1 /*! \page platform Describing the virtual platform
5 In order to run any simulation, SimGrid must be provided with three things:
6 something to run (i.e., your code), a description of the platform on which you want to simulate your application, and
7 information about the deployment of the application: Which process should be executed onto which processor/core?
9 For the last two items, there are essentially three possible ways you can provide
10 this information as an input:
11 \li You can program, if you're using MSG, some of the platform and
12 deployment functions. If you choose to follow this approach, check the dedicated documentation
13 (\ref msg_simulation).
14 \li You can use two XML files: one for the platform description and the other for the deployment.
15 \li You can program the description of your platform in Lua format.
17 For more information on SimGrid's deployment features, please refer to the \ref deployment section.
19 The platform description may be intricate. This documentation is all
20 about how to write this file. You should read about the
21 @ref routing_basics "routing basic concepts" before proceeding. This page
22 first contain a reference guide of the XML. Finally, it gives some hints and tips on how to write a better
25 \section pf_overview Some words about XML and DTD
27 We opted for XML not only because it is extensible but also because many tools (and plugins for existing tools) are
28 available that facilitate editing and validating XML files. Furthermore, libraries that parse XML are often already
29 available and very well tested.
31 The XML checking is done based on the [simgrid.dtd](http://simgrid.gforge.inria.fr/simgrid/simgrid.dtd) Document Type
32 Definition (DTD) file.
34 If you read the DTD, you should notice the following:
35 \li The platform tag has a version attribute. The current version is <b>4</b>. This attribute might be used in the
36 provide backward compatibility.
37 \li The DTD contains definitions for both the platform description and deployment files used by SimGrid.
39 \section pf_netzones Defining a netzone
41 Here is a simplistic example, describing a netzone using the Full
42 routing. Other supported values for the routing attribute can be
43 found below, Section \ref pf_raf.
47 <AS id="netzone0" routing="Full">
50 There is also the ``<route>`` tag; this tag takes two attributes,
51 ``src`` (source) and ``dst`` (destination). Both source and
52 destination must be valid identifiers for routers (these will be
53 introduced later). Contained by the ``<route>`` are network links;
54 these links must be used in order to communicate from the source to
55 the destination specified in the tag. Hence, a route merely describes
56 how to reach a router from another router.
59 More information and (code-)examples can be found in Section \ref pf_rm.
61 A netzone can also contain itself one or more netzone; this allows you to model
62 the hierarchy of your platform.
64 ### Within each AS, the following types of resources exist:
66 %Resource | Documented in Section | Description
67 --------------- | --------------------- | -----------
68 AS | | Every Autonomous System (AS) may contain one or more AS.
69 host | \ref pf_host | This entity carries out the actual computation. For this reason, it contains processors (with potentially multiple cores).
70 router | \ref pf_router | In SimGrid, routers are used to provide helpful information to routing algorithms. Routers may also act as gateways, connecting several autonomous systems with each other.
71 link | \ref pf_link | In SimGrid, (network)links define a connection between two or potentially even more resources. Every link has a bandwidth and a latency and may potentially experience congestion.
72 cluster | \ref pf_cluster | In SimGrid, clusters were introduced to model large and homogenous environments. They are not really a resource by themselves - technically, they are only a shortcut, as they will internally set up all the hosts, network and routing for you, i.e., using this resource, one can easily setup thousands of hosts and links in a few lines of code. Each cluster is itself an AS.
74 As it is desirable to interconnect these resources, a routing has to
75 be defined. The AS is supposed to be Autonomous, hence this has to be
76 done at the AS level. The AS handles two different types of entities
77 (<b>host/router</b> and <b>AS</b>). However, the user is responsible
78 to define routes between those resources, otherwise entities will be
79 unconnected and therefore unreachable from other entities. Although
80 several routing algorithms are built into SimGrid (see \ref pf_rm),
81 you might encounter a case where you want to define routes manually
82 (for instance, due to specific requirements of your platform).
84 There are three tags to use:
85 \li <b>ASroute</b>: to define routes between two <b>AS</b>
86 \li <b>route</b>: to define routes between two <b>host/router</b>
87 \li <b>bypassRoute</b>: to define routes between two <b>AS</b> that
88 will bypass default routing (as specified by the ``routing`` attribute
89 supplied to ``<AS>``, see above).
91 Here is an illustration of these concepts:
93 
94 Circles represent processing units and squares represent network routers. Bold
95 lines represent communication links. AS2 models the core of a national
96 network interconnecting a small flat cluster (AS4) and a larger
97 hierarchical cluster (AS5), a subset of a LAN (AS6), and a set of peers
98 scattered around the world (AS7).
100 \section pf_pftags Resource description
102 \subsection pf_As Platform: The <AS> tag
104 For historical reasons, the XML files use the expression AS for
105 NetZones. Netzones are very important because they group other resources (such
106 as routers/hosts) together (in fact, any such resource must be
107 contained in a NetZone).
109 Available attributes :
111 Attribute name | Mandatory | Values | Description
112 --------------- | --------- | ------ | -----------
113 id | yes | String | The identifier of an AS; facilitates referring to this AS. ID must be unique.
114 routing | yes | Full\| Floyd\| Dijkstra\| DijkstraCache\| None\| Vivaldi\| Cluster | See Section \ref pf_rm for details.
119 <AS id="AS0" routing="Full">
120 <host id="host1" speed="1000000000"/>
121 <host id="host2" speed="1000000000"/>
122 <link id="link1" bandwidth="125000000" latency="0.000100"/>
123 <route src="host1" dst="host2"><link_ctn id="link1"/></route>
127 In this example, AS0 contains two hosts (host1 and host2). The route
128 between the hosts goes through link1.
130 \subsection pf_Cr Computing resources: hosts, clusters and peers.
132 \subsubsection pf_host <host>
134 A <b>host</b> represents a computer/node card. Every host is able to execute
135 code and it can send and receive data to/from other hosts. Most importantly,
136 a host can contain more than 1 core.
140 Attribute name | Mandatory | Values | Description
141 --------------- | --------- | ------ | -----------
142 id | yes | String | The identifier of the host. facilitates referring to this AS.
143 speed | yes | double (must be > 0.0) | Computational power of every core of this host in FLOPS. Must be larger than 0.0.
144 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.
145 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.
146 state_file | no | string | Same mechanism as availability_file.<br /> <b>Note:</b> The filename must be specified with your system's format.
147 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.
148 pstate | no | double (Default: 0.0) | FIXME: Not yet documented.
150 ### Possible children: ###
152 Tag name | Description | Documentation
153 ------------ | ----------- | -------------
154 \<mount/\> | Defines mounting points between some storage resource and the host. | \ref pf_storage_entity_mount
155 \<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
160 <host id="host1" speed="1000000000"/>
161 <host id="host2" speed="1000000000">
162 <prop id="color" value="blue"/>
163 <prop id="rendershape" value="square"/>
168 \anchor pf_host_dynamism
169 ### Expressing dynamism ###
171 SimGrid provides mechanisms to change a hosts' availability over
172 time, using the ``availability_file`` attribute to the ``\<host\>`` tag
173 and a separate text file whose syntax is exemplified below.
175 #### Adding a trace file ####
178 <platform version="4">
179 <host id="bob" speed="500Gf" availability_file="bob.trace" />
183 #### Example of "bob.trace" file ####
192 Let us begin to explain this example by looking at line 2. (Line 1 will become clear soon).
193 The first column describes points in time, in this case, time 0. The second column
194 describes the relative amount of power this host is able to deliver (relative
195 to the maximum performance specified in the ``\<host\>`` tag). (Clearly, the
196 second column needs to contain values that are not smaller than 0 and not larger than 1).
197 In this example, our host will deliver 500 Mflop/s at time 0, as 500 Mflop/s is the
198 maximum performance of this host. At time 11.0, it will
199 deliver half of its maximum performance, i.e., 250 Mflop/s until time 20.0 when it will
200 will start delivering 80\% of its power. In this example, this amounts to 400 Mflop/s.
202 Since the periodicity in line 1 was set to be 1.0, i.e., 1 timestep, this host will
203 continue to provide 500 Mflop/s from time 21. From time 32 it will provide 250 MFlop/s and so on.
205 ### Changing initial state ###
207 It is also possible to specify whether the host is up or down by setting the
208 ``state`` attribute to either <b>ON</b> (default value) or <b>OFF</b>.
210 #### Example: Expliciting the default value "ON" ####
213 <platform version="4">
214 <host id="bob" speed="500Gf" state="ON" />
218 If you want this host to be unavailable, simply substitute ON with OFF.
220 \anchor pf_host_churn
221 ### Expressing churn ###
223 To express the fact that a host can change state over time (as in P2P
224 systems, for instance), it is possible to use a file describing the time
225 at which the host is turned on or off. An example of the content
226 of such a file is presented below.
228 #### Adding a state file ####
231 <platform version="4">
232 <host id="bob" power="500Gf" state_file="bob.fail" />
236 #### Example of "bob.fail" file ####
244 A zero value means <b>down</b> (i.e., OFF) while a positive one means <b>up and
245 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
246 turned off and at time 2.0, it is turned on again until time 12 (2 plus the
247 periodicity 10). It will be turned off again at time 13.0 until time 23.0, and
251 \subsubsection pf_cluster <cluster>
253 ``<cluster />`` represents a machine-cluster. It is most commonly used
254 when one wants to define many hosts and a network quickly. Technically,
255 ``cluster`` is a meta-tag: <b>from the inner SimGrid point of
256 view, a cluster is an AS where some optimized routing is defined</b>.
257 The default inner organization of the cluster is as follow:
263 ____________|__________|_____________ backbone
265 l0| l1| l2| l97| l96 | | l99
271 Here, a set of <b>host</b>s is defined. Each of them has a <b>link</b>
272 to a central backbone (backbone is a link itself, as a link can
273 be used to represent a switch, see the switch / link section
274 below for more details about it). A <b>router</b> allows to connect a
275 <b>cluster</b> to the outside world. Internally,
276 SimGrid treats a cluster as an AS containing all hosts: the router is the default
277 gateway for the cluster.
279 There is an alternative organization, which is as follows:
293 The principle is the same, except that there is no backbone. This representation
294 can be obtained easily: just do not set the bb_* attributes.
297 Attribute name | Mandatory | Values | Description
298 --------------- | --------- | ------ | -----------
299 id | yes | string | The identifier of the cluster. Facilitates referring to this cluster.
300 prefix | yes | string | Each node of the cluster has to have a name. This name will be prefixed with this prefix.
301 suffix | yes | string | Each node of the cluster will be suffixed with this suffix
302 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.
303 speed | yes | int | Same as the ``speed`` attribute of the ``\<host\>`` tag.
304 core | no | int (default: 1) | Same as the ``core`` attribute of the ``\<host\>`` tag.
305 bw | yes | int | Bandwidth for the links between nodes and backbone (if any). See the \ref pf_link "link section" for syntax/details.
306 lat | yes | int | Latency for the links between nodes and backbone (if any). See <b>link</b> section for syntax/details.
307 sharing_policy | no | string | Sharing policy for the links between nodes and backbone (if any). See <b>link</b> section for syntax/details.
308 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>).
309 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>).
310 bb_sharing_policy | no | string | Sharing policy for the backbone (if any). See <b>link</b> section for syntax/details.
311 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>
312 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".
313 loopback_lat | no | int | Latency for loopback (if any). See <b>link</b> section for syntax/details. See loopback_bw for more info.
314 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.
315 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".
318 the router name is defined as the resulting String in the following
322 router_name = prefix + clusterId + "_router" + suffix;
326 #### Cluster example ####
328 Consider the following two (and independent) uses of the ``cluster`` tag:
331 <cluster id="my_cluster_1" prefix="" suffix="" radical="0-262144"
332 speed="1e9" bw="125e6" lat="5E-5"/>
334 <cluster id="my_cluster_2" prefix="c-" suffix=".me" radical="0-99"
335 speed="1e9" bw="125e6" lat="5E-5"
336 bb_bw="2.25e9" bb_lat="5E-4"/>
339 The second example creates one router and 100 machines with the following names:
341 c-my_cluster_2_router.me
349 \subsubsection pf_cabinet <cabinet>
352 This tag is only available when the routing mode of the AS
353 is set to ``Cluster``.
355 The ``<cabinet />`` tag is, like the \ref pf_cluster "<cluster>" tag,
356 a meta-tag. This means that it is simply a shortcut for creating a set of (homogenous) hosts and links quickly;
357 unsurprisingly, this tag was introduced to setup cabinets in data centers quickly. Unlike
358 <cluster>, however, the <cabinet> assumes that you create the backbone
359 and routers yourself; see our examples below.
363 Attribute name | Mandatory | Values | Description
364 --------------- | --------- | ------ | -----------
365 id | yes | string | The identifier of the cabinet. Facilitates referring to this cluster.
366 prefix | yes | string | Each node of the cabinet has to have a name. This name will be prefixed with this prefix.
367 suffix | yes | string | Each node of the cabinet will be suffixed with this suffix
368 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.
369 speed | yes | int | Same as the ``speed`` attribute of the \ref pf_host "<host>" tag.
370 bw | yes | int | Bandwidth for the links between nodes and backbone (if any). See the \ref pf_link "link section" for syntax/details.
371 lat | yes | int | Latency for the links between nodes and backbone (if any). See the \ref pf_link "link section" for syntax/details.
374 Please note that as of now, it is impossible to change attributes such as,
375 amount of cores (always set to 1), the initial state of hosts/links
376 (always set to ON), the sharing policy of the links (always set to \ref pf_sharing_policy_fullduplex "FULLDUPLEX").
380 The following example was taken from ``examples/platforms/meta_cluster.xml`` and
381 shows how to use the cabinet tag.
384 <AS id="my_cluster1" routing="Cluster">
385 <cabinet id="cabinet1" prefix="host-" suffix=".cluster1"
386 speed="1Gf" bw="125MBps" lat="100us" radical="1-10"/>
387 <cabinet id="cabinet2" prefix="host-" suffix=".cluster1"
388 speed="1Gf" bw="125MBps" lat="100us" radical="11-20"/>
389 <cabinet id="cabinet3" prefix="host-" suffix=".cluster1"
390 speed="1Gf" bw="125MBps" lat="100us" radical="21-30"/>
392 <backbone id="backbone1" bandwidth="2.25GBps" latency="500us"/>
397 Please note that you must specify the \ref pf_backbone "<backbone>"
398 tag by yourself; this is not done automatically and there are no checks
399 that ensure this backbone was defined.
401 The hosts generated in the above example are named host-1.cluster, host-2.cluster1
405 \subsubsection pf_peer \<peer\> (Vivaldi netzones only)
407 This tag represents a peer, as in Peer-to-Peer (P2P) networks. This
408 can only be used in Vivaldi NetZones. It creates the following
409 resources to the NetZone:
412 \li Two links: One for download and one for upload. This is
413 convenient to use and simulate stuff under the last mile model (e.g., ADSL peers).
414 \li It connects the two links to the host
418 Attribute name | Mandatory | Values | Description
419 --------------- | --------- | ------ | -----------
420 id | yes | string | The identifier of the peer. Facilitates referring to this peer.
421 speed | yes | int | See the description of the ``host`` tag for this attribute
422 bw_in | yes | int | Bandwidth of the private downstream link
423 bw_out | yes | int | Bandwidth of the private upstream link
424 coordinates | no | string | Coordinates of the gateway for this peer. Example value: 12.8 14.4 6.4
425 sharing_policy | no | SHARED\|FULLDUPLEX (default: FULLDUPLEX) | Sharing policy for links. See <b>link</b> description for details.
426 availability_file| no | string | Availability file for the peer. Same as host availability file. See <b>host</b> description for details.
427 state_file | no | string | State file for the peer. Same as host state file. See <b>host</b> description for details.
430 The communication latency between an host A=(xA,yA,zA) and an host
431 B=(xB,yB,zB) is computed as follows:
433 latency = sqrt( (xA-xB)² + (yA-yB)² ) + zA + zB
435 See the documentation of simgrid::kernel::routing::VivaldiZone for
436 details on how the latency is computed from the coordinate, and on the
437 the up and down bandwidth are used.
439 \subsection pf_ne Network equipments: links and routers
441 There are two tags at all times available to represent network entities and
442 several other tags that are available only in certain contexts.
443 1. ``<link>``: Represents a entity that has a limited bandwidth, a
444 latency, and that can be shared according to TCP way to share this
447 The concept of links in SimGrid may not be intuitive, as links are not
448 limited to connecting (exactly) two entities; in fact, you can have more than
449 two equipments connected to it. (In graph theoretical terms: A link in
450 SimGrid is not an edge, but a hyperedge)
452 2. ``<router/>``: Represents an entity that a message can be routed
453 to, but that is unable to execute any code. In SimGrid, routers have also
454 no impact on the performance: Routers do not limit any bandwidth nor
455 do they increase latency. As a matter of fact, routers are (almost) ignored
456 by the simulator when the simulation has begun.
458 3. ``<backbone/>``: This tag is only available when the containing AS is
459 used as a cluster (i.e., mode="Cluster")
462 If you want to represent an entity like a switch, you must use ``<link>`` (see section). Routers are used
463 to run some routing algorithm and determine routes (see Section \ref pf_routing for details).
465 \subsubsection pf_router <router/>
467 As said before, <b>router</b> is used only to give some information
468 for routing algorithms. So, it does not have any attributes except :
472 Attribute name | Mandatory | Values | Description
473 --------------- | --------- | ------ | -----------
474 id | yes | string | The identifier of the router to be used when referring to it.
475 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.
480 <router id="gw_dc1_horizdist"/>
483 \subsubsection pf_link <link/>
485 Network links can represent one-hop network connections. They are
486 characterized by their id and their bandwidth; links can (but may not) be subject
491 Attribute name | Mandatory | Values | Description
492 --------------- | --------- | ------ | -----------
493 id | yes | string | The identifier of the link to be used when referring to it.
494 bandwidth | yes | int | Maximum bandwidth for this link, given in bytes/s
495 latency | no | double (default: 0.0) | Latency for this link.
496 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.
497 state | no | ON\|OFF (default: ON) | Allows you to to turn this link on or off (working / not working)
498 bandwidth_file | no | string | Allows you to use a file as input for bandwidth.
499 latency_file | no | string | Allows you to use a file as input for latency.
500 state_file | no | string | Allows you to use a file as input for states.
503 #### Possible shortcuts for ``latency`` ####
505 When using the latency attribute, you can specify the latency by using the scientific
506 notation or by using common abbreviations. For instance, the following three tags
510 <link id="LINK1" bandwidth="125000000" latency="5E-6"/>
511 <link id="LINK1" bandwidth="125000000" latency="5us"/>
512 <link id="LINK1" bandwidth="125000000" latency="0.000005"/>
515 Here, the second tag uses "us", meaning "microseconds". Other shortcuts are:
517 Name | Abbreviation | Time (in seconds)
518 ---- | ------------ | -----------------
519 Week | w | 7 * 24 * 60 * 60
520 Day | d | 24 * 60 * 60
524 Millisecond | ms | 0.001 = 10^(-3)
525 Microsecond | us | 0.000001 = 10^(-6)
526 Nanosecond | ns | 0.000000001 = 10^(-9)
527 Picosecond | ps | 0.000000000001 = 10^(-12)
529 #### Sharing policy ####
531 \anchor sharing_policy_shared
532 By default a network link is \b SHARED, i.e., if two or more data flows go
533 through a link, the bandwidth is shared fairly among all data flows. This
534 is similar to the sharing policy TCP uses.
536 \anchor pf_sharing_policy_fatpipe
537 On the other hand, if a link is defined as a \b FATPIPE,
538 each flow going through this link will be provided with the complete bandwidth,
539 i.e., no sharing occurs and the bandwidth is only limiting each flow individually.
540 Please note that this is really on a per-flow basis, not only on a per-host basis!
541 The complete bandwidth provided by this link in this mode
542 is ``number_of_flows*bandwidth``, with at most ``bandwidth`` being available per flow.
544 Using the FATPIPE mode allows to model backbones that won't affect performance
547 \anchor pf_sharing_policy_fullduplex
548 The last mode available is \b FULLDUPLEX. This means that SimGrid will
549 automatically generate two links (one carrying the suffix _UP and the other the
550 suffix _DOWN) for each ``<link>`` tag. This models situations when the direction
551 of traffic is important.
554 Transfers from one side to the other will interact similarly as
555 TCP when ACK returning packets circulate on the other direction. More
556 discussion about it is available in the description of link_ctn description.
558 In other words: The SHARED policy defines a physical limit for the bandwidth.
559 The FATPIPE mode defines a limit for each application,
560 with no upper total limit.
563 Tip: By using the FATPIPE mode, you can model big backbones that
564 won't affect performance (except latency).
569 <link id="SWITCH" bandwidth="125000000" latency="5E-5" sharing_policy="FATPIPE" />
572 #### Expressing dynamism and failures ####
574 Similar to hosts, it is possible to declare links whose state, bandwidth
575 or latency changes over time (see Section \ref pf_host_dynamism for details).
577 In the case of network links, the ``bandwidth`` and ``latency`` attributes are
578 replaced by the ``bandwidth_file`` and ``latency_file`` attributes.
579 The following XML snippet demonstrates how to use this feature in the platform
580 file. The structure of the files "link1.bw" and "link1.lat" is shown below.
583 <link id="LINK1" state_file="link1.fail" bandwidth="80000000" latency=".0001" bandwidth_file="link1.bw" latency_file="link1.lat" />
587 Even if the syntax is the same, the semantic of bandwidth and latency
588 trace files differs from that of host availability files. For bandwidth and
589 latency, the corresponding files do not
590 express availability as a fraction of the available capacity but directly in
591 bytes per seconds for the bandwidth and in seconds for the latency. This is
592 because most tools allowing to capture traces on real platforms (such as NWS)
593 express their results this way.
595 ##### Example of "link1.bw" file #####
603 In this example, the bandwidth changes repeatedly, with all changes
604 being repeated every 12 seconds.
606 At the beginning of the the simulation, the link's bandwidth is 80,000,000
607 B/s (i.e., 80 Mb/s); this value was defined in the XML snippet above.
608 After four seconds, it drops to 40 Mb/s (line 2), and climbs
609 back to 60 Mb/s after another 4 seconds (line 3). The value does not change any
610 more until the end of the period, that is, after 12 seconds have been simulated).
611 At this point, periodicity kicks in and this behavior is repeated: Seconds
612 12-16 will experience 80 Mb/s, 16-20 40 Mb/s etc.).
614 ##### Example of "link1.lat" file #####
623 In this example, the latency varies with a period of 5 seconds.
624 In the xml snippet above, the latency is initialized to be 0.0001s (100µs). This
625 value will be kept during the first second, since the latency_file contains
626 changes to this value at second one, two and three.
627 At second one, the value will be 0.001, i.e., 1ms. One second later it will
628 be adjusted to 0.01 (or 10ms) and one second later it will be set again to 1ms. The
629 value will not change until second 5, when the periodicity defined in line 1
630 kicks in. It then loops back, starting at 100µs (the initial value) for one second.
633 #### The ``<prop/>`` tag ####
635 Similar to the ``<host>`` tag, a link may also contain the ``<prop/>`` tag; see the host
636 documentation (Section \ref pf_host) for an example.
639 \subsubsection pf_backbone <backbone/>
642 This tag is <b>only available</b> when the containing AS uses the "Cluster" routing mode!
644 Using this tag, you can designate an already existing link to be a backbone.
646 Attribute name | Mandatory | Values | Description
647 --------------- | --------- | ------ | -----------
648 id | yes | string | Name of the link that is supposed to act as a backbone.
650 \subsection pf_storage Storage
653 This is a prototype version that should evolve quickly, hence this
654 is just some doc valuable only at the time of writing.
655 This section describes the storage management under SimGrid ; nowadays
656 it's only usable with MSG. It relies basically on linux-like concepts.
657 You also may want to have a look to its corresponding section in
658 @ref msg_file ; access functions are organized as a POSIX-like
661 \subsubsection pf_sto_conc Storage - Main Concepts
663 The storage facilities implemented in SimGrid help to model (and account for)
664 storage devices, such as tapes, hard-drives, CD or DVD devices etc.
665 A typical situation is depicted in the figure below:
667 \image html ./webcruft/storage_sample_scenario.png
668 \image latex ./webcruft/storage_sample_scenario.png "storage_sample_scenario" width=\textwidth
670 In this figure, two hosts called Bob and Alice are interconnected via a network
671 and each host is physically attached to a disk; it is not only possible for each host to
672 mount the disk they are attached to directly, but they can also mount disks
673 that are in a remote location. In this example, Bob mounts Alice's disk remotely
674 and accesses the storage via the network.
676 SimGrid provides 3 different entities that can be used to model setups
677 that include storage facilities:
679 Entity name | Description
680 --------------- | -----------
681 \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).
682 \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.
683 \ref pf_storage_entity_mount "mount" | Must be wrapped by a \ref pf_host tag; declares which storage(s) this host has mounted and where (i.e., the mountpoint).
686 \anchor pf_storage_content_file
687 ### %Storage Content File ###
689 In order to assess exactly how much time is spent reading from the storage,
690 SimGrid needs to know what is stored on the storage device (identified by distinct (file-)name, like in a file system)
691 and what size this content has.
694 The content file is never changed by the simulation; it is parsed once
695 per simulation and kept in memory afterwards. When the content of the
696 storage changes, only the internal SimGrid data structures change.
698 \anchor pf_storage_content_file_structure
699 #### Structure of a %Storage Content File ####
701 Here is an excerpt from two storage content file; if you want to see the whole file, check
702 the file ``examples/platforms/content/storage_content.txt`` that comes with the
705 SimGrid essentially supports two different formats: UNIX-style filepaths should
706 follow the well known format:
709 /lib/libsimgrid.so.3.6.2 12710497
713 /bin/simgrid_update_xml 5018
714 /bin/graphicator 66986
715 /bin/simgrid-colorizer 2993
720 Windows filepaths, unsurprisingly, use the windows style:
723 \Windows\avastSS.scr 41664
724 \Windows\bfsvc.exe 75264
725 \Windows\bootstat.dat 67584
726 \Windows\CoreSingleLanguage.xml 31497
728 \Windows\dchcfg64.exe 335464
729 \Windows\dcmdev64.exe 93288
733 The different file formats come at a cost; in version 3.12 (and most likely
734 in later versions, too), copying files from windows-style storages to unix-style
735 storages (and vice versa) is not supported.
737 \anchor pf_storage_content_file_create
738 #### Generate a %Storage Content File ####
740 If you want to generate a storage content file based on your own filesystem (or at least a filesystem you have access to),
741 try running this command (works only on unix systems):
744 find . -type f -exec ls -1s --block=1 {} \; 2>/dev/null | awk '{ print $2 " " $1}' > ./content.txt
747 \subsubsection pf_storage_entities The Storage Entities
749 These are the entities that you can use in your platform files to include
750 storage in your model. See also the list of our \ref pf_storage_example_files "example files";
751 these might also help you to get started.
753 \anchor pf_storage_entity_storage_type
754 #### \<storage_type\> ####
756 Attribute name | Mandatory | Values | Description
757 --------------- | --------- | ------ | -----------
758 id | yes | string | Identifier of this storage_type; used when referring to it
759 model | yes | string | For reasons of future backwards compatibility only; specifies the name of the model for the storage that should be used
760 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)
761 content | yes | string | Path to a \ref pf_storage_content_file "Storage Content File" on your system. This file must exist.
762 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"!)
764 This tag must contain some predefined model properties, specified via the <model_prop> tag. Here is a list,
765 see below for an example:
767 Property id | Mandatory | Values | Description
768 --------------- | --------- | ------ | -----------
769 Bwrite | yes | string | Bandwidth for write access; in B/s (but you can also specify e.g. "30MBps")
770 Bread | yes | string | Bandwidth for read access; in B/s (but you can also specify e.g. "30MBps")
771 Bconnexion | yes | string | Throughput (of the storage connector) in B/s.
774 A storage_type can also contain the <b><prop></b> tag. The <prop> tag allows you
775 to associate additional information to this <storage_type> and follows the
776 attribute/value schema; see the example below. You may want to use it to give information to
777 the tool you use for rendering your simulation, for example.
779 Here is a complete example for the ``storage_type`` tag:
781 <storage_type id="single_HDD" model="linear_no_lat" size="4000" content_type="txt_unix">
782 <model_prop id="Bwrite" value="30MBps" />
783 <model_prop id="Bread" value="100MBps" />
784 <model_prop id="Bconnection" value="150MBps" />
785 <prop id="Brand" value="Western Digital" />
789 \anchor pf_storage_entity_storage
790 #### <storage> ####
792 ``storage`` attributes:
794 Attribute name | Mandatory | Values | Description
795 -------------- | --------- | ------ | -----------
796 id | yes | string | Identifier of this ``storage``; used when referring to it
797 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.
798 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)
799 content | no | string | When specified, overwrites the content attribute of \ref pf_storage_entity_storage_type "\<storage_type\>"
800 content_type | no | string | When specified, overwrites the content_type attribute of \ref pf_storage_entity_storage_type "\<storage_type\>"
802 Here are two examples:
805 <storage id="Disk1" typeId="single_HDD" attach="bob" />
807 <storage id="Disk2" typeId="single_SSD"
808 content="content/win_storage_content.txt"
809 content_type="txt_windows" attach="alice" />
812 The first example is straightforward: A disk is defined and called "Disk1"; it is
813 of type "single_HDD" (shown as an example of \ref pf_storage_entity_storage_type "\<storage_type\>" above) and attached
814 to a host called "bob" (the definition of this host is omitted here).
816 The second storage is called "Disk2", is still of the same type as Disk1 but
817 now specifies a new content file (so the contents will be different from Disk1)
818 and the filesystem uses the windows style; finally, it is attached to a second host,
819 called alice (which is again not defined here).
821 \anchor pf_storage_entity_mount
822 #### <mount> ####
825 | Attribute name | Mandatory | Values | Description |
826 | ---------------- | ----------- | -------- | ------------- |
827 | id | yes | string | Refers to a \ref pf_storage_entity_storage "<storage>" entity that will be mounted on that computer |
828 | name | yes | string | Path/location to/of the logical reference (mount point) of this disk
830 This tag must be enclosed by a \ref pf_host tag. It then specifies where the mountpoint of a given storage device (defined by the ``id`` attribute)
831 is; this location is specified by the ``name`` attribute.
833 Here is a simple example, taken from the file ``examples/platform/storage.xml``:
836 <storage_type id="single_SSD" model="linear_no_lat" size="500GiB">
837 <model_prop id="Bwrite" value="60MBps" />
838 <model_prop id="Bread" value="200MBps" />
839 <model_prop id="Bconnection" value="220MBps" />
842 <storage id="Disk2" typeId="single_SSD"
843 content="content/win_storage_content.txt"
844 content_type="txt_windows" attach="alice" />
845 <storage id="Disk4" typeId="single_SSD"
846 content="content/small_content.txt"
847 content_type="txt_unix" attach="denise"/>
849 <host id="alice" speed="1Gf">
850 <mount storageId="Disk2" name="c:"/>
853 <host id="denise" speed="1Gf">
854 <mount storageId="Disk2" name="c:"/>
855 <mount storageId="Disk4" name="/home"/>
859 This example is quite interesting, as the same device, called "Disk2", is mounted by
860 two hosts at the same time! Note, however, that the host called ``alice`` is actually
861 attached to this storage, as can be seen in the \ref pf_storage_entity_storage "<storage>"
862 tag. This means that ``denise`` must access this storage through the network, but SimGrid automatically takes
863 care of that for you.
865 Furthermore, this example shows that ``denise`` has mounted two storages with different
866 filesystem types (unix and windows). In general, a host can mount as many storage devices as
870 Again, the difference between ``attach`` and ``mount`` is simply that
871 an attached storage is always physically inside (or connected to) that machine;
872 for instance, a USB stick is attached to one and only one machine (where it's plugged-in)
873 but it can only be mounted on others, as mounted storage can also be a remote location.
875 ###### Example files #####
877 \verbinclude example_filelist_xmltag_mount
879 \subsubsection pf_storage_example_files Example files
881 Several examples were already discussed above; if you're interested in full examples,
882 check the the following platforms:
884 1. ``examples/platforms/storage.xml``
885 2. ``examples/platforms/remote_io.xml``
887 If you're looking for some examplary C code, you may find the source code
888 available in the directory ``examples/msg/io/`` useful.
890 \subsubsection pf_storage_examples_modelling Modelling different situations
892 The storage functionality of SimGrid is type-agnostic, that is, the implementation
893 does not presume any type of storage, such as HDDs/SSDs, RAM,
894 CD/DVD devices, USB sticks etc.
896 This allows the user to apply the simulator for a wide variety of scenarios; one
897 common scenario would be the access of remote RAM.
899 #### Modelling the access of remote RAM ####
901 How can this be achieved in SimGrid? Let's assume we have a setup where three hosts
902 (HostA, HostB, HostC) need to access remote RAM:
912 An easy way to model this scenario is to setup and define the RAM via the
913 \ref pf_storage_entity_storage "storage" and \ref pf_storage_entity_storage_type "storage type"
914 entities and attach it to a remote dummy host; then, every host can have their own links
915 to this host (modelling for instance certain scenarios, such as PCIe ...)
920 RAM - Dummy -- Host B
925 Now, if read from this storage, the host that mounts this storage
926 communicates to the dummy host which reads from RAM and
927 sends the information back.
930 \section pf_routing Routing
932 To achieve high performance, the routing tables used within SimGrid are
933 static. This means that routing between two nodes is calculated once
934 and will not change during execution. The SimGrid team chose to use this
935 approach as it is rare to have a real deficiency of a resource;
936 most of the time, a communication fails because the links experience too much
937 congestion and hence, your connection stops before the timeout or
938 because the computer designated to be the destination of that message
941 We also chose to use shortest paths algorithms in order to emulate
942 routing. Doing so is consistent with the reality: [RIP](https://en.wikipedia.org/wiki/Routing_Information_Protocol),
943 [OSPF](https://en.wikipedia.org/wiki/Open_Shortest_Path_First), [BGP](https://en.wikipedia.org/wiki/Border_Gateway_Protocol)
944 are all calculating shortest paths. They do require some time to converge, but
945 eventually, when the routing tables have stabilized, your packets will follow
948 \subsection pf_rm Routing models
950 For each AS, you must define explicitly which routing model will
951 be used. There are 3 different categories for routing models:
953 1. \ref pf_routing_model_shortest_path "Shortest-path" based models: SimGrid calculates shortest
954 paths and manages them. Behaves more or less like most real life
956 2. \ref pf_routing_model_manual "Manually-entered" route models: you have to define all routes
957 manually in the platform description file; this can become
958 tedious very quickly, as it is very verbose.
959 Consistent with some manually managed real life routing.
960 3. \ref pf_routing_model_simple "Simple/fast models": those models offer fast, low memory routing
961 algorithms. You should consider to use this type of model if
962 you can make some assumptions about your AS.
963 Routing in this case is more or less ignored.
965 \subsubsection pf_raf The router affair
967 Using routers becomes mandatory when using shortest-path based
968 models or when using the bindings to the ns-3 packet-level
969 simulator instead of the native analytical network model implemented
972 For graph-based shortest path algorithms, routers are mandatory, because these
973 algorithms require a graph as input and so we need to have source and
974 destination for each edge.
976 Routers are naturally an important concept ns-3 since the
977 way routers run the packet routing algorithms is actually simulated.
978 SimGrid's analytical models however simply aggregate the routing time
979 with the transfer time.
981 So why did we incorporate routers in SimGrid? Rebuilding a graph representation
982 only from the route information turns out to be a very difficult task, because
983 of the missing information about how routes intersect. That is why we
984 introduced routers, which are simply used to express these intersection points.
985 It is important to understand that routers are only used to provide topological
988 To express this topological information, a <b>route</b> has to be
989 defined in order to declare which link is connected to a router.
992 \subsubsection pf_routing_model_shortest_path Shortest-path based models
994 The following table shows all the models that compute routes using
995 shortest-paths algorithms are currently available in SimGrid. More detail on how
996 to choose the best routing model is given in the Section called \"\ref pf_routing_howto_choose_wisely\".
998 | Name | Description |
999 | --------------------------------------------------- | -------------------------------------------------------------------------- |
1000 | \ref pf_routing_model_floyd "Floyd" | Floyd routing data. Pre-calculates all routes once |
1001 | \ref pf_routing_model_dijkstra "Dijkstra" | Dijkstra routing data. Calculates routes only when needed |
1002 | \ref pf_routing_model_dijkstracache "DijkstraCache" | Dijkstra routing data. Handles some cache for already calculated routes. |
1004 All those shortest-path models are instanciated in the same way and are
1005 completely interchangeable. Here are some examples:
1007 \anchor pf_routing_model_floyd
1012 <AS id="AS0" routing="Floyd">
1014 <cluster id="my_cluster_1" prefix="c-" suffix=""
1015 radical="0-1" speed="1000000000" bw="125000000" lat="5E-5"
1016 router_id="router1"/>
1018 <AS id="AS1" routing="None">
1019 <host id="host1" speed="1000000000"/>
1022 <link id="link1" bandwidth="100000" latency="0.01"/>
1024 <ASroute src="my_cluster_1" dst="AS1"
1027 <link_ctn id="link1"/>
1033 ASroute given at the end gives a topological information: link1 is
1034 between router1 and host1.
1036 #### Example platform files ####
1038 This is an automatically generated list of example files that use the Floyd
1039 routing model (the path is given relative to SimGrid's source directory)
1041 \verbinclude example_filelist_routing_floyd
1043 \anchor pf_routing_model_dijkstra
1046 #### Example platform files ####
1048 This is an automatically generated list of example files that use the Dijkstra
1049 routing model (the path is given relative to SimGrid's source directory)
1051 \verbinclude example_filelist_routing_dijkstra
1055 <AS id="AS_2" routing="Dijkstra">
1056 <host id="AS_2_host1" speed="1000000000"/>
1057 <host id="AS_2_host2" speed="1000000000"/>
1058 <host id="AS_2_host3" speed="1000000000"/>
1059 <link id="AS_2_link1" bandwidth="1250000000" latency="5E-4"/>
1060 <link id="AS_2_link2" bandwidth="1250000000" latency="5E-4"/>
1061 <link id="AS_2_link3" bandwidth="1250000000" latency="5E-4"/>
1062 <link id="AS_2_link4" bandwidth="1250000000" latency="5E-4"/>
1063 <router id="central_router"/>
1064 <router id="AS_2_gateway"/>
1065 <!-- routes providing topological information -->
1066 <route src="central_router" dst="AS_2_host1"><link_ctn id="AS_2_link1"/></route>
1067 <route src="central_router" dst="AS_2_host2"><link_ctn id="AS_2_link2"/></route>
1068 <route src="central_router" dst="AS_2_host3"><link_ctn id="AS_2_link3"/></route>
1069 <route src="central_router" dst="AS_2_gateway"><link_ctn id="AS_2_link4"/></route>
1073 \anchor pf_routing_model_dijkstracache
1074 ### DijkstraCache ###
1076 DijkstraCache example:
1078 <AS id="AS_2" routing="DijkstraCache">
1079 <host id="AS_2_host1" speed="1000000000"/>
1081 (platform unchanged compared to upper example)
1084 #### Example platform files ####
1086 This is an automatically generated list of example files that use the DijkstraCache
1087 routing model (the path is given relative to SimGrid's source directory):
1089 Editor's note: At the time of writing, no platform file used this routing model - so
1090 if there are no example files listed here, this is likely to be correct.
1092 \verbinclude example_filelist_routing_dijkstra_cache
1094 \subsubsection pf_routing_model_manual Manually-entered route models
1096 | Name | Description |
1097 | ---------------------------------- | ------------------------------------------------------------------------------ |
1098 | \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. |
1100 \anchor pf_routing_model_full
1105 <AS id="AS0" routing="Full">
1106 <host id="host1" speed="1000000000"/>
1107 <host id="host2" speed="1000000000"/>
1108 <link id="link1" bandwidth="125000000" latency="0.000100"/>
1109 <route src="host1" dst="host2"><link_ctn id="link1"/></route>
1113 #### Example platform files ####
1115 This is an automatically generated list of example files that use the Full
1116 routing model (the path is given relative to SimGrid's source directory):
1118 \verbinclude example_filelist_routing_full
1120 \subsubsection pf_routing_model_simple Simple/fast models
1122 | Name | Description |
1123 | ---------------------------------------- | ------------------------------------------------------------------------------ |
1124 | \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. |
1125 | \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. |
1126 | \ref pf_routing_model_vivaldi "Vivaldi" | Perfect when you want to use coordinates. Also see the corresponding \ref pf_P2P_tags "P2P section" below. |
1128 \anchor pf_routing_model_cluster
1132 In this mode, the \ref pf_cabinet "<cabinet/>" tag is available.
1134 #### Example platform files ####
1136 This is an automatically generated list of example files that use the Cluster
1137 routing model (the path is given relative to SimGrid's source directory):
1139 \verbinclude example_filelist_routing_cluster
1141 \anchor pf_routing_model_none
1144 This model does exactly what it's name advertises: Nothing. There is no routing
1145 available within this model and if you try to communicate within the AS that
1146 uses this model, SimGrid will fail unless you have explicitly activated the
1147 \ref options_model_select_network_constant "Constant Network Model" (this model charges
1148 the same for every single communication). It should
1149 be noted, however, that you can still attach an \ref pf_routing_tag_asroute "ASroute",
1150 as is demonstrated in the example below:
1152 \verbinclude platforms/cluster_and_one_host.xml
1154 #### Example platform files ####
1156 This is an automatically generated list of example files that use the None
1157 routing model (the path is given relative to SimGrid's source directory):
1159 \verbinclude example_filelist_routing_none
1162 \anchor pf_routing_model_vivaldi
1165 For more information on how to use the [Vivaldi Coordinates](https://en.wikipedia.org/wiki/Vivaldi_coordinates),
1166 see also Section \ref pf_P2P_tags "P2P tags".
1168 Note that it is possible to combine the Vivaldi routing model with other routing models;
1169 an example can be found in the file \c examples/platforms/cloud.xml. This
1170 examples models a NetZone using Vivaldi that contains other NetZones that use different
1173 #### Example platform files ####
1175 This is an automatically generated list of example files that use the None
1176 routing model (the path is given relative to SimGrid's source directory):
1178 \verbinclude example_filelist_routing_vivaldi
1181 \subsection ps_dec Defining routes
1183 There are currently four different ways to define routes:
1185 | Name | Description |
1186 | ------------------------------------------------- | ----------------------------------------------------------------------------------- |
1187 | \ref pf_routing_tag_route "route" | Used to define route between host/router |
1188 | \ref pf_routing_tag_asroute "ASroute" | Used to define route between different AS |
1189 | \ref pf_routing_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. |
1190 | \ref pf_routing_tag_bypassasroute "bypassASroute" | Used in the same way as bypassRoute, but for AS |
1192 Basically all those tags will contain an (ordered) list of references
1193 to link that compose the route you want to define.
1195 Consider the example below:
1198 <route src="Alice" dst="Bob">
1199 <link_ctn id="link1"/>
1200 <link_ctn id="link2"/>
1201 <link_ctn id="link3"/>
1205 The route here from host Alice to Bob will be first link1, then link2,
1206 and finally link3. What about the reverse route? \ref pf_routing_tag_route "Route" and
1207 \ref pf_routing_tag_asroute "ASroute" have an optional attribute \c symmetrical, that can
1208 be either \c YES or \c NO. \c YES means that the reverse route is the same
1209 route in the inverse order, and is set to \c YES by default. Note that
1210 this is not the case for bypass*Route, as it is more probable that you
1211 want to bypass only one default route.
1213 For an \ref pf_routing_tag_asroute "ASroute", things are just slightly more complicated, as you have
1214 to give the id of the gateway which is inside the AS you want to access ...
1215 So it looks like this:
1218 <ASroute src="AS1" dst="AS2"
1219 gw_src="router1" gw_dst="router2">
1220 <link_ctn id="link1"/>
1224 gw == gateway, so when any message are trying to go from AS1 to AS2,
1225 it means that it must pass through router1 to get out of the AS, then
1226 pass through link1, and get into AS2 by being received by router2.
1227 router1 must belong to AS1 and router2 must belong to AS2.
1229 \subsubsection pf_linkctn <link_ctn/>
1231 This entity has only one purpose: Refer to an already existing
1232 \ref pf_link "<link/>" when defining a route, i.e., it
1233 can only occur as a child of \ref pf_routing_tag_route "<route/>"
1235 | Attribute name | Mandatory | Values | Description |
1236 | --------------- | --------- | ------ | ----------- |
1237 | id | yes | String | The identifier of the link that should be added to the route. |
1238 | 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.
1240 #### Example Files ####
1242 This is an automatically generated list of example files that use the \c <link_ctn/>
1243 entity (the path is given relative to SimGrid's source directory):
1245 \verbinclude example_filelist_xmltag_linkctn
1247 \subsubsection pf_routing_tag_asroute ASroute
1249 The purpose of this entity is to define a route between two ASes.
1250 This is mainly useful when you're in the \ref pf_routing_model_full "Full routing model".
1252 #### Attributes ####
1254 | Attribute name | Mandatory | Values | Description |
1255 | --------------- | --------- | ------ | ----------- |
1256 | src | yes | String | The identifier of the source AS |
1257 | dst | yes | String | See the \c src attribute |
1258 | gw_src | yes | String | The gateway that will be used within the src AS; this can be any \ref pf_host "Host" or \ref pf_router "Router" defined within the src AS. |
1259 | gw_dst | yes | String | Same as \c gw_src, but with the dst AS instead. |
1260 | symmetrical | no | YES\|NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1265 <AS id="AS0" routing="Full">
1266 <cluster id="my_cluster_1" prefix="c-" suffix=".me"
1267 radical="0-149" speed="1000000000" bw="125000000" lat="5E-5"
1268 bb_bw="2250000000" bb_lat="5E-4"/>
1270 <cluster id="my_cluster_2" prefix="c-" suffix=".me"
1271 radical="150-299" speed="1000000000" bw="125000000" lat="5E-5"
1272 bb_bw="2250000000" bb_lat="5E-4"/>
1274 <link id="backbone" bandwidth="1250000000" latency="5E-4"/>
1276 <ASroute src="my_cluster_1" dst="my_cluster_2"
1277 gw_src="c-my_cluster_1_router.me"
1278 gw_dst="c-my_cluster_2_router.me">
1279 <link_ctn id="backbone"/>
1281 <ASroute src="my_cluster_2" dst="my_cluster_1"
1282 gw_src="c-my_cluster_2_router.me"
1283 gw_dst="c-my_cluster_1_router.me">
1284 <link_ctn id="backbone"/>
1289 \subsubsection pf_routing_tag_route route
1291 The principle is the same as for
1292 \ref pf_routing_tag_asroute "ASroute": The route contains a list of links that
1293 provide a path from \c src to \c dst. Here, \c src and \c dst can both be either a
1294 \ref pf_host "host" or \ref pf_router "router". This is mostly useful for the
1295 \ref pf_routing_model_full "Full routing model" as well as for the
1296 \ref pf_routing_model_shortest_path "shortest-paths" based models (as they require
1297 topological information).
1300 | Attribute name | Mandatory | Values | Description |
1301 | --------------- | --------- | ---------------------- | ----------- |
1302 | src | yes | String | The value given to the source's "id" attribute |
1303 | dst | yes | String | The value given to the destination's "id" attribute. |
1304 | symmetrical | no | YES\| NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1309 A route in the \ref pf_routing_model_full "Full routing model" could look like this:
1311 <route src="Tremblay" dst="Bourassa">
1312 <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"/>
1316 A route in the \ref pf_routing_model_shortest_path "Shortest-Path routing model" could look like this:
1318 <route src="Tremblay" dst="Bourassa">
1323 You must only have one link in your routes when you're using them to provide
1324 topological information, as the routes here are simply the edges of the
1325 (network-)graph and the employed algorithms need to know which edge connects
1326 which pair of entities.
1328 \subsubsection pf_routing_tag_bypassasroute bypassASroute
1330 As said before, once you choose
1331 a model, it (most likely; the constant network model, for example, doesn't) calculates routes for you. But maybe you want to
1332 define some of your routes, which will be specific. You may also want
1333 to bypass some routes defined in lower level AS at an upper stage:
1334 <b>bypassASroute</b> is the tag you're looking for. It allows to
1335 bypass routes defined between already defined between AS (if you want
1336 to bypass route for a specific host, you should just use byPassRoute).
1337 The principle is the same as ASroute : <b>bypassASroute</b> contains
1338 list of links that are in the path between src and dst.
1340 #### Attributes ####
1342 | Attribute name | Mandatory | Values | Description |
1343 | --------------- | --------- | ---------------------- | ----------- |
1344 | src | yes | String | The value given to the source AS's "id" attribute |
1345 | dst | yes | String | The value given to the destination AS's "id" attribute. |
1346 | 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 |
1347 | 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|
1348 | symmetrical | no | YES\| NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1353 <bypassASRoute src="my_cluster_1" dst="my_cluster_2"
1354 gw_src="my_cluster_1_router"
1355 gw_dst="my_cluster_2_router">
1356 <link_ctn id="link_tmp"/>
1360 This example shows that link \c link_tmp (definition not displayed here) directly
1361 connects the router \c my_cluster_1_router in the source cluster to the router
1362 \c my_cluster_2_router in the destination router. Additionally, as the \c symmetrical
1363 attribute was not given, this route is presumed to be symmetrical.
1365 \subsubsection pf_routing_tag_bypassroute bypassRoute
1367 As said before, once you choose
1368 a model, it (most likely; the constant network model, for example, doesn't) calculates routes for you. But maybe you want to
1369 define some of your routes, which will be specific. You may also want
1370 to bypass some routes defined in lower level AS at an upper stage :
1371 <b>bypassRoute</b> is the tag you're looking for. It allows to bypass
1372 routes defined between <b>host/router</b>. The principle is the same
1373 as route : <b>bypassRoute</b> contains list of links references of
1374 links that are in the path between src and dst.
1376 #### Attributes ####
1378 | Attribute name | Mandatory | Values | Description |
1379 | --------------- | --------- | ---------------------- | ----------- |
1380 | src | yes | String | The value given to the source AS's "id" attribute |
1381 | dst | yes | String | The value given to the destination AS's "id" attribute. |
1382 | symmetrical | no | YES \| NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1387 <bypassRoute src="host_1" dst="host_2">
1388 <link_ctn id="link_tmp"/>
1392 This example shows that link \c link_tmp (definition not displayed here) directly
1393 connects host \c host_1 to host \c host_2. Additionally, as the \c symmetrical
1394 attribute was not given, this route is presumed to be symmetrical.
1396 \subsection pb_baroex Basic Routing Example
1398 Let's say you have an AS named AS_Big that contains two other AS, AS_1
1399 and AS_2. If you want to make a host (h1) from AS_1 with another one
1400 (h2) from AS_2 then you'll have to proceed as follows:
1401 \li First, you have to ensure that a route is defined from h1 to the
1402 AS_1's exit gateway and from h2 to AS_2's exit gateway.
1403 \li Then, you'll have to define a route between AS_1 to AS_2. As those
1404 AS are both resources belonging to AS_Big, then it has to be done
1405 at AS_big level. To define such a route, you have to give the
1406 source AS (AS_1), the destination AS (AS_2), and their respective
1407 gateway (as the route is effectively defined between those two
1408 entry/exit points). Elements of this route can only be elements
1409 belonging to AS_Big, so links and routers in this route should be
1410 defined inside AS_Big. If you choose some shortest-path model,
1411 this route will be computed automatically.
1413 As said before, there are mainly 2 tags for routing :
1414 \li <b>ASroute</b>: to define routes between two <b>AS</b>
1415 \li <b>route</b>: to define routes between two <b>host/router</b>
1417 As we are dealing with routes between AS, it means that those we'll
1418 have some definition at AS_Big level. Let consider AS_1 contains 1
1419 host, 1 link and one router and AS_2 3 hosts, 4 links and one router.
1420 There will be a central router, and a cross-like topology. At the end
1421 of the crosses arms, you'll find the 3 hosts and the router that will
1422 act as a gateway. We have to define routes inside those two AS. Let
1423 say that AS_1 contains full routes, and AS_2 contains some Floyd
1424 routing (as we don't want to bother with defining all routes). As
1425 we're using some shortest path algorithms to route into AS_2, we'll
1426 then have to define some <b>route</b> to gives some topological
1427 information to SimGrid. Here is a file doing it all :
1430 <AS id="AS_Big" routing="Dijkstra">
1431 <AS id="AS_1" routing="Full">
1432 <host id="AS_1_host1" speed="1000000000"/>
1433 <link id="AS_1_link" bandwidth="1250000000" latency="5E-4"/>
1434 <router id="AS_1_gateway"/>
1435 <route src="AS_1_host1" dst="AS_1_gateway">
1436 <link_ctn id="AS_1_link"/>
1439 <AS id="AS_2" routing="Floyd">
1440 <host id="AS_2_host1" speed="1000000000"/>
1441 <host id="AS_2_host2" speed="1000000000"/>
1442 <host id="AS_2_host3" speed="1000000000"/>
1443 <link id="AS_2_link1" bandwidth="1250000000" latency="5E-4"/>
1444 <link id="AS_2_link2" bandwidth="1250000000" latency="5E-4"/>
1445 <link id="AS_2_link3" bandwidth="1250000000" latency="5E-4"/>
1446 <link id="AS_2_link4" bandwidth="1250000000" latency="5E-4"/>
1447 <router id="central_router"/>
1448 <router id="AS_2_gateway"/>
1449 <!-- routes providing topological information -->
1450 <route src="central_router" dst="AS_2_host1"><link_ctn id="AS_2_link1"/></route>
1451 <route src="central_router" dst="AS_2_host2"><link_ctn id="AS_2_link2"/></route>
1452 <route src="central_router" dst="AS_2_host3"><link_ctn id="AS_2_link3"/></route>
1453 <route src="central_router" dst="AS_2_gateway"><link_ctn id="AS_2_link4"/></route>
1455 <link id="backbone" bandwidth="1250000000" latency="5E-4"/>
1457 <ASroute src="AS_1" dst="AS_2"
1458 gw_src="AS_1_gateway"
1459 gw_dst="AS_2_gateway">
1460 <link_ctn id="backbone"/>
1465 \section pf_other_tags Tags not (directly) describing the platform
1467 The following tags can be used inside a \<platform\> tag even if they are not
1468 directly describing the platform:
1469 \li \ref pf_config "config": it allows you to pass some configuration stuff like, for
1470 example, the network model and so on. It follows the
1471 \li \ref pf_include "include": allows you to include another file into the current one.
1473 \subsection pf_config config
1475 The only purpose of this tag is to contain the \c prop tags, as described below.
1476 These tags will then configure the options as described by Section \ref options.
1479 #### Attributes ####
1481 | Attribute name | Mandatory | Values | Description |
1482 | --------------- | --------- | ---------------------- | ----------- |
1483 | id | yes | String | The identifier of the config tag when referring to id; this is basically useless, though. |
1485 #### Possible children ####
1487 Tag name | Description | Documentation
1488 ------------ | ----------- | -------------
1489 \<prop/\> | The prop tag allows you to define different configuration options following the attribute/value schema. See the \ref options page. | N/A
1494 <?xml version='1.0'?>
1495 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid.dtd">
1496 <platform version="4">
1497 <config id="General">
1498 <prop id="maxmin/precision" value="0.000010"></prop>
1499 <prop id="cpu/optim" value="TI"></prop>
1500 <prop id="host/model" value="compound"></prop>
1501 <prop id="network/model" value="SMPI"></prop>
1502 <prop id="path" value="~/"></prop>
1503 <prop id="smpi/bw-factor" value="65472:0.940694;15424:0.697866;9376:0.58729"></prop>
1506 <AS id="AS0" routing="Full">
1510 \subsection pf_include include
1512 Even if it can be used in other contexts, this tag was originally created
1513 to be used with \ref pf_trace. The idea was to have a file describing the
1514 platform, and another file attaching traces of a given period to the platform.
1516 The drawback is that the file chuncks that will be included do not
1517 constitute valid XML files. This may explain why this feature was never really
1518 used in practice (as far as we know). Other mechanisms, such as the ability to load
1519 several platform files one after the other, could be considered in the future.
1521 In the meanwhile, the \c include tag allows you to import other platforms into your
1522 local file. This is done with the intention to help people
1523 combine their different AS and provide new platforms. Those files
1524 should contain XML that consists of
1525 \ref pf_include "include", \ref pf_cluster "cluster", \ref pf_peer "peer", \ref pf_As "AS", \ref pf_trace "trace", \ref pf_trace "tags".
1527 Do not forget to close the tag to make it work, or you will end up with an invalid XML file.
1529 #### Attributes ####
1531 | Attribute name | Mandatory | Values | Description |
1532 | --------------- | --------- | ---------------------- | ----------- |
1533 | file | yes | String | Filename of the path you want to include with either relative or absolute path. |
1538 The following example includes two files, clusterA.xml and clusterB.xml and
1539 combines them two one platform file; all hosts, routers etc. defined in
1540 each of them will then be usable.
1543 <?xml version='1.0'?>
1544 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid/simgrid.dtd">
1545 <platform version="4">
1546 <AS id="main" routing="Full">
1547 <include file="clusterA.xml"></include>
1548 <include file="clusterB.xml"></include>
1553 \subsection pf_trace trace and trace_connect
1555 Both tags are an alternate way to pass files containing information on
1556 availability, state etc. to an entity. (See also, for instance, Section \ref
1557 pf_host_churn "Churn", as described for the host entity.) Instead of referring
1558 to the file directly in the host, link, or cluster tag, you proceed by defining
1559 a trace with an id corresponding to a file, later a host/link/cluster, and
1560 finally using trace_connect you say that the file trace must be used by the
1567 <AS id="AS0" routing="Full">
1568 <host id="bob" speed="1000000000"/>
1570 <trace id="myTrace" file="bob.trace" periodicity="1.0"/>
1571 <trace_connect trace="myTrace" element="bob" kind="POWER"/>
1575 The order here is important. \c trace_connect must come
1576 after the elements \c trace and \c host, as both the host
1577 and the trace definition must be known when \c trace_connect
1578 is parsed; the order of \c trace and \c host is arbitrary.
1581 #### \c trace attributes ####
1584 | Attribute name | Mandatory | Values | Description |
1585 | --------------- | --------- | ---------------------- | ----------- |
1586 | id | yes | String | Identifier of this trace; this is the name you pass on to \c trace_connect. |
1587 | 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. |
1588 | trace_periodicity | yes | String | This is the same as for \ref pf_host "hosts" (see there for details) |
1590 Here is an example of trace when no file name is provided:
1593 <trace id="myTrace" periodicity="1.0">
1600 #### \c trace_connect attributes ####
1602 | Attribute name | Mandatory | Values | Description |
1603 | --------------- | --------- | ---------------------- | ----------- |
1604 | kind | no | HOST_AVAIL\|POWER\|<br/>LINK_AVAIL\|BANDWIDTH\|LATENCY (Default: HOST_AVAIL) | Describes the kind of trace. |
1605 | trace | yes | String | Identifier of the referenced trace (specified of the trace's \c id attribute) |
1606 | element | yes | String | The identifier of the referenced entity as given by its \c id attribute |
1608 \section pf_hints Hints, tips and frequently requested features
1610 Now you should know at least the syntax and be able to create a
1611 platform by your own. However, after having ourselves wrote some platforms, there
1612 are some best practices you should pay attention to in order to
1613 produce good platform and some choices you can make in order to have
1614 faster simulations. Here's some hints and tips, then.
1616 @subsection Finding the platform example that you need
1618 Most platform files that we ship are in the @c examples/platforms
1619 folder. The good old @c grep tool can find the examples you need when
1620 wondering on a specific XML tag. Here is an example session searching
1621 for @ref pf_trace "trace_connect":
1624 % cd examples/platforms
1625 % grep -R -i -n --include="*.xml" "trace_connect" .
1626 ./two_hosts_platform_with_availability_included.xml:26:<trace_connect kind="SPEED" trace="A" element="Cpu A"/>
1627 ./two_hosts_platform_with_availability_included.xml:27:<trace_connect kind="HOST_AVAIL" trace="A_failure" element="Cpu A"/>
1628 ./two_hosts_platform_with_availability_included.xml:28:<trace_connect kind="SPEED" trace="B" element="Cpu B"/>
1629 ./two_hosts.xml:17: <trace_connect trace="Tremblay_power" element="Tremblay" kind="SPEED"/>
1632 \subsection pf_as_h AS Hierarchy
1633 The AS design allows SimGrid to go fast, because computing route is
1634 done only for the set of resources defined in this AS. If you're using
1635 only a big AS containing all resource with no AS into it and you're
1636 using Full model, then ... you'll loose all interest into it. On the
1637 other hand, designing a binary tree of AS with, at the lower level,
1638 only one host, then you'll also loose all the good AS hierarchy can
1639 give you. Remind you should always be "reasonable" in your platform
1640 definition when choosing the hierarchy. A good choice if you try to
1641 describe a real life platform is to follow the AS described in
1642 reality, since this kind of trade-off works well for real life
1645 \subsection pf_exit_as Exit AS: why and how
1646 Users that have looked at some of our platforms may have notice a
1647 non-intuitive schema ... Something like that :
1651 <AS id="AS_4" routing="Full">
1652 <AS id="exitAS_4" routing="Full">
1653 <router id="router_4"/>
1655 <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"/>
1656 <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"/>
1657 <link id="4_1" bandwidth="2250000000" latency="5E-5"/>
1658 <link id="4_2" bandwidth="2250000000" latency="5E-5"/>
1659 <link id="bb_4" bandwidth="2250000000" latency="5E-4"/>
1660 <ASroute src="cl_4_1"
1662 gw_src="c_4_1-cl_4_1_router"
1663 gw_dst="c_4_2-cl_4_2_router">
1664 <link_ctn id="4_1"/>
1665 <link_ctn id="bb_4"/>
1666 <link_ctn id="4_2"/>
1668 <ASroute src="cl_4_1"
1670 gw_src="c_4_1-cl_4_1_router"
1672 <link_ctn id="4_1"/>
1673 <link_ctn id="bb_4"/>
1675 <ASroute src="cl_4_2"
1677 gw_src="c_4_2-cl_4_2_router"
1679 <link_ctn id="4_2"/>
1680 <link_ctn id="bb_4"/>
1685 In the AS_4, you have an exitAS_4 defined, containing only one router,
1686 and routes defined to that AS from all other AS (as cluster is only a
1687 shortcut for an AS, see cluster description for details). If there was
1688 an upper AS, it would define routes to and from AS_4 with the gateway
1689 router_4. It's just because, as we did not allowed (for performances
1690 issues) to have routes from an AS to a single host/router, you have to
1691 enclose your gateway, when you have AS included in your AS, within an
1692 AS to define routes to it.
1694 \subsection pf_P2P_tags P2P or how to use coordinates
1695 SimGrid allows you to use some coordinated-based system, like vivaldi,
1696 to describe a platform. The main concept is that you have some peers
1697 that are located somewhere: this is the function of the
1698 <b>coordinates</b> of the \<peer\> or \<host\> tag. There's nothing
1699 complicated in using it, here is an example:
1702 <?xml version='1.0'?>
1703 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid.dtd">
1704 <platform version="4">
1706 <AS id="AS0" routing="Vivaldi">
1707 <host id="100030591" coordinates="25.5 9.4 1.4" speed="1.5Gf" />
1708 <host id="100036570" coordinates="-12.7 -9.9 2.1" speed="7.3Gf" />
1710 <host id="100429957" coordinates="17.5 6.7 18.8" speed="8.3Gf" />
1715 Coordinates are then used to calculate latency (in microseconds)
1716 between two hosts by calculating the distance between the two hosts
1717 coordinates with the following formula: distance( (x1, y1, z1), (x2,
1718 y2, z2) ) = euclidian( (x1,y1), (x2,y2) ) + abs(z1) + abs(z2)
1720 In other words, we take the euclidian distance on the two first
1721 dimensions, and then add the absolute values found on the third
1722 dimension. This may seem strange, but it was found to allow better
1723 approximations of the latency matrices (see the paper describing
1726 Note that the previous example defines a routing directly between hosts but it could be also used to define a routing between AS.
1727 That is for example what is commonly done when using peers (see Section \ref pf_peer).
1729 <?xml version='1.0'?>
1730 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid.dtd">
1731 <platform version="4">
1733 <AS id="AS0" routing="Vivaldi">
1734 <peer id="peer-0" coordinates="173.0 96.8 0.1" speed="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us"/>
1735 <peer id="peer-1" coordinates="247.0 57.3 0.6" speed="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us" />
1736 <peer id="peer-2" coordinates="243.4 58.8 1.4" speed="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us" />
1740 In such a case though, we connect the AS created by the <b>peer</b> tag with the Vivaldi routing mechanism.
1741 This means that to route between AS1 and AS2, it will use the coordinates of router_AS1 and router_AS2.
1742 This is currently a convention and we may offer to change this convention in the DTD later if needed.
1743 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.
1746 \subsection pf_routing_howto_choose_wisely Choosing wisely the routing model to use
1749 Choosing wisely the routing model to use can significantly fasten your
1750 simulation/save your time when writing the platform/save tremendous
1751 disk space. Here is the list of available model and their
1752 characteristics (lookup : time to resolve a route):
1754 \li <b>Full</b>: Full routing data (fast, large memory requirements,
1756 \li <b>Floyd</b>: Floyd routing data (slow initialization, fast
1757 lookup, lesser memory requirements, shortest path routing only).
1758 Calculates all routes at once at the beginning.
1759 \li <b>Dijkstra</b>: Dijkstra routing data (fast initialization, slow
1760 lookup, small memory requirements, shortest path routing only).
1761 Calculates a route when necessary.
1762 \li <b>DijkstraCache</b>: Dijkstra routing data (fast initialization,
1763 fast lookup, small memory requirements, shortest path routing
1764 only). Same as Dijkstra, except it handles a cache for latest used
1766 \li <b>None</b>: No routing (usable with Constant network only).
1767 Defines that there is no routes, so if you try to determine a
1768 route without constant network within this AS, SimGrid will raise
1770 \li <b>Vivaldi</b>: Vivaldi routing, so when you want to use coordinates
1771 \li <b>Cluster</b>: Cluster routing, specific to cluster tag, should
1774 \subsection pf_switch I want to describe a switch but there is no switch tag!
1776 Actually we did not include switch tag. But when you're trying to
1777 simulate a switch, assuming
1778 fluid bandwidth models are used (which SimGrid uses by default unless
1779 ns-3 or constant network models are activated), the limiting factor is
1780 switch backplane bandwidth. So, essentially, at least from
1781 the simulation perspective, a switch is similar to a
1782 link: some device that is traversed by flows and with some latency and
1783 so,e maximum bandwidth. Thus, you can simply simulate a switch as a
1785 can be connected to this "switch", which is then included in routes just
1789 \subsection pf_multicabinets I want to describe multi-cabinets clusters!
1791 You have several possibilities, as usual when modeling things. If your
1792 cabinets are homogeneous and the intercabinet network negligible for
1793 your study, you should just create a larger cluster with all hosts at
1796 In the rare case where your hosts are not homogeneous between the
1797 cabinets, you can create your cluster completely manually. For that,
1798 create an As using the Cluster routing, and then use one
1799 <cabinet> for each cabinet. This cabinet tag can only be used an
1800 As using the Cluster routing schema, and creating
1802 Be warned that creating a cluster manually from the XML with
1803 <cabinet>, <backbone> and friends is rather tedious. The
1804 easiest way to retrieve some control of your model without diving into
1805 the <cluster> internals is certainly to create one separate
1806 <cluster> per cabinet and interconnect them together. This is
1807 what we did in the G5K example platform for the Graphen cluster.
1809 \subsection pf_platform_multipath I want to express multipath routing in platform files!
1811 It is unfortunately impossible to express the fact that there is more
1812 than one routing path between two given hosts. Let's consider the
1813 following platform file:
1816 <route src="A" dst="B">
1819 <route src="B" dst="C">
1822 <route src="A" dst="C">
1827 Although it is perfectly valid, it does not mean that data traveling
1828 from A to C can either go directly (using link 3) or through B (using
1829 links 1 and 2). It simply means that the routing on the graph is not
1830 trivial, and that data do not following the shortest path in number of
1831 hops on this graph. Another way to say it is that there is no implicit
1832 in these routing descriptions. The system will only use the routes you
1833 declare (such as <route src="A" dst="C"><link_ctn
1834 id="3"/></route>), without trying to build new routes by aggregating
1837 You are also free to declare platform where the routing is not
1838 symmetrical. For example, add the following to the previous file:
1841 <route src="C" dst="A">
1847 This makes sure that data from C to A go through B where data from A
1848 to C go directly. Don't worry about realism of such settings since
1849 we've seen ways more weird situation in real settings (in fact, that's
1850 the realism of very regular platforms which is questionable, but
1851 that's another story).
