1 /*! \page platform Describing the virtual platform
5 As @ref starting_components "explained in the introduction," any
6 SimGrid study must entail the description of the platform on which you
7 want to simulate your application. You have to describe **each element
8 of your platform**, such as computing hosts, clusters, each disks,
9 links, etc. You must also define the **routing on your platform**, ie
10 which path is taken between two hosts. Finally, you may also describe
11 an **experimental scenario**, with qualitative changes (e.g.,
12 bandwidth changes representing an external load) and qualitative
13 changes (representing how some elements fail and restart over time).
15 You should really separate your application from the platform
16 description, as it will ease your experimental campain afterward.
17 Mixing them is seen as a really bad experimental practice. The easiest
18 to enforce this split is to put the platform description in a XML
19 file. Many example platforms are provided in the archive, and this
20 page gives all needed details to write such files, as well as some
21 hints and tricks about describing your platform.
23 On the other side, XML is sometimes not expressive enough for some
24 platforms, in particular large platforms exhibiting repetitive
25 patterns that are not simply expressed in XML. In practice, many
26 users end up generating their XML platform files from some sort of
27 scripts. It is probably preferable to rewrite your XML @ref
28 platform_lua "platform using the lua scripting language" instead.
29 In the future, it should be possible to describe the platform directly
30 in C++, but this is not possible yet.
32 As usual, SimGrid is a versatile framework, and you should find the
33 way of describing your platform that best fits your experimental
36 \section pf_overview Describing the platform with XML
38 Your platform description should follow the specification presented in
39 the [simgrid.dtd](http://simgrid.gforge.inria.fr/simgrid/simgrid.dtd)
40 DTD file. The same DTD is used for both the platform and deployment
43 From time to time, this DTD evolves to introduce possibly
44 backward-incompatible changes. That is why each platform desciption is
45 enclosed within a @c platform tag, that have a @c version attribute.
46 The current version is <b>4.1</b>. The @c simgrid_update_xml program can
47 upgrade most of the past platform files to the recent formalism.
49 \section pf_first_example First Platform Example
51 Here is a very simple platform file, containing 3 resources (two hosts
52 and one link), and explicitly giving the route between the hosts.
56 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid/simgrid.dtd">
57 <platform version="4.1">
58 <zone id="first zone" routing="Full">
59 <!-- the resources -->
60 <host id="host1" speed="1Mf"/>
61 <host id="host2" speed="2Mf"/>
62 <link id="link1" bandwidth="125MBps" latency="100us"/>
63 <!-- the routing: specify how the hosts are interconnected -->
64 <route src="host1" dst="host2">
65 <link_ctn id="link1"/>
71 As we said, the englobing @ref pf_overview "<platform>" tag is
72 used to specify the dtd version used for this file.
74 Then, every resource (specified with @ref pf_tag_host, @ref
75 pf_tag_link or others) must be located within a given **networking
76 zone**. Each zone is in charge of the routing between its
77 resources. It means that when an host wants to communicate with
78 another host of the same zone, it is the zone's duty to find the list
79 of links that are involved in the communication. Here, since the @ref
80 pf_tag_zone tag has **Full** as a **routing attribute**, all routes
81 must be explicitely given using the @ref pf_tag_route and @ref
82 pf_tag_linkctn tags (this @ref pf_rm "routing model" is both simple
83 and inefficient :) It is OK to not specify the route between two
84 hosts, as long as the processes located on them never try to
87 A zone can contain several zones itself, leading to a hierarchical
88 decomposition of the platform. This can be more efficient (as the
89 inter-zone routing gets factorized with @ref pf_tag_zoneroute), and
90 allows to have more than one routing model in your platform. For
91 example, you could have a coordinate-based routing for the WAN parts
92 of your platforms, a full routing within each datacenter, and a highly
93 optimized routing within each cluster of the datacenter. In this
94 case, determining the route between two given hosts gets @ref
95 routing_basics "somewhat more complex" but SimGrid still computes
96 these routes for you in a time- and space-efficient manner.
97 Here is an illustration of these concepts:
99 ![A hierarchy of networking zones.](AS_hierarchy.png)
101 Circles represent processing units and squares represent network
102 routers. Bold lines represent communication links. The zone "AS2"
103 models the core of a national network interconnecting a small flat
104 cluster (AS4) and a larger hierarchical cluster (AS5), a subset of a
105 LAN (AS6), and a set of peers scattered around the world (AS7).
107 \section pf_pftags Resource description
109 \subsection pf_tag_zone <zone>
111 For historical reasons, the XML files use the expression AS for
112 NetZones. Netzones are very important because they group other resources (such
113 as routers/hosts) together (in fact, any such resource must be
114 contained in a NetZone).
116 Available attributes :
118 Attribute name | Mandatory | Values | Description
119 --------------- | --------- | ------ | -----------
120 id | yes | String | The identifier of an AS; facilitates referring to this AS. ID must be unique.
121 routing | yes | Full\| Floyd\| Dijkstra\| DijkstraCache\| None\| Vivaldi\| Cluster | See Section \ref pf_rm for details.
126 <AS id="AS0" routing="Full">
127 <host id="host1" speed="1000000000"/>
128 <host id="host2" speed="1000000000"/>
129 <link id="link1" bandwidth="125000000" latency="0.000100"/>
130 <route src="host1" dst="host2"><link_ctn id="link1"/></route>
134 In this example, AS0 contains two hosts (host1 and host2). The route
135 between the hosts goes through link1.
137 \subsection pf_Cr Computing resources: hosts, clusters and peers.
139 \subsubsection pf_tag_host <host>
141 A <b>host</b> represents a computer/node card. Every host is able to execute
142 code and it can send and receive data to/from other hosts. Most importantly,
143 a host can contain more than 1 core.
147 Attribute name | Mandatory | Values | Description
148 --------------- | --------- | ------ | -----------
149 id | yes | String | The identifier of the host. facilitates referring to this AS.
150 speed | yes | double (must be > 0.0) | Computational power of every core of this host in FLOPS. Must be larger than 0.0.
151 core | no | int (Default: 1) | The number of cores of this host. If more than one core is specified, the "speed" parameter refers to every core, i.e., the total computational power is no_of_cores*speed.<br /> If 6 cores are specified, up to 6 tasks can be executed without sharing the computational power; if more than 6 tasks are executed, computational power will be shared among these tasks. <br /> <b>Warning:</b> Although functional, this model was never scientifically assessed.
152 availability_file| no | string | (Relative or absolute) filename to use as input; must contain availability traces for this host. The syntax of this file is defined below. <br /> <b>Note:</b> The filename must be specified with your system's format.
153 state_file | no | string | Same mechanism as availability_file.<br /> <b>Note:</b> The filename must be specified with your system's format.
154 coordinates | no | string | Must be provided when choosing the Vivaldi, coordinate-based routing model for the AS the host belongs to. More details can be found in the Section \ref pf_P2P_tags.
155 pstate | no | double (Default: 0.0) | FIXME: Not yet documented.
157 ### Possible children: ###
159 Tag name | Description | Documentation
160 ------------ | ----------- | -------------
161 \<mount/\> | Defines mounting points between some storage resource and the host. | \ref pf_storage_entity_mount
162 \<prop/\> | The prop tag allows you to define additional information on this host following the attribute/value schema. You may want to use it to give information to the tool you use for rendering your simulation, for example. | N/A
167 <host id="host1" speed="1000000000"/>
168 <host id="host2" speed="1000000000">
169 <prop id="color" value="blue"/>
170 <prop id="rendershape" value="square"/>
175 \anchor pf_host_dynamism
176 ### Expressing dynamism ###
178 SimGrid provides mechanisms to change a hosts' availability over
179 time, using the ``availability_file`` attribute to the ``\<host\>`` tag
180 and a separate text file whose syntax is exemplified below.
182 #### Adding a trace file ####
185 <platform version="4">
186 <host id="bob" speed="500Gf" availability_file="bob.trace" />
190 #### Example of "bob.trace" file ####
199 Let us begin to explain this example by looking at line 2. (Line 1 will become clear soon).
200 The first column describes points in time, in this case, time 0. The second column
201 describes the relative amount of power this host is able to deliver (relative
202 to the maximum performance specified in the ``\<host\>`` tag). (Clearly, the
203 second column needs to contain values that are not smaller than 0 and not larger than 1).
204 In this example, our host will deliver 500 Mflop/s at time 0, as 500 Mflop/s is the
205 maximum performance of this host. At time 11.0, it will
206 deliver half of its maximum performance, i.e., 250 Mflop/s until time 20.0 when it will
207 will start delivering 80\% of its power. In this example, this amounts to 400 Mflop/s.
209 Since the periodicity in line 1 was set to be 1.0, i.e., 1 timestep, this host will
210 continue to provide 500 Mflop/s from time 21. From time 32 it will provide 250 MFlop/s and so on.
212 ### Changing initial state ###
214 It is also possible to specify whether the host is up or down by setting the
215 ``state`` attribute to either <b>ON</b> (default value) or <b>OFF</b>.
217 #### Example: Expliciting the default value "ON" ####
220 <platform version="4">
221 <host id="bob" speed="500Gf" state="ON" />
225 If you want this host to be unavailable, simply substitute ON with OFF.
227 \anchor pf_host_churn
228 ### Expressing churn ###
230 To express the fact that a host can change state over time (as in P2P
231 systems, for instance), it is possible to use a file describing the time
232 at which the host is turned on or off. An example of the content
233 of such a file is presented below.
235 #### Adding a state file ####
238 <platform version="4">
239 <host id="bob" power="500Gf" state_file="bob.fail" />
243 #### Example of "bob.fail" file ####
251 A zero value means <b>down</b> (i.e., OFF) while a positive one means <b>up and
252 running</b> (i.e., ON). From time 0.0 to time 1.0, the host is on as usual. At time 1.0, it is
253 turned off and at time 2.0, it is turned on again until time 12 (2 plus the
254 periodicity 10). It will be turned off again at time 13.0 until time 23.0, and
258 \subsubsection pf_cluster <cluster>
260 ``<cluster />`` represents a machine-cluster. It is most commonly used
261 when one wants to define many hosts and a network quickly. Technically,
262 ``cluster`` is a meta-tag: <b>from the inner SimGrid point of
263 view, a cluster is an AS where some optimized routing is defined</b>.
264 The default inner organization of the cluster is as follow:
270 ____________|__________|_____________ backbone
272 l0| l1| l2| l97| l96 | | l99
278 Here, a set of <b>host</b>s is defined. Each of them has a <b>link</b>
279 to a central backbone (backbone is a link itself, as a link can
280 be used to represent a switch, see the switch / link section
281 below for more details about it). A <b>router</b> allows to connect a
282 <b>cluster</b> to the outside world. Internally,
283 SimGrid treats a cluster as an AS containing all hosts: the router is the default
284 gateway for the cluster.
286 There is an alternative organization, which is as follows:
300 The principle is the same, except that there is no backbone. This representation
301 can be obtained easily: just do not set the bb_* attributes.
304 Attribute name | Mandatory | Values | Description
305 --------------- | --------- | ------ | -----------
306 id | yes | string | The identifier of the cluster. Facilitates referring to this cluster.
307 prefix | yes | string | Each node of the cluster has to have a name. This name will be prefixed with this prefix.
308 suffix | yes | string | Each node of the cluster will be suffixed with this suffix
309 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.
310 speed | yes | int | Same as the ``speed`` attribute of the ``\<host\>`` tag.
311 core | no | int (default: 1) | Same as the ``core`` attribute of the ``\<host\>`` tag.
312 bw | yes | int | Bandwidth for the links between nodes and backbone (if any). See the \ref pf_tag_link "link section" for syntax/details.
313 lat | yes | int | Latency for the links between nodes and backbone (if any). See <b>link</b> section for syntax/details.
314 sharing_policy | no | string | Sharing policy for the links between nodes and backbone (if any). See <b>link</b> section for syntax/details.
315 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>).
316 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>).
317 bb_sharing_policy | no | string | Sharing policy for the backbone (if any). See <b>link</b> section for syntax/details.
318 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 fullduplex 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>
319 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".
320 loopback_lat | no | int | Latency for loopback (if any). See <b>link</b> section for syntax/details. See loopback_bw for more info.
321 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.
322 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".
325 the router name is defined as the resulting String in the following
329 router_name = prefix + clusterId + "_router" + suffix;
333 #### Cluster example ####
335 Consider the following two (and independent) uses of the ``cluster`` tag:
338 <cluster id="my_cluster_1" prefix="" suffix="" radical="0-262144"
339 speed="1e9" bw="125e6" lat="5E-5"/>
341 <cluster id="my_cluster_2" prefix="c-" suffix=".me" radical="0-99"
342 speed="1e9" bw="125e6" lat="5E-5"
343 bb_bw="2.25e9" bb_lat="5E-4"/>
346 The second example creates one router and 100 machines with the following names:
348 c-my_cluster_2_router.me
356 \subsubsection pf_cabinet <cabinet>
359 This tag is only available when the routing mode of the AS
360 is set to ``Cluster``.
362 The ``<cabinet />`` tag is, like the \ref pf_cluster "<cluster>" tag,
363 a meta-tag. This means that it is simply a shortcut for creating a set of (homogenous) hosts and links quickly;
364 unsurprisingly, this tag was introduced to setup cabinets in data centers quickly. Unlike
365 <cluster>, however, the <cabinet> assumes that you create the backbone
366 and routers yourself; see our examples below.
370 Attribute name | Mandatory | Values | Description
371 --------------- | --------- | ------ | -----------
372 id | yes | string | The identifier of the cabinet. Facilitates referring to this cluster.
373 prefix | yes | string | Each node of the cabinet has to have a name. This name will be prefixed with this prefix.
374 suffix | yes | string | Each node of the cabinet will be suffixed with this suffix
375 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.
376 speed | yes | int | Same as the ``speed`` attribute of the \ref pf_tag_host "<host>" tag.
377 bw | yes | int | Bandwidth for the links between nodes and backbone (if any). See the \ref pf_tag_link "link section" for syntax/details.
378 lat | yes | int | Latency for the links between nodes and backbone (if any). See the \ref pf_tag_link "link section" for syntax/details.
381 Please note that as of now, it is impossible to change attributes such as,
382 amount of cores (always set to 1), the initial state of hosts/links
383 (always set to ON), the sharing policy of the links (always set to \ref pf_sharing_policy_fullduplex "FULLDUPLEX").
387 The following example was taken from ``examples/platforms/meta_cluster.xml`` and
388 shows how to use the cabinet tag.
391 <AS id="my_cluster1" routing="Cluster">
392 <cabinet id="cabinet1" prefix="host-" suffix=".cluster1"
393 speed="1Gf" bw="125MBps" lat="100us" radical="1-10"/>
394 <cabinet id="cabinet2" prefix="host-" suffix=".cluster1"
395 speed="1Gf" bw="125MBps" lat="100us" radical="11-20"/>
396 <cabinet id="cabinet3" prefix="host-" suffix=".cluster1"
397 speed="1Gf" bw="125MBps" lat="100us" radical="21-30"/>
399 <backbone id="backbone1" bandwidth="2.25GBps" latency="500us"/>
404 Please note that you must specify the \ref pf_backbone "<backbone>"
405 tag by yourself; this is not done automatically and there are no checks
406 that ensure this backbone was defined.
408 The hosts generated in the above example are named host-1.cluster, host-2.cluster1
412 \subsubsection pf_peer \<peer\> (Vivaldi netzones only)
414 This tag represents a peer, as in Peer-to-Peer (P2P) networks. This
415 can only be used in Vivaldi NetZones. It creates the following
416 resources to the NetZone:
419 \li Two links: One for download and one for upload. This is
420 convenient to use and simulate stuff under the last mile model (e.g., ADSL peers).
421 \li It connects the two links to the host
425 Attribute name | Mandatory | Values | Description
426 --------------- | --------- | ------ | -----------
427 id | yes | string | The identifier of the peer. Facilitates referring to this peer.
428 speed | yes | int | See the description of the ``host`` tag for this attribute
429 bw_in | yes | int | Bandwidth of the private downstream link
430 bw_out | yes | int | Bandwidth of the private upstream link
431 coordinates | no | string | Coordinates of the gateway for this peer. Example value: 12.8 14.4 6.4
432 sharing_policy | no | SHARED\|FULLDUPLEX (default: FULLDUPLEX) | Sharing policy for links. See <b>link</b> description for details.
433 availability_file| no | string | Availability file for the peer. Same as host availability file. See <b>host</b> description for details.
434 state_file | no | string | State file for the peer. Same as host state file. See <b>host</b> description for details.
437 The communication latency between an host A=(xA,yA,zA) and an host
438 B=(xB,yB,zB) is computed as follows:
440 latency = sqrt( (xA-xB)² + (yA-yB)² ) + zA + zB
442 See the documentation of simgrid::kernel::routing::VivaldiZone for
443 details on how the latency is computed from the coordinate, and on the
444 the up and down bandwidth are used.
446 \subsection pf_ne Network equipments: links and routers
448 There are two tags at all times available to represent network entities and
449 several other tags that are available only in certain contexts.
450 1. ``<link>``: Represents a entity that has a limited bandwidth, a
451 latency, and that can be shared according to TCP way to share this
454 The concept of links in SimGrid may not be intuitive, as links are not
455 limited to connecting (exactly) two entities; in fact, you can have more than
456 two equipments connected to it. (In graph theoretical terms: A link in
457 SimGrid is not an edge, but a hyperedge)
459 2. ``<router/>``: Represents an entity that a message can be routed
460 to, but that is unable to execute any code. In SimGrid, routers have also
461 no impact on the performance: Routers do not limit any bandwidth nor
462 do they increase latency. As a matter of fact, routers are (almost) ignored
463 by the simulator when the simulation has begun.
465 3. ``<backbone/>``: This tag is only available when the containing AS is
466 used as a cluster (i.e., mode="Cluster")
469 If you want to represent an entity like a switch, you must use ``<link>`` (see section). Routers are used
470 to run some routing algorithm and determine routes (see Section \ref pf_routing for details).
472 \subsubsection pf_router <router/>
474 As said before, <b>router</b> is used only to give some information
475 for routing algorithms. So, it does not have any attributes except :
479 Attribute name | Mandatory | Values | Description
480 --------------- | --------- | ------ | -----------
481 id | yes | string | The identifier of the router to be used when referring to it.
482 coordinates | no | string | Must be provided when choosing the Vivaldi, coordinate-based routing model for the AS the router belongs to. More details can be found in the Section \ref pf_P2P_tags.
487 <router id="gw_dc1_horizdist"/>
490 \subsubsection pf_tag_link <link>
492 Network links can represent one-hop network connections. They are
493 characterized by their id and their bandwidth; links can (but may not) be subject
498 Attribute name | Mandatory | Values | Description
499 --------------- | --------- | ------ | -----------
500 id | yes | string | The identifier of the link to be used when referring to it.
501 bandwidth | yes | int | Maximum bandwidth for this link, given in bytes/s
502 latency | no | double (default: 0.0) | Latency for this link.
503 sharing_policy | no | \ref sharing_policy_shared "SHARED"\|\ref pf_sharing_policy_fatpipe "FATPIPE"\|\ref pf_sharing_policy_fullduplex "FULLDUPLEX" (default: SHARED) | Sharing policy for the link.
504 state | no | ON\|OFF (default: ON) | Allows you to to turn this link on or off (working / not working)
505 bandwidth_file | no | string | Allows you to use a file as input for bandwidth.
506 latency_file | no | string | Allows you to use a file as input for latency.
507 state_file | no | string | Allows you to use a file as input for states.
510 #### Possible shortcuts for ``latency`` ####
512 When using the latency attribute, you can specify the latency by using the scientific
513 notation or by using common abbreviations. For instance, the following three tags
517 <link id="LINK1" bandwidth="125000000" latency="5E-6"/>
518 <link id="LINK1" bandwidth="125000000" latency="5us"/>
519 <link id="LINK1" bandwidth="125000000" latency="0.000005"/>
522 Here, the second tag uses "us", meaning "microseconds". Other shortcuts are:
524 Name | Abbreviation | Time (in seconds)
525 ---- | ------------ | -----------------
526 Week | w | 7 * 24 * 60 * 60
527 Day | d | 24 * 60 * 60
531 Millisecond | ms | 0.001 = 10^(-3)
532 Microsecond | us | 0.000001 = 10^(-6)
533 Nanosecond | ns | 0.000000001 = 10^(-9)
534 Picosecond | ps | 0.000000000001 = 10^(-12)
536 #### Sharing policy ####
538 \anchor sharing_policy_shared
539 By default a network link is \b SHARED, i.e., if two or more data flows go
540 through a link, the bandwidth is shared fairly among all data flows. This
541 is similar to the sharing policy TCP uses.
543 \anchor pf_sharing_policy_fatpipe
544 On the other hand, if a link is defined as a \b FATPIPE,
545 each flow going through this link will be provided with the complete bandwidth,
546 i.e., no sharing occurs and the bandwidth is only limiting each flow individually.
547 Please note that this is really on a per-flow basis, not only on a per-host basis!
548 The complete bandwidth provided by this link in this mode
549 is ``number_of_flows*bandwidth``, with at most ``bandwidth`` being available per flow.
551 Using the FATPIPE mode allows to model backbones that won't affect performance
554 \anchor pf_sharing_policy_fullduplex
555 The last mode available is \b FULLDUPLEX. This means that SimGrid will
556 automatically generate two links (one carrying the suffix _UP and the other the
557 suffix _DOWN) for each ``<link>`` tag. This models situations when the direction
558 of traffic is important.
561 Transfers from one side to the other will interact similarly as
562 TCP when ACK returning packets circulate on the other direction. More
563 discussion about it is available in the description of link_ctn description.
565 In other words: The SHARED policy defines a physical limit for the bandwidth.
566 The FATPIPE mode defines a limit for each application,
567 with no upper total limit.
570 Tip: By using the FATPIPE mode, you can model big backbones that
571 won't affect performance (except latency).
576 <link id="SWITCH" bandwidth="125000000" latency="5E-5" sharing_policy="FATPIPE" />
579 #### Expressing dynamism and failures ####
581 Similar to hosts, it is possible to declare links whose state, bandwidth
582 or latency changes over time (see Section \ref pf_host_dynamism for details).
584 In the case of network links, the ``bandwidth`` and ``latency`` attributes are
585 replaced by the ``bandwidth_file`` and ``latency_file`` attributes.
586 The following XML snippet demonstrates how to use this feature in the platform
587 file. The structure of the files "link1.bw" and "link1.lat" is shown below.
590 <link id="LINK1" state_file="link1.fail" bandwidth="80000000" latency=".0001" bandwidth_file="link1.bw" latency_file="link1.lat" />
594 Even if the syntax is the same, the semantic of bandwidth and latency
595 trace files differs from that of host availability files. For bandwidth and
596 latency, the corresponding files do not
597 express availability as a fraction of the available capacity but directly in
598 bytes per seconds for the bandwidth and in seconds for the latency. This is
599 because most tools allowing to capture traces on real platforms (such as NWS)
600 express their results this way.
602 ##### Example of "link1.bw" file #####
610 In this example, the bandwidth changes repeatedly, with all changes
611 being repeated every 12 seconds.
613 At the beginning of the the simulation, the link's bandwidth is 80,000,000
614 B/s (i.e., 80 Mb/s); this value was defined in the XML snippet above.
615 After four seconds, it drops to 40 Mb/s (line 2), and climbs
616 back to 60 Mb/s after another 4 seconds (line 3). The value does not change any
617 more until the end of the period, that is, after 12 seconds have been simulated).
618 At this point, periodicity kicks in and this behavior is repeated: Seconds
619 12-16 will experience 80 Mb/s, 16-20 40 Mb/s etc.).
621 ##### Example of "link1.lat" file #####
630 In this example, the latency varies with a period of 5 seconds.
631 In the xml snippet above, the latency is initialized to be 0.0001s (100µs). This
632 value will be kept during the first second, since the latency_file contains
633 changes to this value at second one, two and three.
634 At second one, the value will be 0.001, i.e., 1ms. One second later it will
635 be adjusted to 0.01 (or 10ms) and one second later it will be set again to 1ms. The
636 value will not change until second 5, when the periodicity defined in line 1
637 kicks in. It then loops back, starting at 100µs (the initial value) for one second.
640 #### The ``<prop/>`` tag ####
642 Similar to the ``<host>`` tag, a link may also contain the ``<prop/>`` tag; see the host
643 documentation (Section \ref pf_host) for an example.
646 \subsubsection pf_backbone <backbone/>
649 This tag is <b>only available</b> when the containing AS uses the "Cluster" routing mode!
651 Using this tag, you can designate an already existing link to be a backbone.
653 Attribute name | Mandatory | Values | Description
654 --------------- | --------- | ------ | -----------
655 id | yes | string | Name of the link that is supposed to act as a backbone.
657 \subsection pf_storage Storage
660 This is a prototype version that should evolve quickly, hence this
661 is just some doc valuable only at the time of writing.
662 This section describes the storage management under SimGrid ; nowadays
663 it's only usable with MSG. It relies basically on linux-like concepts.
664 You also may want to have a look to its corresponding section in
665 @ref msg_file ; access functions are organized as a POSIX-like
668 \subsubsection pf_sto_conc Storage - Main Concepts
670 The storage facilities implemented in SimGrid help to model (and account for)
671 storage devices, such as tapes, hard-drives, CD or DVD devices etc.
672 A typical situation is depicted in the figure below:
674 \image html ./webcruft/storage_sample_scenario.png
675 \image latex ./webcruft/storage_sample_scenario.png "storage_sample_scenario" width=\textwidth
677 In this figure, two hosts called Bob and Alice are interconnected via a network
678 and each host is physically attached to a disk; it is not only possible for each host to
679 mount the disk they are attached to directly, but they can also mount disks
680 that are in a remote location. In this example, Bob mounts Alice's disk remotely
681 and accesses the storage via the network.
683 SimGrid provides 3 different entities that can be used to model setups
684 that include storage facilities:
686 Entity name | Description
687 --------------- | -----------
688 \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).
689 \ref pf_storage_entity_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.
690 \ref pf_storage_entity_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).
693 \anchor pf_storage_content_file
694 ### %Storage Content File ###
696 In order to assess exactly how much time is spent reading from the storage,
697 SimGrid needs to know what is stored on the storage device (identified by distinct (file-)name, like in a file system)
698 and what size this content has.
701 The content file is never changed by the simulation; it is parsed once
702 per simulation and kept in memory afterwards. When the content of the
703 storage changes, only the internal SimGrid data structures change.
705 \anchor pf_storage_content_file_structure
706 #### Structure of a %Storage Content File ####
708 Here is an excerpt from two storage content file; if you want to see the whole file, check
709 the file ``examples/platforms/content/storage_content.txt`` that comes with the
712 SimGrid essentially supports two different formats: UNIX-style filepaths should
713 follow the well known format:
716 /lib/libsimgrid.so.3.6.2 12710497
720 /bin/simgrid_update_xml 5018
721 /bin/graphicator 66986
722 /bin/simgrid-colorizer 2993
727 Windows filepaths, unsurprisingly, use the windows style:
730 \Windows\avastSS.scr 41664
731 \Windows\bfsvc.exe 75264
732 \Windows\bootstat.dat 67584
733 \Windows\CoreSingleLanguage.xml 31497
735 \Windows\dchcfg64.exe 335464
736 \Windows\dcmdev64.exe 93288
740 The different file formats come at a cost; in version 3.12 (and most likely
741 in later versions, too), copying files from windows-style storages to unix-style
742 storages (and vice versa) is not supported.
744 \anchor pf_storage_content_file_create
745 #### Generate a %Storage Content File ####
747 If you want to generate a storage content file based on your own filesystem (or at least a filesystem you have access to),
748 try running this command (works only on unix systems):
751 find . -type f -exec ls -1s --block=1 {} \; 2>/dev/null | awk '{ print $2 " " $1}' > ./content.txt
754 \subsubsection pf_storage_entities The Storage Entities
756 These are the entities that you can use in your platform files to include
757 storage in your model. See also the list of our \ref pf_storage_example_files "example files";
758 these might also help you to get started.
760 \anchor pf_storage_entity_storage_type
761 #### \<storage_type\> ####
763 Attribute name | Mandatory | Values | Description
764 --------------- | --------- | ------ | -----------
765 id | yes | string | Identifier of this storage_type; used when referring to it
766 model | yes | string | For reasons of future backwards compatibility only; specifies the name of the model for the storage that should be used
767 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)
768 content | yes | string | Path to a \ref pf_storage_content_file "Storage Content File" on your system. This file must exist.
769 content_type | no | ("txt_unix"\|"txt_win") | Determines which kind of filesystem you're using; make sure the filenames (stored in that file, see \ref pf_storage_content_file_structure "Storage Content File Structure"!)
771 This tag must contain some predefined model properties, specified via the <model_prop> tag. Here is a list,
772 see below for an example:
774 Property id | Mandatory | Values | Description
775 --------------- | --------- | ------ | -----------
776 Bwrite | yes | string | Bandwidth for write access; in B/s (but you can also specify e.g. "30MBps")
777 Bread | yes | string | Bandwidth for read access; in B/s (but you can also specify e.g. "30MBps")
778 Bconnexion | yes | string | Throughput (of the storage connector) in B/s.
781 A storage_type can also contain the <b><prop></b> tag. The <prop> tag allows you
782 to associate additional information to this <storage_type> and follows the
783 attribute/value schema; see the example below. You may want to use it to give information to
784 the tool you use for rendering your simulation, for example.
786 Here is a complete example for the ``storage_type`` tag:
788 <storage_type id="single_HDD" model="linear_no_lat" size="4000" content_type="txt_unix">
789 <model_prop id="Bwrite" value="30MBps" />
790 <model_prop id="Bread" value="100MBps" />
791 <model_prop id="Bconnection" value="150MBps" />
792 <prop id="Brand" value="Western Digital" />
796 \anchor pf_storage_entity_storage
797 #### <storage> ####
799 ``storage`` attributes:
801 Attribute name | Mandatory | Values | Description
802 -------------- | --------- | ------ | -----------
803 id | yes | string | Identifier of this ``storage``; used when referring to it
804 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.
805 attach | yes | string | Name of a host (see Section \ref pf_host) to which this storage is <i>physically</i> attached to (e.g., a hard drive in a computer)
806 content | no | string | When specified, overwrites the content attribute of \ref pf_storage_entity_storage_type "\<storage_type\>"
807 content_type | no | string | When specified, overwrites the content_type attribute of \ref pf_storage_entity_storage_type "\<storage_type\>"
809 Here are two examples:
812 <storage id="Disk1" typeId="single_HDD" attach="bob" />
814 <storage id="Disk2" typeId="single_SSD"
815 content="content/win_storage_content.txt"
816 content_type="txt_windows" attach="alice" />
819 The first example is straightforward: A disk is defined and called "Disk1"; it is
820 of type "single_HDD" (shown as an example of \ref pf_storage_entity_storage_type "\<storage_type\>" above) and attached
821 to a host called "bob" (the definition of this host is omitted here).
823 The second storage is called "Disk2", is still of the same type as Disk1 but
824 now specifies a new content file (so the contents will be different from Disk1)
825 and the filesystem uses the windows style; finally, it is attached to a second host,
826 called alice (which is again not defined here).
828 \anchor pf_storage_entity_mount
829 #### <mount> ####
832 | Attribute name | Mandatory | Values | Description |
833 | ---------------- | ----------- | -------- | ------------- |
834 | id | yes | string | Refers to a \ref pf_storage_entity_storage "<storage>" entity that will be mounted on that computer |
835 | name | yes | string | Path/location to/of the logical reference (mount point) of this disk
837 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)
838 is; this location is specified by the ``name`` attribute.
840 Here is a simple example, taken from the file ``examples/platform/storage.xml``:
843 <storage_type id="single_SSD" model="linear_no_lat" size="500GiB">
844 <model_prop id="Bwrite" value="60MBps" />
845 <model_prop id="Bread" value="200MBps" />
846 <model_prop id="Bconnection" value="220MBps" />
849 <storage id="Disk2" typeId="single_SSD"
850 content="content/win_storage_content.txt"
851 content_type="txt_windows" attach="alice" />
852 <storage id="Disk4" typeId="single_SSD"
853 content="content/small_content.txt"
854 content_type="txt_unix" attach="denise"/>
856 <host id="alice" speed="1Gf">
857 <mount storageId="Disk2" name="c:"/>
860 <host id="denise" speed="1Gf">
861 <mount storageId="Disk2" name="c:"/>
862 <mount storageId="Disk4" name="/home"/>
866 This example is quite interesting, as the same device, called "Disk2", is mounted by
867 two hosts at the same time! Note, however, that the host called ``alice`` is actually
868 attached to this storage, as can be seen in the \ref pf_storage_entity_storage "<storage>"
869 tag. This means that ``denise`` must access this storage through the network, but SimGrid automatically takes
870 care of that for you.
872 Furthermore, this example shows that ``denise`` has mounted two storages with different
873 filesystem types (unix and windows). In general, a host can mount as many storage devices as
877 Again, the difference between ``attach`` and ``mount`` is simply that
878 an attached storage is always physically inside (or connected to) that machine;
879 for instance, a USB stick is attached to one and only one machine (where it's plugged-in)
880 but it can only be mounted on others, as mounted storage can also be a remote location.
882 ###### Example files #####
884 \verbinclude example_filelist_xmltag_mount
886 \subsubsection pf_storage_example_files Example files
888 Several examples were already discussed above; if you're interested in full examples,
889 check the the following platforms:
891 1. ``examples/platforms/storage.xml``
892 2. ``examples/platforms/remote_io.xml``
894 If you're looking for some examplary C code, you may find the source code
895 available in the directory ``examples/msg/io/`` useful.
897 \subsubsection pf_storage_examples_modelling Modelling different situations
899 The storage functionality of SimGrid is type-agnostic, that is, the implementation
900 does not presume any type of storage, such as HDDs/SSDs, RAM,
901 CD/DVD devices, USB sticks etc.
903 This allows the user to apply the simulator for a wide variety of scenarios; one
904 common scenario would be the access of remote RAM.
906 #### Modelling the access of remote RAM ####
908 How can this be achieved in SimGrid? Let's assume we have a setup where three hosts
909 (HostA, HostB, HostC) need to access remote RAM:
919 An easy way to model this scenario is to setup and define the RAM via the
920 \ref pf_storage_entity_storage "storage" and \ref pf_storage_entity_storage_type "storage type"
921 entities and attach it to a remote dummy host; then, every host can have their own links
922 to this host (modelling for instance certain scenarios, such as PCIe ...)
927 RAM - Dummy -- Host B
932 Now, if read from this storage, the host that mounts this storage
933 communicates to the dummy host which reads from RAM and
934 sends the information back.
937 \section pf_routing Routing
939 To achieve high performance, the routing tables used within SimGrid are
940 static. This means that routing between two nodes is calculated once
941 and will not change during execution. The SimGrid team chose to use this
942 approach as it is rare to have a real deficiency of a resource;
943 most of the time, a communication fails because the links experience too much
944 congestion and hence, your connection stops before the timeout or
945 because the computer designated to be the destination of that message
948 We also chose to use shortest paths algorithms in order to emulate
949 routing. Doing so is consistent with the reality: [RIP](https://en.wikipedia.org/wiki/Routing_Information_Protocol),
950 [OSPF](https://en.wikipedia.org/wiki/Open_Shortest_Path_First), [BGP](https://en.wikipedia.org/wiki/Border_Gateway_Protocol)
951 are all calculating shortest paths. They do require some time to converge, but
952 eventually, when the routing tables have stabilized, your packets will follow
955 \subsection pf_rm Routing models
957 For each AS, you must define explicitly which routing model will
958 be used. There are 3 different categories for routing models:
960 1. \ref pf_routing_model_shortest_path "Shortest-path" based models: SimGrid calculates shortest
961 paths and manages them. Behaves more or less like most real life
963 2. \ref pf_routing_model_manual "Manually-entered" route models: you have to define all routes
964 manually in the platform description file; this can become
965 tedious very quickly, as it is very verbose.
966 Consistent with some manually managed real life routing.
967 3. \ref pf_routing_model_simple "Simple/fast models": those models offer fast, low memory routing
968 algorithms. You should consider to use this type of model if
969 you can make some assumptions about your AS.
970 Routing in this case is more or less ignored.
972 \subsubsection pf_raf The router affair
974 Using routers becomes mandatory when using shortest-path based
975 models or when using the bindings to the ns-3 packet-level
976 simulator instead of the native analytical network model implemented
979 For graph-based shortest path algorithms, routers are mandatory, because these
980 algorithms require a graph as input and so we need to have source and
981 destination for each edge.
983 Routers are naturally an important concept ns-3 since the
984 way routers run the packet routing algorithms is actually simulated.
985 SimGrid's analytical models however simply aggregate the routing time
986 with the transfer time.
988 So why did we incorporate routers in SimGrid? Rebuilding a graph representation
989 only from the route information turns out to be a very difficult task, because
990 of the missing information about how routes intersect. That is why we
991 introduced routers, which are simply used to express these intersection points.
992 It is important to understand that routers are only used to provide topological
995 To express this topological information, a <b>route</b> has to be
996 defined in order to declare which link is connected to a router.
999 \subsubsection pf_routing_model_shortest_path Shortest-path based models
1001 The following table shows all the models that compute routes using
1002 shortest-paths algorithms are currently available in SimGrid. More detail on how
1003 to choose the best routing model is given in the Section called \"\ref pf_routing_howto_choose_wisely\".
1005 | Name | Description |
1006 | --------------------------------------------------- | -------------------------------------------------------------------------- |
1007 | \ref pf_routing_model_floyd "Floyd" | Floyd routing data. Pre-calculates all routes once |
1008 | \ref pf_routing_model_dijkstra "Dijkstra" | Dijkstra routing data. Calculates routes only when needed |
1009 | \ref pf_routing_model_dijkstracache "DijkstraCache" | Dijkstra routing data. Handles some cache for already calculated routes. |
1011 All those shortest-path models are instanciated in the same way and are
1012 completely interchangeable. Here are some examples:
1014 \anchor pf_routing_model_floyd
1019 <AS id="AS0" routing="Floyd">
1021 <cluster id="my_cluster_1" prefix="c-" suffix=""
1022 radical="0-1" speed="1000000000" bw="125000000" lat="5E-5"
1023 router_id="router1"/>
1025 <AS id="AS1" routing="None">
1026 <host id="host1" speed="1000000000"/>
1029 <link id="link1" bandwidth="100000" latency="0.01"/>
1031 <ASroute src="my_cluster_1" dst="AS1"
1034 <link_ctn id="link1"/>
1040 ASroute given at the end gives a topological information: link1 is
1041 between router1 and host1.
1043 #### Example platform files ####
1045 This is an automatically generated list of example files that use the Floyd
1046 routing model (the path is given relative to SimGrid's source directory)
1048 \verbinclude example_filelist_routing_floyd
1050 \anchor pf_routing_model_dijkstra
1053 #### Example platform files ####
1055 This is an automatically generated list of example files that use the Dijkstra
1056 routing model (the path is given relative to SimGrid's source directory)
1058 \verbinclude example_filelist_routing_dijkstra
1062 <AS id="AS_2" routing="Dijkstra">
1063 <host id="AS_2_host1" speed="1000000000"/>
1064 <host id="AS_2_host2" speed="1000000000"/>
1065 <host id="AS_2_host3" speed="1000000000"/>
1066 <link id="AS_2_link1" bandwidth="1250000000" latency="5E-4"/>
1067 <link id="AS_2_link2" bandwidth="1250000000" latency="5E-4"/>
1068 <link id="AS_2_link3" bandwidth="1250000000" latency="5E-4"/>
1069 <link id="AS_2_link4" bandwidth="1250000000" latency="5E-4"/>
1070 <router id="central_router"/>
1071 <router id="AS_2_gateway"/>
1072 <!-- routes providing topological information -->
1073 <route src="central_router" dst="AS_2_host1"><link_ctn id="AS_2_link1"/></route>
1074 <route src="central_router" dst="AS_2_host2"><link_ctn id="AS_2_link2"/></route>
1075 <route src="central_router" dst="AS_2_host3"><link_ctn id="AS_2_link3"/></route>
1076 <route src="central_router" dst="AS_2_gateway"><link_ctn id="AS_2_link4"/></route>
1080 \anchor pf_routing_model_dijkstracache
1081 ### DijkstraCache ###
1083 DijkstraCache example:
1085 <AS id="AS_2" routing="DijkstraCache">
1086 <host id="AS_2_host1" speed="1000000000"/>
1088 (platform unchanged compared to upper example)
1091 #### Example platform files ####
1093 This is an automatically generated list of example files that use the DijkstraCache
1094 routing model (the path is given relative to SimGrid's source directory):
1096 Editor's note: At the time of writing, no platform file used this routing model - so
1097 if there are no example files listed here, this is likely to be correct.
1099 \verbinclude example_filelist_routing_dijkstra_cache
1101 \subsubsection pf_routing_model_manual Manually-entered route models
1103 | Name | Description |
1104 | ---------------------------------- | ------------------------------------------------------------------------------ |
1105 | \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. |
1107 \anchor pf_routing_model_full
1112 <AS id="AS0" routing="Full">
1113 <host id="host1" speed="1000000000"/>
1114 <host id="host2" speed="1000000000"/>
1115 <link id="link1" bandwidth="125000000" latency="0.000100"/>
1116 <route src="host1" dst="host2"><link_ctn id="link1"/></route>
1120 #### Example platform files ####
1122 This is an automatically generated list of example files that use the Full
1123 routing model (the path is given relative to SimGrid's source directory):
1125 \verbinclude example_filelist_routing_full
1127 \subsubsection pf_routing_model_simple Simple/fast models
1129 | Name | Description |
1130 | ---------------------------------------- | ------------------------------------------------------------------------------ |
1131 | \ref pf_routing_model_cluster "Cluster" | This is specific to the \ref pf_cluster "<cluster/>" tag and should not be used by the user, as several assumptions are made. |
1132 | \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. |
1133 | \ref pf_routing_model_vivaldi "Vivaldi" | Perfect when you want to use coordinates. Also see the corresponding \ref pf_P2P_tags "P2P section" below. |
1135 \anchor pf_routing_model_cluster
1139 In this mode, the \ref pf_cabinet "<cabinet/>" tag is available.
1141 #### Example platform files ####
1143 This is an automatically generated list of example files that use the Cluster
1144 routing model (the path is given relative to SimGrid's source directory):
1146 \verbinclude example_filelist_routing_cluster
1148 \anchor pf_routing_model_none
1151 This model does exactly what it's name advertises: Nothing. There is no routing
1152 available within this model and if you try to communicate within the AS that
1153 uses this model, SimGrid will fail unless you have explicitly activated the
1154 \ref options_model_select_network_constant "Constant Network Model" (this model charges
1155 the same for every single communication). It should
1156 be noted, however, that you can still attach an \ref pf_tag_asroute "ASroute",
1157 as is demonstrated in the example below:
1159 \verbinclude platforms/cluster_and_one_host.xml
1161 #### Example platform files ####
1163 This is an automatically generated list of example files that use the None
1164 routing model (the path is given relative to SimGrid's source directory):
1166 \verbinclude example_filelist_routing_none
1169 \anchor pf_routing_model_vivaldi
1172 For more information on how to use the [Vivaldi Coordinates](https://en.wikipedia.org/wiki/Vivaldi_coordinates),
1173 see also Section \ref pf_P2P_tags "P2P tags".
1175 Note that it is possible to combine the Vivaldi routing model with other routing models;
1176 an example can be found in the file \c examples/platforms/cloud.xml. This
1177 examples models a NetZone using Vivaldi that contains other NetZones that use different
1180 #### Example platform files ####
1182 This is an automatically generated list of example files that use the None
1183 routing model (the path is given relative to SimGrid's source directory):
1185 \verbinclude example_filelist_routing_vivaldi
1188 \subsection ps_dec Defining routes
1190 There are currently four different ways to define routes:
1192 | Name | Description |
1193 | ------------------------------------------------- | ----------------------------------------------------------------------------------- |
1194 | \ref pf_tag_route "route" | Used to define route between host/router |
1195 | \ref pf_tag_zoneroute "zoneRoute" | Used to define route between different zones |
1196 | \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. |
1197 | \ref pf_tag_bypassasroute "bypassZoneRoute" | Used in the same way as bypassRoute, but for zones |
1199 Basically all those tags will contain an (ordered) list of references
1200 to link that compose the route you want to define.
1202 Consider the example below:
1205 <route src="Alice" dst="Bob">
1206 <link_ctn id="link1"/>
1207 <link_ctn id="link2"/>
1208 <link_ctn id="link3"/>
1212 The route here from host Alice to Bob will be first link1, then link2,
1213 and finally link3. What about the reverse route? \ref pf_tag_route "Route" and
1214 \ref pf_tag_zoneroute "ASroute" have an optional attribute \c symmetrical, that can
1215 be either \c YES or \c NO. \c YES means that the reverse route is the same
1216 route in the inverse order, and is set to \c YES by default. Note that
1217 this is not the case for bypass*Route, as it is more probable that you
1218 want to bypass only one default route.
1220 For an \ref pf_tag_zoneroute "ASroute", things are just slightly more complicated, as you have
1221 to give the id of the gateway which is inside the AS you want to access ...
1222 So it looks like this:
1225 <ASroute src="AS1" dst="AS2"
1226 gw_src="router1" gw_dst="router2">
1227 <link_ctn id="link1"/>
1231 gw == gateway, so when any message are trying to go from AS1 to AS2,
1232 it means that it must pass through router1 to get out of the AS, then
1233 pass through link1, and get into AS2 by being received by router2.
1234 router1 must belong to AS1 and router2 must belong to AS2.
1236 \subsubsection pf_tag_linkctn <link_ctn>
1238 This entity has only one purpose: Refer to an already existing
1239 \ref pf_tag_link "<link/>" when defining a route, i.e., it
1240 can only occur as a child of \ref pf_tag_route "<route/>"
1242 | Attribute name | Mandatory | Values | Description |
1243 | --------------- | --------- | ------ | ----------- |
1244 | id | yes | String | The identifier of the link that should be added to the route. |
1245 | direction | maybe | UP\|DOWN | If the link referenced by \c id has been declared as \ref pf_sharing_policy_fullduplex "FULLDUPLEX", this indicates which direction the route traverses through this link: UP or DOWN. If you don't use FULLDUPLEX, do not use this attribute or SimGrid will not find the right link.
1247 #### Example Files ####
1249 This is an automatically generated list of example files that use the \c <link_ctn/>
1250 entity (the path is given relative to SimGrid's source directory):
1252 \verbinclude example_filelist_xmltag_linkctn
1254 \subsubsection pf_tag_zoneroute <zoneRoute>
1256 The purpose of this entity is to define a route between two ASes.
1257 This is mainly useful when you're in the \ref pf_routing_model_full "Full routing model".
1259 #### Attributes ####
1261 | Attribute name | Mandatory | Values | Description |
1262 | --------------- | --------- | ------ | ----------- |
1263 | src | yes | String | The identifier of the source AS |
1264 | dst | yes | String | See the \c src attribute |
1265 | gw_src | yes | String | The gateway that will be used within the src AS; this can be any \ref pf_tag_host "Host" or \ref pf_router "Router" defined within the src AS. |
1266 | gw_dst | yes | String | Same as \c gw_src, but with the dst AS instead. |
1267 | symmetrical | no | YES\|NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1272 <AS id="AS0" routing="Full">
1273 <cluster id="my_cluster_1" prefix="c-" suffix=".me"
1274 radical="0-149" speed="1000000000" bw="125000000" lat="5E-5"
1275 bb_bw="2250000000" bb_lat="5E-4"/>
1277 <cluster id="my_cluster_2" prefix="c-" suffix=".me"
1278 radical="150-299" speed="1000000000" bw="125000000" lat="5E-5"
1279 bb_bw="2250000000" bb_lat="5E-4"/>
1281 <link id="backbone" bandwidth="1250000000" latency="5E-4"/>
1283 <ASroute src="my_cluster_1" dst="my_cluster_2"
1284 gw_src="c-my_cluster_1_router.me"
1285 gw_dst="c-my_cluster_2_router.me">
1286 <link_ctn id="backbone"/>
1288 <ASroute src="my_cluster_2" dst="my_cluster_1"
1289 gw_src="c-my_cluster_2_router.me"
1290 gw_dst="c-my_cluster_1_router.me">
1291 <link_ctn id="backbone"/>
1296 \subsubsection pf_tag_route <route>
1298 The principle is the same as for
1299 \ref pf_tag_zoneroute "ASroute": The route contains a list of links that
1300 provide a path from \c src to \c dst. Here, \c src and \c dst can both be either a
1301 \ref pf_tag_host "host" or \ref pf_router "router". This is mostly useful for the
1302 \ref pf_routing_model_full "Full routing model" as well as for the
1303 \ref pf_routing_model_shortest_path "shortest-paths" based models (as they require
1304 topological information).
1307 | Attribute name | Mandatory | Values | Description |
1308 | --------------- | --------- | ---------------------- | ----------- |
1309 | src | yes | String | The value given to the source's "id" attribute |
1310 | dst | yes | String | The value given to the destination's "id" attribute. |
1311 | symmetrical | no | YES\| NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1316 A route in the \ref pf_routing_model_full "Full routing model" could look like this:
1318 <route src="Tremblay" dst="Bourassa">
1319 <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"/>
1323 A route in the \ref pf_routing_model_shortest_path "Shortest-Path routing model" could look like this:
1325 <route src="Tremblay" dst="Bourassa">
1330 You must only have one link in your routes when you're using them to provide
1331 topological information, as the routes here are simply the edges of the
1332 (network-)graph and the employed algorithms need to know which edge connects
1333 which pair of entities.
1335 \subsubsection pf_tag_bypassasroute bypassASroute
1337 As said before, once you choose
1338 a model, it (most likely; the constant network model, for example, doesn't) calculates routes for you. But maybe you want to
1339 define some of your routes, which will be specific. You may also want
1340 to bypass some routes defined in lower level AS at an upper stage:
1341 <b>bypassASroute</b> is the tag you're looking for. It allows to
1342 bypass routes defined between already defined between AS (if you want
1343 to bypass route for a specific host, you should just use byPassRoute).
1344 The principle is the same as ASroute : <b>bypassASroute</b> contains
1345 list of links that are in the path between src and dst.
1347 #### Attributes ####
1349 | Attribute name | Mandatory | Values | Description |
1350 | --------------- | --------- | ---------------------- | ----------- |
1351 | src | yes | String | The value given to the source AS's "id" attribute |
1352 | dst | yes | String | The value given to the destination AS's "id" attribute. |
1353 | gw_src | yes | String | The value given to the source gateway's "id" attribute; this can be any host or router within the src AS |
1354 | gw_dst | yes | String | The value given to the destination gateway's "id" attribute; this can be any host or router within the dst AS|
1355 | symmetrical | no | YES\| NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1360 <bypassASRoute src="my_cluster_1" dst="my_cluster_2"
1361 gw_src="my_cluster_1_router"
1362 gw_dst="my_cluster_2_router">
1363 <link_ctn id="link_tmp"/>
1367 This example shows that link \c link_tmp (definition not displayed here) directly
1368 connects the router \c my_cluster_1_router in the source cluster to the router
1369 \c my_cluster_2_router in the destination router. Additionally, as the \c symmetrical
1370 attribute was not given, this route is presumed to be symmetrical.
1372 \subsubsection pf_tag_bypassroute bypassRoute
1374 As said before, once you choose
1375 a model, it (most likely; the constant network model, for example, doesn't) calculates routes for you. But maybe you want to
1376 define some of your routes, which will be specific. You may also want
1377 to bypass some routes defined in lower level AS at an upper stage :
1378 <b>bypassRoute</b> is the tag you're looking for. It allows to bypass
1379 routes defined between <b>host/router</b>. The principle is the same
1380 as route : <b>bypassRoute</b> contains list of links references of
1381 links that are in the path between src and dst.
1383 #### Attributes ####
1385 | Attribute name | Mandatory | Values | Description |
1386 | --------------- | --------- | ---------------------- | ----------- |
1387 | src | yes | String | The value given to the source AS's "id" attribute |
1388 | dst | yes | String | The value given to the destination AS's "id" attribute. |
1389 | symmetrical | no | YES \| NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1394 <bypassRoute src="host_1" dst="host_2">
1395 <link_ctn id="link_tmp"/>
1399 This example shows that link \c link_tmp (definition not displayed here) directly
1400 connects host \c host_1 to host \c host_2. Additionally, as the \c symmetrical
1401 attribute was not given, this route is presumed to be symmetrical.
1403 \subsection pb_baroex Basic Routing Example
1405 Let's say you have an AS named AS_Big that contains two other AS, AS_1
1406 and AS_2. If you want to make a host (h1) from AS_1 with another one
1407 (h2) from AS_2 then you'll have to proceed as follows:
1408 \li First, you have to ensure that a route is defined from h1 to the
1409 AS_1's exit gateway and from h2 to AS_2's exit gateway.
1410 \li Then, you'll have to define a route between AS_1 to AS_2. As those
1411 AS are both resources belonging to AS_Big, then it has to be done
1412 at AS_big level. To define such a route, you have to give the
1413 source AS (AS_1), the destination AS (AS_2), and their respective
1414 gateway (as the route is effectively defined between those two
1415 entry/exit points). Elements of this route can only be elements
1416 belonging to AS_Big, so links and routers in this route should be
1417 defined inside AS_Big. If you choose some shortest-path model,
1418 this route will be computed automatically.
1420 As said before, there are mainly 2 tags for routing :
1421 \li <b>ASroute</b>: to define routes between two <b>AS</b>
1422 \li <b>route</b>: to define routes between two <b>host/router</b>
1424 As we are dealing with routes between AS, it means that those we'll
1425 have some definition at AS_Big level. Let consider AS_1 contains 1
1426 host, 1 link and one router and AS_2 3 hosts, 4 links and one router.
1427 There will be a central router, and a cross-like topology. At the end
1428 of the crosses arms, you'll find the 3 hosts and the router that will
1429 act as a gateway. We have to define routes inside those two AS. Let
1430 say that AS_1 contains full routes, and AS_2 contains some Floyd
1431 routing (as we don't want to bother with defining all routes). As
1432 we're using some shortest path algorithms to route into AS_2, we'll
1433 then have to define some <b>route</b> to gives some topological
1434 information to SimGrid. Here is a file doing it all :
1437 <AS id="AS_Big" routing="Dijkstra">
1438 <AS id="AS_1" routing="Full">
1439 <host id="AS_1_host1" speed="1000000000"/>
1440 <link id="AS_1_link" bandwidth="1250000000" latency="5E-4"/>
1441 <router id="AS_1_gateway"/>
1442 <route src="AS_1_host1" dst="AS_1_gateway">
1443 <link_ctn id="AS_1_link"/>
1446 <AS id="AS_2" routing="Floyd">
1447 <host id="AS_2_host1" speed="1000000000"/>
1448 <host id="AS_2_host2" speed="1000000000"/>
1449 <host id="AS_2_host3" speed="1000000000"/>
1450 <link id="AS_2_link1" bandwidth="1250000000" latency="5E-4"/>
1451 <link id="AS_2_link2" bandwidth="1250000000" latency="5E-4"/>
1452 <link id="AS_2_link3" bandwidth="1250000000" latency="5E-4"/>
1453 <link id="AS_2_link4" bandwidth="1250000000" latency="5E-4"/>
1454 <router id="central_router"/>
1455 <router id="AS_2_gateway"/>
1456 <!-- routes providing topological information -->
1457 <route src="central_router" dst="AS_2_host1"><link_ctn id="AS_2_link1"/></route>
1458 <route src="central_router" dst="AS_2_host2"><link_ctn id="AS_2_link2"/></route>
1459 <route src="central_router" dst="AS_2_host3"><link_ctn id="AS_2_link3"/></route>
1460 <route src="central_router" dst="AS_2_gateway"><link_ctn id="AS_2_link4"/></route>
1462 <link id="backbone" bandwidth="1250000000" latency="5E-4"/>
1464 <ASroute src="AS_1" dst="AS_2"
1465 gw_src="AS_1_gateway"
1466 gw_dst="AS_2_gateway">
1467 <link_ctn id="backbone"/>
1472 \section pf_other_tags Tags not (directly) describing the platform
1474 The following tags can be used inside a \<platform\> tag even if they are not
1475 directly describing the platform:
1476 \li \ref pf_config "config": it allows you to pass some configuration stuff like, for
1477 example, the network model and so on. It follows the
1478 \li \ref pf_include "include": allows you to include another file into the current one.
1480 \subsection pf_config config
1482 The only purpose of this tag is to contain the \c prop tags, as described below.
1483 These tags will then configure the options as described by Section \ref options.
1486 #### Attributes ####
1488 | Attribute name | Mandatory | Values | Description |
1489 | --------------- | --------- | ---------------------- | ----------- |
1490 | id | yes | String | The identifier of the config tag when referring to id; this is basically useless, though. |
1492 #### Possible children ####
1494 Tag name | Description | Documentation
1495 ------------ | ----------- | -------------
1496 \<prop/\> | The prop tag allows you to define different configuration options following the attribute/value schema. See the \ref options page. | N/A
1501 <?xml version='1.0'?>
1502 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid.dtd">
1503 <platform version="4">
1504 <config id="General">
1505 <prop id="maxmin/precision" value="0.000010"></prop>
1506 <prop id="cpu/optim" value="TI"></prop>
1507 <prop id="host/model" value="compound"></prop>
1508 <prop id="network/model" value="SMPI"></prop>
1509 <prop id="path" value="~/"></prop>
1510 <prop id="smpi/bw-factor" value="65472:0.940694;15424:0.697866;9376:0.58729"></prop>
1513 <AS id="AS0" routing="Full">
1517 \subsection pf_include include
1519 Even if it can be used in other contexts, this tag was originally created
1520 to be used with \ref pf_trace. The idea was to have a file describing the
1521 platform, and another file attaching traces of a given period to the platform.
1523 The drawback is that the file chuncks that will be included do not
1524 constitute valid XML files. This may explain why this feature was never really
1525 used in practice (as far as we know). Other mechanisms, such as the ability to load
1526 several platform files one after the other, could be considered in the future.
1528 In the meanwhile, the \c include tag allows you to import other platforms into your
1529 local file. This is done with the intention to help people
1530 combine their different AS and provide new platforms. Those files
1531 should contain XML that consists of
1532 \ref pf_include "include", \ref pf_cluster "cluster", \ref pf_peer "peer", \ref pf_As "AS", \ref pf_trace "trace", \ref pf_trace "tags".
1534 Do not forget to close the tag to make it work, or you will end up with an invalid XML file.
1536 #### Attributes ####
1538 | Attribute name | Mandatory | Values | Description |
1539 | --------------- | --------- | ---------------------- | ----------- |
1540 | file | yes | String | Filename of the path you want to include with either relative or absolute path. |
1545 The following example includes two files, clusterA.xml and clusterB.xml and
1546 combines them two one platform file; all hosts, routers etc. defined in
1547 each of them will then be usable.
1550 <?xml version='1.0'?>
1551 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid/simgrid.dtd">
1552 <platform version="4">
1553 <AS id="main" routing="Full">
1554 <include file="clusterA.xml"></include>
1555 <include file="clusterB.xml"></include>
1560 \subsection pf_trace trace and trace_connect
1562 Both tags are an alternate way to pass files containing information on
1563 availability, state etc. to an entity. (See also, for instance, Section \ref
1564 pf_host_churn "Churn", as described for the host entity.) Instead of referring
1565 to the file directly in the host, link, or cluster tag, you proceed by defining
1566 a trace with an id corresponding to a file, later a host/link/cluster, and
1567 finally using trace_connect you say that the file trace must be used by the
1574 <AS id="AS0" routing="Full">
1575 <host id="bob" speed="1000000000"/>
1577 <trace id="myTrace" file="bob.trace" periodicity="1.0"/>
1578 <trace_connect trace="myTrace" element="bob" kind="POWER"/>
1582 The order here is important. \c trace_connect must come
1583 after the elements \c trace and \c host, as both the host
1584 and the trace definition must be known when \c trace_connect
1585 is parsed; the order of \c trace and \c host is arbitrary.
1588 #### \c trace attributes ####
1591 | Attribute name | Mandatory | Values | Description |
1592 | --------------- | --------- | ---------------------- | ----------- |
1593 | id | yes | String | Identifier of this trace; this is the name you pass on to \c trace_connect. |
1594 | 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. |
1595 | trace_periodicity | yes | String | This is the same as for \ref pf_tag_host "hosts" (see there for details) |
1597 Here is an example of trace when no file name is provided:
1600 <trace id="myTrace" periodicity="1.0">
1607 #### \c trace_connect attributes ####
1609 | Attribute name | Mandatory | Values | Description |
1610 | --------------- | --------- | ---------------------- | ----------- |
1611 | kind | no | HOST_AVAIL\|POWER\|<br/>LINK_AVAIL\|BANDWIDTH\|LATENCY (Default: HOST_AVAIL) | Describes the kind of trace. |
1612 | trace | yes | String | Identifier of the referenced trace (specified of the trace's \c id attribute) |
1613 | element | yes | String | The identifier of the referenced entity as given by its \c id attribute |
1615 \section pf_hints Hints, tips and frequently requested features
1617 Now you should know at least the syntax and be able to create a
1618 platform by your own. However, after having ourselves wrote some platforms, there
1619 are some best practices you should pay attention to in order to
1620 produce good platform and some choices you can make in order to have
1621 faster simulations. Here's some hints and tips, then.
1623 @subsection Finding the platform example that you need
1625 Most platform files that we ship are in the @c examples/platforms
1626 folder. The good old @c grep tool can find the examples you need when
1627 wondering on a specific XML tag. Here is an example session searching
1628 for @ref pf_trace "trace_connect":
1631 % cd examples/platforms
1632 % grep -R -i -n --include="*.xml" "trace_connect" .
1633 ./two_hosts_platform_with_availability_included.xml:26:<trace_connect kind="SPEED" trace="A" element="Cpu A"/>
1634 ./two_hosts_platform_with_availability_included.xml:27:<trace_connect kind="HOST_AVAIL" trace="A_failure" element="Cpu A"/>
1635 ./two_hosts_platform_with_availability_included.xml:28:<trace_connect kind="SPEED" trace="B" element="Cpu B"/>
1636 ./two_hosts.xml:17: <trace_connect trace="Tremblay_power" element="Tremblay" kind="SPEED"/>
1639 \subsection pf_as_h AS Hierarchy
1640 The AS design allows SimGrid to go fast, because computing route is
1641 done only for the set of resources defined in this AS. If you're using
1642 only a big AS containing all resource with no AS into it and you're
1643 using Full model, then ... you'll loose all interest into it. On the
1644 other hand, designing a binary tree of AS with, at the lower level,
1645 only one host, then you'll also loose all the good AS hierarchy can
1646 give you. Remind you should always be "reasonable" in your platform
1647 definition when choosing the hierarchy. A good choice if you try to
1648 describe a real life platform is to follow the AS described in
1649 reality, since this kind of trade-off works well for real life
1652 \subsection pf_exit_as Exit AS: why and how
1653 Users that have looked at some of our platforms may have notice a
1654 non-intuitive schema ... Something like that :
1658 <AS id="AS_4" routing="Full">
1659 <AS id="exitAS_4" routing="Full">
1660 <router id="router_4"/>
1662 <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"/>
1663 <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"/>
1664 <link id="4_1" bandwidth="2250000000" latency="5E-5"/>
1665 <link id="4_2" bandwidth="2250000000" latency="5E-5"/>
1666 <link id="bb_4" bandwidth="2250000000" latency="5E-4"/>
1667 <ASroute src="cl_4_1"
1669 gw_src="c_4_1-cl_4_1_router"
1670 gw_dst="c_4_2-cl_4_2_router">
1671 <link_ctn id="4_1"/>
1672 <link_ctn id="bb_4"/>
1673 <link_ctn id="4_2"/>
1675 <ASroute src="cl_4_1"
1677 gw_src="c_4_1-cl_4_1_router"
1679 <link_ctn id="4_1"/>
1680 <link_ctn id="bb_4"/>
1682 <ASroute src="cl_4_2"
1684 gw_src="c_4_2-cl_4_2_router"
1686 <link_ctn id="4_2"/>
1687 <link_ctn id="bb_4"/>
1692 In the AS_4, you have an exitAS_4 defined, containing only one router,
1693 and routes defined to that AS from all other AS (as cluster is only a
1694 shortcut for an AS, see cluster description for details). If there was
1695 an upper AS, it would define routes to and from AS_4 with the gateway
1696 router_4. It's just because, as we did not allowed (for performances
1697 issues) to have routes from an AS to a single host/router, you have to
1698 enclose your gateway, when you have AS included in your AS, within an
1699 AS to define routes to it.
1701 \subsection pf_P2P_tags P2P or how to use coordinates
1702 SimGrid allows you to use some coordinated-based system, like vivaldi,
1703 to describe a platform. The main concept is that you have some peers
1704 that are located somewhere: this is the function of the
1705 <b>coordinates</b> of the \<peer\> or \<host\> tag. There's nothing
1706 complicated in using it, here is an example:
1709 <?xml version='1.0'?>
1710 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid.dtd">
1711 <platform version="4">
1713 <AS id="AS0" routing="Vivaldi">
1714 <host id="100030591" coordinates="25.5 9.4 1.4" speed="1.5Gf" />
1715 <host id="100036570" coordinates="-12.7 -9.9 2.1" speed="7.3Gf" />
1717 <host id="100429957" coordinates="17.5 6.7 18.8" speed="8.3Gf" />
1722 Coordinates are then used to calculate latency (in microseconds)
1723 between two hosts by calculating the distance between the two hosts
1724 coordinates with the following formula: distance( (x1, y1, z1), (x2,
1725 y2, z2) ) = euclidian( (x1,y1), (x2,y2) ) + abs(z1) + abs(z2)
1727 In other words, we take the euclidian distance on the two first
1728 dimensions, and then add the absolute values found on the third
1729 dimension. This may seem strange, but it was found to allow better
1730 approximations of the latency matrices (see the paper describing
1733 Note that the previous example defines a routing directly between hosts but it could be also used to define a routing between AS.
1734 That is for example what is commonly done when using peers (see Section \ref pf_peer).
1736 <?xml version='1.0'?>
1737 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid.dtd">
1738 <platform version="4">
1740 <AS id="AS0" routing="Vivaldi">
1741 <peer id="peer-0" coordinates="173.0 96.8 0.1" speed="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us"/>
1742 <peer id="peer-1" coordinates="247.0 57.3 0.6" speed="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us" />
1743 <peer id="peer-2" coordinates="243.4 58.8 1.4" speed="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us" />
1747 In such a case though, we connect the AS created by the <b>peer</b> tag with the Vivaldi routing mechanism.
1748 This means that to route between AS1 and AS2, it will use the coordinates of router_AS1 and router_AS2.
1749 This is currently a convention and we may offer to change this convention in the DTD later if needed.
1750 You may have noted that conveniently, a peer named FOO defines an AS named FOO and a router named router_FOO, which is why it works seamlessly with the <b>peer</b> tag.
1753 \subsection pf_routing_howto_choose_wisely Choosing wisely the routing model to use
1756 Choosing wisely the routing model to use can significantly fasten your
1757 simulation/save your time when writing the platform/save tremendous
1758 disk space. Here is the list of available model and their
1759 characteristics (lookup : time to resolve a route):
1761 \li <b>Full</b>: Full routing data (fast, large memory requirements,
1763 \li <b>Floyd</b>: Floyd routing data (slow initialization, fast
1764 lookup, lesser memory requirements, shortest path routing only).
1765 Calculates all routes at once at the beginning.
1766 \li <b>Dijkstra</b>: Dijkstra routing data (fast initialization, slow
1767 lookup, small memory requirements, shortest path routing only).
1768 Calculates a route when necessary.
1769 \li <b>DijkstraCache</b>: Dijkstra routing data (fast initialization,
1770 fast lookup, small memory requirements, shortest path routing
1771 only). Same as Dijkstra, except it handles a cache for latest used
1773 \li <b>None</b>: No routing (usable with Constant network only).
1774 Defines that there is no routes, so if you try to determine a
1775 route without constant network within this AS, SimGrid will raise
1777 \li <b>Vivaldi</b>: Vivaldi routing, so when you want to use coordinates
1778 \li <b>Cluster</b>: Cluster routing, specific to cluster tag, should
1781 \subsection pf_switch I want to describe a switch but there is no switch tag!
1783 Actually we did not include switch tag. But when you're trying to
1784 simulate a switch, assuming
1785 fluid bandwidth models are used (which SimGrid uses by default unless
1786 ns-3 or constant network models are activated), the limiting factor is
1787 switch backplane bandwidth. So, essentially, at least from
1788 the simulation perspective, a switch is similar to a
1789 link: some device that is traversed by flows and with some latency and
1790 so,e maximum bandwidth. Thus, you can simply simulate a switch as a
1792 can be connected to this "switch", which is then included in routes just
1796 \subsection pf_multicabinets I want to describe multi-cabinets clusters!
1798 You have several possibilities, as usual when modeling things. If your
1799 cabinets are homogeneous and the intercabinet network negligible for
1800 your study, you should just create a larger cluster with all hosts at
1803 In the rare case where your hosts are not homogeneous between the
1804 cabinets, you can create your cluster completely manually. For that,
1805 create an As using the Cluster routing, and then use one
1806 <cabinet> for each cabinet. This cabinet tag can only be used an
1807 As using the Cluster routing schema, and creating
1809 Be warned that creating a cluster manually from the XML with
1810 <cabinet>, <backbone> and friends is rather tedious. The
1811 easiest way to retrieve some control of your model without diving into
1812 the <cluster> internals is certainly to create one separate
1813 <cluster> per cabinet and interconnect them together. This is
1814 what we did in the G5K example platform for the Graphen cluster.
1816 \subsection pf_platform_multipath I want to express multipath routing in platform files!
1818 It is unfortunately impossible to express the fact that there is more
1819 than one routing path between two given hosts. Let's consider the
1820 following platform file:
1823 <route src="A" dst="B">
1826 <route src="B" dst="C">
1829 <route src="A" dst="C">
1834 Although it is perfectly valid, it does not mean that data traveling
1835 from A to C can either go directly (using link 3) or through B (using
1836 links 1 and 2). It simply means that the routing on the graph is not
1837 trivial, and that data do not following the shortest path in number of
1838 hops on this graph. Another way to say it is that there is no implicit
1839 in these routing descriptions. The system will only use the routes you
1840 declare (such as <route src="A" dst="C"><link_ctn
1841 id="3"/></route>), without trying to build new routes by aggregating
1844 You are also free to declare platform where the routing is not
1845 symmetrical. For example, add the following to the previous file:
1848 <route src="C" dst="A">
1854 This makes sure that data from C to A go through B where data from A
1855 to C go directly. Don't worry about realism of such settings since
1856 we've seen ways more weird situation in real settings (in fact, that's
1857 the realism of very regular platforms which is questionable, but
1858 that's another story).