" latency="0.000100"/>
803 <route src="host1" dst="host2"><link_ctn id="link1"/></route>
807 #### Example platform files ####
809 This is an automatically generated list of example files that use the Full
810 routing model (the path is given relative to SimGrid's source directory):
812 @verbinclude example_filelist_routing_full
814 @subsubsection pf_routing_model_simple Simple/fast models
816 | Name | Description |
817 | ---------------------------------------- | ------------------------------------------------------------------------------ |
818 | @ref pf_routing_model_cluster "Cluster" | This is specific to the @ref pf_tag_cluster "<cluster/>" tag and should not be used by the user, as several assumptions are made. |
819 | @ref pf_routing_model_none "None" | No routing at all. Unless you know what you're doing, avoid using this mode in combination with a non-constant network model. |
820 | @ref pf_routing_model_vivaldi "Vivaldi" | Perfect when you want to use coordinates. Also see the corresponding @ref pf_P2P_tags "P2P section" below. |
822 @anchor pf_routing_model_cluster
826 In this mode, the @ref pf_cabinet "<cabinet/>" tag is available.
828 #### Example platform files ####
830 This is an automatically generated list of example files that use the Cluster
831 routing model (the path is given relative to SimGrid's source directory):
833 @verbinclude example_filelist_routing_cluster
835 @anchor pf_routing_model_none
839 This model does exactly what it's name advertises: Nothing. There is no routing
840 available within this model and if you try to communicate within the zone that
841 uses this model, SimGrid will fail unless you have explicitly activated the
842 @ref options_model_select_network_constant "Constant Network Model" (this model charges
843 the same for every single communication). It should
844 be noted, however, that you can still attach an @ref pf_tag_zoneroute "ZoneRoute",
845 as is demonstrated in the example below:
847 @verbinclude platforms/cluster_and_one_host.xml
849 #### Example platform files ####
851 This is an automatically generated list of example files that use the None
852 routing model (the path is given relative to SimGrid's source directory):
854 @verbinclude example_filelist_routing_none
857 @anchor pf_routing_model_vivaldi
860 For more information on how to use the [Vivaldi Coordinates](https://en.wikipedia.org/wiki/Vivaldi_coordinates),
861 see also Section @ref pf_P2P_tags "P2P tags".
863 Note that it is possible to combine the Vivaldi routing model with other routing models;
864 an example can be found in the file @c examples/platforms/cloud.xml. This
865 examples models a NetZone using Vivaldi that contains other NetZones that use different
868 #### Example platform files ####
870 This is an automatically generated list of example files that use the None
871 routing model (the path is given relative to SimGrid's source directory):
873 @verbinclude example_filelist_routing_vivaldi
876 @subsection ps_dec Defining routes
878 There are currently four different ways to define routes:
880 | Name | Description |
881 | ------------------------------------------------- | ----------------------------------------------------------------------------------- |
882 | @ref pf_tag_route "route" | Used to define route between host/router |
883 | @ref pf_tag_zoneroute "zoneRoute" | Used to define route between different zones |
884 | @ref pf_tag_bypassroute "bypassRoute" | Used to supersede normal routes as calculated by the network model between host/router; e.g., can be used to use a route that is not the shortest path for any of the shortest-path routing models. |
885 | @ref pf_tag_bypassasroute "bypassZoneRoute" | Used in the same way as bypassRoute, but for zones |
887 Basically all those tags will contain an (ordered) list of references
888 to link that compose the route you want to define.
890 Consider the example below:
893 <route src="Alice" dst="Bob">
894 <link_ctn id="link1"/>
895 <link_ctn id="link2"/>
896 <link_ctn id="link3"/>
900 The route here from host Alice to Bob will be first link1, then link2,
901 and finally link3. What about the reverse route? @ref pf_tag_route "Route" and
902 @ref pf_tag_zoneroute "zoneroute" have an optional attribute @c symmetrical, that can
903 be either @c YES or @c NO. @c YES means that the reverse route is the same
904 route in the inverse order, and is set to @c YES by default. Note that
905 this is not the case for bypass*Route, as it is more probable that you
906 want to bypass only one default route.
908 For an @ref pf_tag_zoneroute "zoneroute", things are just slightly more complicated, as you have
909 to give the id of the gateway which is inside the zone you want to access ...
910 So it looks like this:
913 <zoneroute src="zone1" dst="zone2"
914 gw_src="router1" gw_dst="router2">
915 <link_ctn id="link1"/>
919 gw == gateway, so when any message are trying to go from zone1 to zone2,
920 it means that it must pass through router1 to get out of the zone, then
921 pass through link1, and get into zone2 by being received by router2.
922 router1 must belong to zone1 and router2 must belong to zone2.
924 @subsubsection pf_tag_linkctn <link_ctn>
926 This entity has only one purpose: Refer to an already existing
927 @ref pf_tag_link "<link/>" when defining a route, i.e., it
928 can only occur as a child of @ref pf_tag_route "<route/>"
930 | Attribute name | Mandatory | Values | Description |
931 | --------------- | --------- | ------ | ----------- |
932 | id | yes | String | The identifier of the link that should be added to the route. |
933 | direction | maybe | UP@|DOWN | If the link referenced by @c id has been declared as @ref pf_sharing_policy_splitduplex "SPLITDUPLEX", this indicates which direction the route traverses through this link: UP or DOWN. If you don't use SPLITDUPLEX, do not use this attribute or SimGrid will not find the right link.
935 #### Example Files ####
937 This is an automatically generated list of example files that use the @c <link_ctn/>
938 entity (the path is given relative to SimGrid's source directory):
940 @verbinclude example_filelist_xmltag_linkctn
942 @subsubsection pf_tag_zoneroute <zoneRoute>
944 The purpose of this entity is to define a route between two
945 NetZones. Recall that all zones form a tree, so to connect two
946 sibiling zones, you must give such a zoneRoute specifying the source
947 and destination zones, along with the gateway in each zone (ie, the
948 point to reach within that zone to reach the netzone), and the list of
949 links in the ancestor zone to go from one zone to another.
951 So, to go from an host @c src_host that is within zone @c src, to an
952 host @c dst_host that is within @c dst, you need to:
954 - move within zone @c src, from @c src_host to the specified @c gw_src;
955 - traverse all links specified by the zoneRoute (they are supposed to be within the common ancestor);
956 - move within zone @c dst, from @c gw_dst to @c dst_host.
960 | Attribute name | Mandatory | Values | Description |
961 | --------------- | --------- | ------ | ----------- |
962 | src | yes | String | The identifier of the source zone |
963 | dst | yes | String | See the @c src attribute |
964 | gw_src | yes | String | The gateway that will be used within the src zone; this can be any @ref pf_tag_host "Host" or @ref pf_router "Router" defined within the src zone. |
965 | gw_dst | yes | String | Same as @c gw_src, but with the dst zone instead. |
966 | symmetrical | no | YES@|NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
971 <zone id="zone0" routing="Full">
972 <cluster id="my_cluster_1" prefix="c-" suffix=".me"
973 radical="0-149" speed="1000000000" bw="125000000" lat="5E-5"
974 bb_bw="2250000000" bb_lat="5E-4"/>
976 <cluster id="my_cluster_2" prefix="c-" suffix=".me"
977 radical="150-299" speed="1000000000" bw="125000000" lat="5E-5"
978 bb_bw="2250000000" bb_lat="5E-4"/>
980 <link id="backbone" bandwidth="1250000000" latency="5E-4"/>
982 <zoneroute src="my_cluster_1" dst="my_cluster_2"
983 gw_src="c-my_cluster_1_router.me"
984 gw_dst="c-my_cluster_2_router.me">
985 <link_ctn id="backbone"/>
987 <zoneroute src="my_cluster_2" dst="my_cluster_1"
988 gw_src="c-my_cluster_2_router.me"
989 gw_dst="c-my_cluster_1_router.me">
990 <link_ctn id="backbone"/>
995 @subsubsection pf_tag_route <route>
997 The principle is the same as for
998 @ref pf_tag_zoneroute "ZoneRoute": The route contains a list of links that
999 provide a path from @c src to @c dst. Here, @c src and @c dst can both be either a
1000 @ref pf_tag_host "host" or @ref pf_router "router". This is mostly useful for the
1001 @ref pf_routing_model_full "Full routing model" as well as for the
1002 @ref pf_routing_model_shortest_path "shortest-paths" based models (as they require
1003 topological information).
1006 | Attribute name | Mandatory | Values | Description |
1007 | --------------- | --------- | ---------------------- | ----------- |
1008 | src | yes | String | The value given to the source's "id" attribute |
1009 | dst | yes | String | The value given to the destination's "id" attribute. |
1010 | symmetrical | no | YES@| NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1015 A route in the @ref pf_routing_model_full "Full routing model" could look like this:
1017 <route src="Tremblay" dst="Bourassa">
1018 <link_ctn id="4"/><link_ctn id="3"/><link_ctn id="2"/><link_ctn id="0"/><link_ctn id="1"/><link_ctn id="6"/><link_ctn id="7"/>
1022 A route in the @ref pf_routing_model_shortest_path "Shortest-Path routing model" could look like this:
1024 <route src="Tremblay" dst="Bourassa">
1029 You must only have one link in your routes when you're using them to provide
1030 topological information, as the routes here are simply the edges of the
1031 (network-)graph and the employed algorithms need to know which edge connects
1032 which pair of entities.
1034 @subsubsection pf_tag_bypassasroute bypasszoneroute
1036 As said before, once you choose
1037 a model, it (most likely; the constant network model, for example, doesn't) calculates routes for you. But maybe you want to
1038 define some of your routes, which will be specific. You may also want
1039 to bypass some routes defined in lower level zone at an upper stage:
1040 <b>bypasszoneroute</b> is the tag you're looking for. It allows to
1041 bypass routes defined between already defined between zone (if you want
1042 to bypass route for a specific host, you should just use byPassRoute).
1043 The principle is the same as zoneroute: <b>bypasszoneroute</b> contains
1044 list of links that are in the path between src and dst.
1046 #### Attributes ####
1048 | Attribute name | Mandatory | Values | Description |
1049 | --------------- | --------- | ---------------------- | ----------- |
1050 | src | yes | String | The value given to the source zone's "id" attribute |
1051 | dst | yes | String | The value given to the destination zone's "id" attribute. |
1052 | gw_src | yes | String | The value given to the source gateway's "id" attribute; this can be any host or router within the src zone |
1053 | gw_dst | yes | String | The value given to the destination gateway's "id" attribute; this can be any host or router within the dst zone|
1054 | symmetrical | no | YES@| NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1059 <bypasszoneRoute src="my_cluster_1" dst="my_cluster_2"
1060 gw_src="my_cluster_1_router"
1061 gw_dst="my_cluster_2_router">
1062 <link_ctn id="link_tmp"/>
1066 This example shows that link @c link_tmp (definition not displayed here) directly
1067 connects the router @c my_cluster_1_router in the source cluster to the router
1068 @c my_cluster_2_router in the destination router. Additionally, as the @c symmetrical
1069 attribute was not given, this route is presumed to be symmetrical.
1071 @subsubsection pf_tag_bypassroute bypassRoute
1073 As said before, once you choose
1074 a model, it (most likely; the constant network model, for example, doesn't) calculates routes for you. But maybe you want to
1075 define some of your routes, which will be specific. You may also want
1076 to bypass some routes defined in lower level zone at an upper stage:
1077 <b>bypassRoute</b> is the tag you're looking for. It allows to bypass
1078 routes defined between <b>host/router</b>. The principle is the same
1079 as route: <b>bypassRoute</b> contains list of links references of
1080 links that are in the path between src and dst.
1082 #### Attributes ####
1084 | Attribute name | Mandatory | Values | Description |
1085 | --------------- | --------- | ---------------------- | ----------- |
1086 | src | yes | String | The value given to the source zone's "id" attribute |
1087 | dst | yes | String | The value given to the destination zone's "id" attribute. |
1088 | symmetrical | no | YES @| NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1093 <bypassRoute src="host_1" dst="host_2">
1094 <link_ctn id="link_tmp"/>
1098 This example shows that link @c link_tmp (definition not displayed here) directly
1099 connects host @c host_1 to host @c host_2. Additionally, as the @c symmetrical
1100 attribute was not given, this route is presumed to be symmetrical.
1102 @subsection pb_baroex Basic Routing Example
1104 Let's say you have an zone named zone_Big that contains two other zone, zone_1
1105 and zone_2. If you want to make a host (h1) from zone_1 with another one
1106 (h2) from zone_2 then you'll have to proceed as follows:
1107 @li First, you have to ensure that a route is defined from h1 to the
1108 zone_1's exit gateway and from h2 to zone_2's exit gateway.
1109 @li Then, you'll have to define a route between zone_1 to zone_2. As those
1110 zone are both resources belonging to zone_Big, then it has to be done
1111 at zone_big level. To define such a route, you have to give the
1112 source zone (zone_1), the destination zone (zone_2), and their respective
1113 gateway (as the route is effectively defined between those two
1114 entry/exit points). Elements of this route can only be elements
1115 belonging to zone_Big, so links and routers in this route should be
1116 defined inside zone_Big. If you choose some shortest-path model,
1117 this route will be computed automatically.
1119 As said before, there are mainly 2 tags for routing:
1120 @li <b>zoneroute</b>: to define routes between two <b>zone</b>
1121 @li <b>route</b>: to define routes between two <b>host/router</b>
1123 As we are dealing with routes between zone, it means that those we'll
1124 have some definition at zone_Big level. Let consider zone_1 contains 1
1125 host, 1 link and one router and zone_2 3 hosts, 4 links and one router.
1126 There will be a central router, and a cross-like topology. At the end
1127 of the crosses arms, you'll find the 3 hosts and the router that will
1128 act as a gateway. We have to define routes inside those two zone. Let
1129 say that zone_1 contains full routes, and zone_2 contains some Floyd
1130 routing (as we don't want to bother with defining all routes). As
1131 we're using some shortest path algorithms to route into zone_2, we'll
1132 then have to define some <b>route</b> to gives some topological
1133 information to SimGrid. Here is a file doing it all:
1136 <zone id="zone_Big" routing="Dijkstra">
1137 <zone id="zone_1" routing="Full">
1138 <host id="zone_1_host1" speed="1000000000"/>
1139 <link id="zone_1_link" bandwidth="1250000000" latency="5E-4"/>
1140 <router id="zone_1_gateway"/>
1141 <route src="zone_1_host1" dst="zone_1_gateway">
1142 <link_ctn id="zone_1_link"/>
1145 <zone id="zone_2" routing="Floyd">
1146 <host id="zone_2_host1" speed="1000000000"/>
1147 <host id="zone_2_host2" speed="1000000000"/>
1148 <host id="zone_2_host3" speed="1000000000"/>
1149 <link id="zone_2_link1" bandwidth="1250000000" latency="5E-4"/>
1150 <link id="zone_2_link2" bandwidth="1250000000" latency="5E-4"/>
1151 <link id="zone_2_link3" bandwidth="1250000000" latency="5E-4"/>
1152 <link id="zone_2_link4" bandwidth="1250000000" latency="5E-4"/>
1153 <router id="central_router"/>
1154 <router id="zone_2_gateway"/>
1155 <!-- routes providing topological information -->
1156 <route src="central_router" dst="zone_2_host1"><link_ctn id="zone_2_link1"/></route>
1157 <route src="central_router" dst="zone_2_host2"><link_ctn id="zone_2_link2"/></route>
1158 <route src="central_router" dst="zone_2_host3"><link_ctn id="zone_2_link3"/></route>
1159 <route src="central_router" dst="zone_2_gateway"><link_ctn id="zone_2_link4"/></route>
1161 <link id="backbone" bandwidth="1250000000" latency="5E-4"/>
1163 <zoneroute src="zone_1" dst="zone_2"
1164 gw_src="zone_1_gateway"
1165 gw_dst="zone_2_gateway">
1166 <link_ctn id="backbone"/>
1171 @section pf_other Other tags
1173 The following tags can be used inside a @<platform@> tag even if they are not
1174 directly describing the platform:
1176 - @ref pf_tag_config passes configuration options, e.g. to change the network model;
1177 - @ref pf_tag_prop gives user-defined properties to various elements
1179 @subsection pf_tag_config <config>
1181 Adding configuration flags into the platform file is particularly
1182 useful when the described platform is best used with specific
1183 flags. For example, you could finely tune SMPI in your platform file directly.
1185 | Attribute | Values | Description |
1186 | ---------- | ------------------- | ---------------------------------------------- |
1187 | id | String (optional) | This optional identifier is ignored by SimGrid |
1189 * **Included tags:** @ref pf_tag_prop to specify a given configuration item (see @ref options).
1191 Any such configuration must be given at the very top of the platform file.
1196 <?xml version='1.0'?>
1197 <!DOCTYPE platform SYSTEM "https://simgrid.org/simgrid.dtd">
1198 <platform version="4">
1200 <prop id="maxmin/precision" value="0.000010" />
1201 <prop id="cpu/optim" value="TI" />
1202 <prop id="network/model" value="SMPI" />
1203 <prop id="smpi/bw-factor" value="65472:0.940694;15424:0.697866;9376:0.58729" />
1206 <zone id="zone0" routing="Full">
1210 @subsection pf_tag_prop <prop>
1212 Defines a user-defined property, identified with a name and having a
1213 value. You can specify such properties to most kind of resources:
1214 @ref pf_tag_zone, @ref pf_tag_host, @ref pf_tag_storage,
1215 @ref pf_tag_cluster and @ref pf_tag_link. These values can be retrieved
1216 at runtime with MSG_zone_property() or simgrid::s4u::NetZone::property(),
1217 or similar functions.
1219 | Attribute | Values | Description |
1220 | --------- | ---------------------- | ----------------------------------------------------------------------------------------- |
1221 | id | String (mandatory) | Identifier of this property. Must be unique for a given property holder, eg host or link. |
1222 | value | String (mandatory) | Value of this property; The semantic is completely up to you. |
1224 * **Included tags:** none.
1229 <prop id="Operating System" value="Linux" />
1233 @subsection pf_trace trace and trace_connect
1235 Both tags are an alternate way to pass files containing information on
1236 availability, state etc. to an entity. (See also @ref howto_churn).
1237 Instead of referring to the file directly in the host, link, or
1238 cluster tag, you proceed by defining a trace with an id corresponding
1239 to a file, later a host/link/cluster, and finally using trace_connect
1240 you say that the file trace must be used by the entity.
1246 <zone id="zone0" routing="Full">
1247 <host id="bob" speed="1000000000"/>
1249 <trace id="myTrace" file="bob.trace" periodicity="1.0"/>
1250 <trace_connect trace="myTrace" element="bob" kind="POWER"/>
1254 The order here is important. @c trace_connect must come
1255 after the elements @c trace and @c host, as both the host
1256 and the trace definition must be known when @c trace_connect
1257 is parsed; the order of @c trace and @c host is arbitrary.
1260 #### @c trace attributes ####
1263 | Attribute name | Mandatory | Values | Description |
1264 | --------------- | --------- | ---------------------- | ----------- |
1265 | id | yes | String | Identifier of this trace; this is the name you pass on to @c trace_connect. |
1266 | file | no | String | Filename of the file that contains the information - the path must follow the style of your OS. You can omit this, but then you must specifiy the values inside of <trace> and </trace> - see the example below. |
1267 | trace_periodicity | yes | String | This is the same as for @ref pf_tag_host "hosts" (see there for details) |
1269 Here is an example of trace when no file name is provided:
1272 <trace id="myTrace" periodicity="1.0">
1279 #### @c trace_connect attributes ####
1281 | Attribute name | Mandatory | Values | Description |
1282 | --------------- | --------- | ---------------------- | ----------- |
1283 | kind | no | HOST_AVAIL@|POWER@|<br/>LINK_AVAIL@|BANDWIDTH@|LATENCY (Default: HOST_AVAIL) | Describes the kind of trace. |
1284 | trace | yes | String | Identifier of the referenced trace (specified of the trace's @c id attribute) |
1285 | element | yes | String | The identifier of the referenced entity as given by its @c id attribute |
1287 @section pf_hints Hints, tips and frequently requested features
1289 Now you should know at least the syntax and be able to create a
1290 platform by your own. However, after having ourselves wrote some platforms, there
1291 are some best practices you should pay attention to in order to
1292 produce good platform and some choices you can make in order to have
1293 faster simulations. Here's some hints and tips, then.
1295 @subsection pf_hints_search Finding the platform example that you need
1297 Most platform files that we ship are in the @c examples/platforms
1298 folder. The good old @c grep tool can find the examples you need when
1299 wondering on a specific XML tag. Here is an example session searching
1300 for @ref pf_trace "trace_connect":
1303 % cd examples/platforms
1304 % grep -R -i -n --include="*.xml" "trace_connect" .
1305 ./two_hosts_platform_with_availability_included.xml:26:<trace_connect kind="SPEED" trace="A" element="Cpu A"/>
1306 ./two_hosts_platform_with_availability_included.xml:27:<trace_connect kind="HOST_AVAIL" trace="A_failure" element="Cpu A"/>
1307 ./two_hosts_platform_with_availability_included.xml:28:<trace_connect kind="SPEED" trace="B" element="Cpu B"/>
1308 ./two_hosts.xml:17: <trace_connect trace="Tremblay_power" element="Tremblay" kind="SPEED"/>
1311 @subsection pf_hint_generating How to generate different platform files?
1313 This is actually a good idea to search for a better platform file,
1314 that better fit the need of your study. To be honest, the provided
1315 examples are not representative of anything. They exemplify our XML
1316 syntax, but that's all. small_platform.xml for example was generated
1317 without much thought beyond that.
1319 The best thing to do when possible is to write your own platform file,
1320 that model the platform on which you run your code. For that, you
1321 could use <a href="https://gitlab.inria.fr/simgrid/platform-calibration">our
1322 calibration scripts</a>. This leads to very good fits between the
1323 platform, the model and the needs. The g5k.xml example resulted of
1324 such an effort, which also lead to <a href="https://github.com/lpouillo/topo5k/">an
1325 ongoing attempt</a> to automatically extract the SimGrid platform from
1326 the <a href="http://grid5000.fr/">Grid'5000</a> experimental platform.
1327 But it's hard to come up with generic models. Don't take these files
1328 too seriously. Actually, you should always challenge our models and
1329 your instanciation if the accuracy really matters to you (see <a
1330 href="https://hal.inria.fr/hal-00907887">this discussion</a>).
1332 But such advices only hold if you have a real platform and a real
1333 application at hand. It's moot for more abstract studies working on
1334 ideas and algorithms instead of technical artefacts. Well, in this
1335 case, there unfortunately is nothing better than this old and rusty
1336 <a href="http://pda.gforge.inria.fr/tools/download.html">simulacrum</a>.
1337 This project is dormant since over 10 years (and you will have to
1338 update the generated platforms with <tt>bin/simgrid_update_xml</tt> to
1339 use them), but that's the best we have for this right now....
1341 @subsection pf_zone_h Zone Hierarchy
1342 The network zone design allows SimGrid to go fast, because computing route is
1343 done only for the set of resources defined in the current zone. If you're using
1344 only a big zone containing all resource with no zone into it and you're
1345 using Full model, then ... you'll loose all interest into it. On the
1346 other hand, designing a binary tree of zone with, at the lower level,
1347 only one host, then you'll also loose all the good zone hierarchy can
1348 give you. Remind you should always be "reasonable" in your platform
1349 definition when choosing the hierarchy. A good choice if you try to
1350 describe a real life platform is to follow the zone described in
1351 reality, since this kind of trade-off works well for real life
1354 @subsection pf_exit_zone Exit Zone: why and how
1355 Users that have looked at some of our platforms may have notice a
1356 non-intuitive schema ... Something like that:
1360 <zone id="zone_4" routing="Full">
1361 <zone id="exitzone_4" routing="Full">
1362 <router id="router_4"/>
1364 <cluster id="cl_4_1" prefix="c_4_1-" suffix="" radical="1-20" speed="1000000000" bw="125000000" lat="5E-5" bb_bw="2250000000" bb_lat="5E-4"/>
1365 <cluster id="cl_4_2" prefix="c_4_2-" suffix="" radical="1-20" speed="1000000000" bw="125000000" lat="5E-5" bb_bw="2250000000" bb_lat="5E-4"/>
1366 <link id="4_1" bandwidth="2250000000" latency="5E-5"/>
1367 <link id="4_2" bandwidth="2250000000" latency="5E-5"/>
1368 <link id="bb_4" bandwidth="2250000000" latency="5E-4"/>
1369 <zoneroute src="cl_4_1"
1371 gw_src="c_4_1-cl_4_1_router"
1372 gw_dst="c_4_2-cl_4_2_router">
1373 <link_ctn id="4_1"/>
1374 <link_ctn id="bb_4"/>
1375 <link_ctn id="4_2"/>
1377 <zoneroute src="cl_4_1"
1379 gw_src="c_4_1-cl_4_1_router"
1381 <link_ctn id="4_1"/>
1382 <link_ctn id="bb_4"/>
1384 <zoneroute src="cl_4_2"
1386 gw_src="c_4_2-cl_4_2_router"
1388 <link_ctn id="4_2"/>
1389 <link_ctn id="bb_4"/>
1394 In the zone_4, you have an exitzone_4 defined, containing only one router,
1395 and routes defined to that zone from all other zone (as cluster is only a
1396 shortcut for an zone, see cluster description for details). If there was
1397 an upper zone, it would define routes to and from zone_4 with the gateway
1398 router_4. It's just because, as we did not allowed (for performances
1399 issues) to have routes from an zone to a single host/router, you have to
1400 enclose your gateway, when you have zone included in your zone, within an
1401 zone to define routes to it.
1403 @subsection pf_P2P_tags P2P or how to use coordinates
1404 SimGrid allows you to use some coordinated-based system, like vivaldi,
1405 to describe a platform. The main concept is that you have some peers
1406 that are located somewhere: this is the function of the
1407 <b>coordinates</b> of the @<peer@> or @<host@> tag. There's nothing
1408 complicated in using it, here is an example:
1411 <?xml version='1.0'?>
1412 <!DOCTYPE platform SYSTEM "https://simgrid.org/simgrid.dtd">
1413 <platform version="4">
1415 <zone id="zone0" routing="Vivaldi">
1416 <host id="100030591" coordinates="25.5 9.4 1.4" speed="1.5Gf" />
1417 <host id="100036570" coordinates="-12.7 -9.9 2.1" speed="7.3Gf" />
1419 <host id="100429957" coordinates="17.5 6.7 18.8" speed="8.3Gf" />
1424 Coordinates are then used to calculate latency (in microseconds)
1425 between two hosts by calculating the distance between the two hosts
1426 coordinates with the following formula: distance( (x1, y1, z1), (x2,
1427 y2, z2) ) = euclidian( (x1,y1), (x2,y2) ) + abs(z1) + abs(z2)
1429 In other words, we take the euclidian distance on the two first
1430 dimensions, and then add the absolute values found on the third
1431 dimension. This may seem strange, but it was found to allow better
1432 approximations of the latency matrices (see the paper describing
1435 Note that the previous example defines a routing directly between hosts but it could be also used to define a routing between zone.
1436 That is for example what is commonly done when using peers (see Section @ref pf_peer).
1438 <?xml version='1.0'?>
1439 <!DOCTYPE platform SYSTEM "https://simgrid.org/simgrid.dtd">
1440 <platform version="4">
1442 <zone id="zone0" routing="Vivaldi">
1443 <peer id="peer-0" coordinates="173.0 96.8 0.1" speed="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us"/>
1444 <peer id="peer-1" coordinates="247.0 57.3 0.6" speed="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us" />
1445 <peer id="peer-2" coordinates="243.4 58.8 1.4" speed="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us" />
1449 In such a case though, we connect the zone created by the <b>peer</b> tag with the Vivaldi routing mechanism.
1450 This means that to route between zone1 and zone2, it will use the coordinates of router_zone1 and router_zone2.
1451 This is currently a convention and we may offer to change this convention in the DTD later if needed.
1452 You may have noted that conveniently, a peer named FOO defines an zone named FOO and a router named router_FOO, which is why it works seamlessly with the <b>peer</b> tag.
1455 @subsection pf_routing_howto_choose_wisely Choosing wisely the routing model to use
1458 Choosing wisely the routing model to use can significantly fasten your
1459 simulation/save your time when writing the platform/save tremendous
1460 disk space. Here is the list of available model and their
1461 characteristics (lookup: time to resolve a route):
1463 @li <b>Full</b>: Full routing data (fast, large memory requirements,
1465 @li <b>Floyd</b>: Floyd routing data (slow initialization, fast
1466 lookup, lesser memory requirements, shortest path routing only).
1467 Calculates all routes at once at the beginning.
1468 @li <b>Dijkstra</b>: Dijkstra routing data (fast initialization, slow
1469 lookup, small memory requirements, shortest path routing only).
1470 Calculates a route when necessary.
1471 @li <b>DijkstraCache</b>: Dijkstra routing data (fast initialization,
1472 fast lookup, small memory requirements, shortest path routing
1473 only). Same as Dijkstra, except it handles a cache for latest used
1475 @li <b>None</b>: No routing (usable with Constant network only).
1476 Defines that there is no routes, so if you try to determine a
1477 route without constant network within this zone, SimGrid will raise
1479 @li <b>Vivaldi</b>: Vivaldi routing, so when you want to use coordinates
1480 @li <b>Cluster</b>: Cluster routing, specific to cluster tag, should
1483 @subsection pf_loopback I want to specify the characteristics of the loopback link!
1485 Each routing model automatically adds a loopback link for each declared host, i.e.,
1486 a network route from the host to itself, if no such route is declared in the XML
1487 file. This default link has a bandwidth of 498 Mb/s, a latency of 15 microseconds,
1488 and is <b>not</b> shared among network flows.
1490 If you want to specify the characteristics of the loopback link for a given host, you
1491 just have to specify a route from this host to itself with the desired characteristics
1492 in the XML file. This will prevent the routing model to add and use the default
1495 @subsection pf_switch I want to describe a switch but there is no switch tag!
1497 Actually we did not include switch tag. But when you're trying to
1498 simulate a switch, assuming
1499 fluid bandwidth models are used (which SimGrid uses by default unless
1500 ns-3 or constant network models are activated), the limiting factor is
1501 switch backplane bandwidth. So, essentially, at least from
1502 the simulation perspective, a switch is similar to a
1503 link: some device that is traversed by flows and with some latency and
1504 so,e maximum bandwidth. Thus, you can simply simulate a switch as a
1506 can be connected to this "switch", which is then included in routes just
1510 @subsection pf_multicabinets I want to describe multi-cabinets clusters!
1512 You have several possibilities, as usual when modeling things. If your
1513 cabinets are homogeneous and the intercabinet network negligible for
1514 your study, you should just create a larger cluster with all hosts at
1517 In the rare case where your hosts are not homogeneous between the
1518 cabinets, you can create your cluster completely manually. For that,
1519 create an As using the Cluster routing, and then use one
1520 <cabinet> for each cabinet. This cabinet tag can only be used an
1521 As using the Cluster routing schema, and creating
1523 Be warned that creating a cluster manually from the XML with
1524 <cabinet>, <backbone> and friends is rather tedious. The
1525 easiest way to retrieve some control of your model without diving into
1526 the <cluster> internals is certainly to create one separate
1527 <cluster> per cabinet and interconnect them together. This is
1528 what we did in the G5K example platform for the Graphen cluster.
1530 @subsection pf_platform_multipath I want to express multipath routing in platform files!
1532 It is unfortunately impossible to express the fact that there is more
1533 than one routing path between two given hosts. Let's consider the
1534 following platform file:
1537 <route src="A" dst="B">
1540 <route src="B" dst="C">
1543 <route src="A" dst="C">
1548 Although it is perfectly valid, it does not mean that data traveling
1549 from A to C can either go directly (using link 3) or through B (using
1550 links 1 and 2). It simply means that the routing on the graph is not
1551 trivial, and that data do not following the shortest path in number of
1552 hops on this graph. Another way to say it is that there is no implicit
1553 in these routing descriptions. The system will only use the routes you
1554 declare (such as <route src="A" dst="C"><link_ctn
1555 id="3"/></route>), without trying to build new routes by aggregating
1558 You are also free to declare platform where the routing is not
1559 symmetrical. For example, add the following to the previous file:
1562 <route src="C" dst="A">
1568 This makes sure that data from C to A go through B where data from A
1569 to C go directly. Don't worry about realism of such settings since
1570 we've seen ways more weird situation in real settings (in fact, that's
1571 the realism of very regular platforms which is questionable, but
1572 that's another story).