1 /*! \page platform Describing the virtual platform
8 \htmlinclude graphical-toc.svg
12 document.getElementById("VirtualPlatform").style="opacity:0.93999999;fill:#ff0000;fill-opacity:0.1;stroke:#000000;stroke-width:0.35277778;stroke-linecap:round;stroke-linejoin:round;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1";
16 As @ref starting_components "explained in the introduction," any
17 SimGrid study must entail the description of the platform on which you
18 want to simulate your application. You have to describe **each element
19 of your platform**, such as computing hosts, clusters, each disks,
20 links, etc. You must also define the **routing on your platform**, ie
21 which path is taken between two hosts. Finally, you may also describe
22 an **experimental scenario**, with qualitative changes (e.g.,
23 bandwidth changes representing an external load) and qualitative
24 changes (representing how some elements fail and restart over time).
26 You should really separate your application from the platform
27 description, as it will ease your experimental campain afterward.
28 Mixing them is seen as a really bad experimental practice. The easiest
29 to enforce this split is to put the platform description in a XML
30 file. Many example platforms are provided in the archive, and this
31 page gives all needed details to write such files, as well as some
32 hints and tricks about describing your platform.
34 On the other side, XML is sometimes not expressive enough for some
35 platforms, in particular large platforms exhibiting repetitive
36 patterns that are not simply expressed in XML. In practice, many
37 users end up generating their XML platform files from some sort of
38 scripts. It is probably preferable to rewrite your XML @ref
39 platform_lua "platform using the lua scripting language" instead.
40 In the future, it should be possible to describe the platform directly
41 in C++, but this is not possible yet.
43 As usual, SimGrid is a versatile framework, and you should find the
44 way of describing your platform that best fits your experimental
47 \section pf_overview Describing the platform with XML
49 Your platform description should follow the specification presented in
50 the [simgrid.dtd](http://simgrid.gforge.inria.fr/simgrid/simgrid.dtd)
51 DTD file. The same DTD is used for both the platform and deployment
54 From time to time, this DTD evolves to introduce possibly
55 backward-incompatible changes. That is why each platform desciption is
56 enclosed within a @c platform tag, that have a @c version attribute.
57 The current version is <b>4.1</b>. The @c simgrid_update_xml program can
58 upgrade most of the past platform files to the recent formalism.
60 \section pf_first_example First Platform Example
62 Here is a very simple platform file, containing 3 resources (two hosts
63 and one link), and explicitly giving the route between the hosts.
67 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid/simgrid.dtd">
68 <platform version="4.1">
69 <zone id="first zone" routing="Full">
70 <!-- the resources -->
71 <host id="host1" speed="1Mf"/>
72 <host id="host2" speed="2Mf"/>
73 <link id="link1" bandwidth="125MBps" latency="100us"/>
74 <!-- the routing: specify how the hosts are interconnected -->
75 <route src="host1" dst="host2">
76 <link_ctn id="link1"/>
82 As we said, the englobing @ref pf_overview "<platform>" tag is
83 used to specify the dtd version used for this file.
85 Then, every resource (specified with @ref pf_tag_host, @ref
86 pf_tag_link or others) must be located within a given **networking
87 zone**. Each zone is in charge of the routing between its
88 resources. It means that when an host wants to communicate with
89 another host of the same zone, it is the zone's duty to find the list
90 of links that are involved in the communication. Here, since the @ref
91 pf_tag_zone tag has **Full** as a **routing attribute**, all routes
92 must be explicitely given using the @ref pf_tag_route and @ref
93 pf_tag_linkctn tags (this @ref pf_rm "routing model" is both simple
94 and inefficient :) It is OK to not specify the route between two
95 hosts, as long as the processes located on them never try to
98 A zone can contain several zones itself, leading to a hierarchical
99 decomposition of the platform. This can be more efficient (as the
100 inter-zone routing gets factorized with @ref pf_tag_zoneroute), and
101 allows to have more than one routing model in your platform. For
102 example, you could have a coordinate-based routing for the WAN parts
103 of your platforms, a full routing within each datacenter, and a highly
104 optimized routing within each cluster of the datacenter. In this
105 case, determining the route between two given hosts gets @ref
106 routing_basics "somewhat more complex" but SimGrid still computes
107 these routes for you in a time- and space-efficient manner.
108 Here is an illustration of these concepts:
110 
112 Circles represent processing units and squares represent network
113 routers. Bold lines represent communication links. The zone "AS2"
114 models the core of a national network interconnecting a small flat
115 cluster (AS4) and a larger hierarchical cluster (AS5), a subset of a
116 LAN (AS6), and a set of peers scattered around the world (AS7).
118 \section pf_res Resource description
120 \subsection pf_res_computing Computing Resources
122 \subsubsection pf_tag_host <host>
124 An host is the computing resource on which an actor can execute.
126 Attribute | Values | Description
127 ----------------- | -------------------------------------- | -----------
128 id | String (mandatory) | The identifier of the host. facilitates referring to this AS.
129 speed | double (mandatory) | Computational power of every core of this host in FLOPS (must be positive)
130 core | int (defaults to 1) | Number of cores (see @ref howto_multicore)
131 state | optionally "OFF" | If set to OFF, the host is initially turned off.
132 availability_file | File name (optional) | (Relative or absolute) filename to use as input; must contain availability traces for this host. The syntax of this file is defined below.
133 state_file | File name (optional) | File to use as a state profile (see @ref howto_churn)
134 coordinates | String (mandatory when using Vivaldi routing) | The coordinates of this host (see @ref pf_P2P_tags).
135 pstate | Double (Defaults to 0) | FIXME: Not yet documented.
137 #### Included tags ####
139 - @ref pf_tag_mount Specifies the storages mounted on that host
140 - @ref pf_tag_prop Specifies a user-defined property of that host, that you can retrieve with MSG_host_get_property_value() or simgrid::s4u::Host::property().
145 <host id="host1" speed="1000000000"/>
146 <host id="host2" speed="1000000000">
147 <prop id="color" value="blue"/>
148 <prop id="rendershape" value="square"/>
152 \anchor pf_host_dynamism
153 ### Expressing dynamism ###
155 SimGrid provides mechanisms to change a hosts' availability over
156 time, using the ``availability_file`` attribute to the ``\<host\>`` tag
157 and a separate text file whose syntax is exemplified below.
159 #### Adding a trace file ####
162 <platform version="4">
163 <host id="bob" speed="500Gf" availability_file="bob.trace" />
167 #### Example of "bob.trace" file ####
176 Let us begin to explain this example by looking at line 2. (Line 1 will become clear soon).
177 The first column describes points in time, in this case, time 0. The second column
178 describes the relative amount of power this host is able to deliver (relative
179 to the maximum performance specified in the ``\<host\>`` tag). (Clearly, the
180 second column needs to contain values that are not smaller than 0 and not larger than 1).
181 In this example, our host will deliver 500 Mflop/s at time 0, as 500 Mflop/s is the
182 maximum performance of this host. At time 11.0, it will
183 deliver half of its maximum performance, i.e., 250 Mflop/s until time 20.0 when it will
184 will start delivering 80\% of its power. In this example, this amounts to 400 Mflop/s.
186 Since the periodicity in line 1 was set to be 1.0, i.e., 1 timestep, this host will
187 continue to provide 500 Mflop/s from time 21. From time 32 it will provide 250 MFlop/s and so on.
189 \subsubsection pf_tag_cluster <cluster>
191 ``<cluster />`` represents a machine-cluster. It is most commonly used
192 when one wants to define many hosts and a network quickly. Technically,
193 ``cluster`` is a meta-tag: <b>from the inner SimGrid point of
194 view, a cluster is a network zone where some optimized routing is defined</b>.
195 The default inner organization of the cluster is as follow:
201 ____________|__________|_____________ backbone
203 l0| l1| l2| l97| l96 | | l99
209 Here, a set of <b>host</b>s is defined. Each of them has a <b>link</b>
210 to a central backbone (backbone is a link itself, as a link can
211 be used to represent a switch, see the switch / link section
212 below for more details about it). A <b>router</b> allows to connect a
213 <b>cluster</b> to the outside world. Internally,
214 SimGrid treats a cluster as a network zone containing all hosts: the router is the default
215 gateway for the cluster.
217 There is an alternative organization, which is as follows:
231 The principle is the same, except that there is no backbone. This representation
232 can be obtained easily: just do not set the bb_* attributes.
235 Attribute name | Mandatory | Values | Description
236 --------------- | --------- | ------ | -----------
237 id | yes | string | The identifier of the cluster. Facilitates referring to this cluster.
238 prefix | yes | string | Each node of the cluster has to have a name. This name will be prefixed with this prefix.
239 suffix | yes | string | Each node of the cluster will be suffixed with this suffix
240 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.
241 speed | yes | int | Same as the ``speed`` attribute of the ``\<host\>`` tag.
242 core | no | int (default: 1) | Same as the ``core`` attribute of the ``\<host\>`` tag.
243 bw | yes | int | Bandwidth for the links between nodes and backbone (if any). See the \ref pf_tag_link "link section" for syntax/details.
244 lat | yes | int | Latency for the links between nodes and backbone (if any). See <b>link</b> section for syntax/details.
245 sharing_policy | no | string | Sharing policy for the links between nodes and backbone (if any). See <b>link</b> section for syntax/details.
246 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>).
247 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>).
248 bb_sharing_policy | no | string | Sharing policy for the backbone (if any). See <b>link</b> section for syntax/details.
249 limiter_link | no | int | Bandwidth for limiter link (if any). This adds a specific link for each node, to set the maximum bandwidth reached when communicating in both directions at the same time. In theory this value should be 2*bw for splitduplex links, but in reality this might be less. This value will depend heavily on the communication model, and on the cluster's hardware, so no default value can be set, this has to be measured. More details can be obtained in <a href="https://hal.inria.fr/hal-00919507/"> "Toward Better Simulation of MPI Applications on Ethernet/TCP Networks"</a>
250 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".
251 loopback_lat | no | int | Latency for loopback (if any). See <b>link</b> section for syntax/details. See loopback_bw for more info.
252 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.
253 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".
256 the router name is defined as the resulting String in the following
260 router_name = prefix + clusterId + "_router" + suffix;
264 #### Cluster example ####
266 Consider the following two (and independent) uses of the ``cluster`` tag:
269 <cluster id="my_cluster_1" prefix="" suffix="" radical="0-262144"
270 speed="1e9" bw="125e6" lat="5E-5"/>
272 <cluster id="my_cluster_2" prefix="c-" suffix=".me" radical="0-99"
273 speed="1e9" bw="125e6" lat="5E-5"
274 bb_bw="2.25e9" bb_lat="5E-4"/>
277 The second example creates one router and 100 machines with the following names:
279 c-my_cluster_2_router.me
287 \subsubsection pf_cabinet <cabinet>
290 This tag is only available when the routing mode of the network zone
291 is set to ``Cluster``.
293 The ``<cabinet />`` tag is, like the \ref pf_tag_cluster "<cluster>" tag,
294 a meta-tag. This means that it is simply a shortcut for creating a set of (homogenous) hosts and links quickly;
295 unsurprisingly, this tag was introduced to setup cabinets in data centers quickly. Unlike
296 <cluster>, however, the <cabinet> assumes that you create the backbone
297 and routers yourself; see our examples below.
301 Attribute name | Mandatory | Values | Description
302 --------------- | --------- | ------ | -----------
303 id | yes | string | The identifier of the cabinet. Facilitates referring to this cluster.
304 prefix | yes | string | Each node of the cabinet has to have a name. This name will be prefixed with this prefix.
305 suffix | yes | string | Each node of the cabinet will be suffixed with this suffix
306 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.
307 speed | yes | int | Same as the ``speed`` attribute of the \ref pf_tag_host "<host>" tag.
308 bw | yes | int | Bandwidth for the links between nodes and backbone (if any). See the \ref pf_tag_link "link section" for syntax/details.
309 lat | yes | int | Latency for the links between nodes and backbone (if any). See the \ref pf_tag_link "link section" for syntax/details.
312 Please note that as of now, it is impossible to change attributes such as,
313 amount of cores (always set to 1), the initial state of hosts/links
314 (always set to ON), the sharing policy of the links (always set to \ref pf_sharing_policy_splitduplex "SPLITDUPLEX").
318 The following example was taken from ``examples/platforms/meta_cluster.xml`` and
319 shows how to use the cabinet tag.
322 <zone id="my_cluster1" routing="Cluster">
323 <cabinet id="cabinet1" prefix="host-" suffix=".cluster1"
324 speed="1Gf" bw="125MBps" lat="100us" radical="1-10"/>
325 <cabinet id="cabinet2" prefix="host-" suffix=".cluster1"
326 speed="1Gf" bw="125MBps" lat="100us" radical="11-20"/>
327 <cabinet id="cabinet3" prefix="host-" suffix=".cluster1"
328 speed="1Gf" bw="125MBps" lat="100us" radical="21-30"/>
330 <backbone id="backbone1" bandwidth="2.25GBps" latency="500us"/>
335 Please note that you must specify the \ref pf_backbone "<backbone>"
336 tag by yourself; this is not done automatically and there are no checks
337 that ensure this backbone was defined.
339 The hosts generated in the above example are named host-1.cluster, host-2.cluster1
343 \subsubsection pf_peer \<peer\> (Vivaldi netzones only)
345 This tag represents a peer, as in Peer-to-Peer (P2P) networks. This
346 can only be used in Vivaldi NetZones. It creates the following
347 resources to the NetZone:
350 \li Two links: One for download and one for upload. This is
351 convenient to use and simulate stuff under the last mile model (e.g., ADSL peers).
352 \li It connects the two links to the host
356 Attribute name | Mandatory | Values | Description
357 --------------- | --------- | ------ | -----------
358 id | yes | string | The identifier of the peer. Facilitates referring to this peer.
359 speed | yes | int | See the description of the ``host`` tag for this attribute
360 bw_in | yes | int | Bandwidth of the private downstream link
361 bw_out | yes | int | Bandwidth of the private upstream link
362 coordinates | no | string | Coordinates of the gateway for this peer. Example value: 12.8 14.4 6.4
363 sharing_policy | no | SHARED\|SPLITDUPLEX (default: SPLITDUPLEX) | Sharing policy for links. See <b>link</b> description for details.
364 availability_file| no | string | Availability file for the peer. Same as host availability file. See <b>host</b> description for details.
365 state_file | no | string | State file for the peer. Same as host state file. See <b>host</b> description for details.
368 The communication latency between an host A=(xA,yA,zA) and an host
369 B=(xB,yB,zB) is computed as follows:
371 latency = sqrt( (xA-xB)² + (yA-yB)² ) + zA + zB
373 See the documentation of simgrid::kernel::routing::VivaldiZone for
374 details on how the latency is computed from the coordinate, and on the
375 the up and down bandwidth are used.
377 \subsection pf_ne Network equipments
379 There are two tags at all times available to represent network entities and
380 several other tags that are available only in certain contexts.
381 1. ``<link>``: Represents a entity that has a limited bandwidth, a
382 latency, and that can be shared according to TCP way to share this
385 The concept of links in SimGrid may not be intuitive, as links are not
386 limited to connecting (exactly) two entities; in fact, you can have more than
387 two equipments connected to it. (In graph theoretical terms: A link in
388 SimGrid is not an edge, but a hyperedge)
390 2. ``<router/>``: Represents an entity that a message can be routed
391 to, but that is unable to execute any code. In SimGrid, routers have also
392 no impact on the performance: Routers do not limit any bandwidth nor
393 do they increase latency. As a matter of fact, routers are (almost) ignored
394 by the simulator when the simulation has begun.
396 3. ``<backbone/>``: This tag is only available when the containing network zone is
397 used as a cluster (i.e., mode="Cluster")
400 If you want to represent an entity like a switch, you must use ``<link>`` (see section). Routers are used
401 to run some routing algorithm and determine routes (see Section \ref pf_routing for details).
403 \subsubsection pf_router <router/>
405 As said before, <b>router</b> is used only to give some information
406 for routing algorithms. So, it does not have any attributes except :
410 Attribute name | Mandatory | Values | Description
411 --------------- | --------- | ------ | -----------
412 id | yes | string | The identifier of the router to be used when referring to it.
413 coordinates | no | string | Must be provided when choosing the Vivaldi, coordinate-based routing model for the network zone the router belongs to. More details can be found in the Section \ref pf_P2P_tags.
418 <router id="gw_dc1_horizdist"/>
421 \subsubsection pf_tag_link <link>
423 Network links can represent one-hop network connections. They are
424 characterized by their id and their bandwidth; links can (but may not) be subject
429 Attribute name | Mandatory | Values | Description
430 --------------- | --------- | ------ | -----------
431 id | yes | string | The identifier of the link to be used when referring to it.
432 bandwidth | yes | int | Maximum bandwidth for this link, given in bytes/s
433 latency | no | double (default: 0.0) | Latency for this link.
434 sharing_policy | no | \ref sharing_policy_shared "SHARED"\|\ref pf_sharing_policy_fatpipe "FATPIPE"\|\ref pf_sharing_policy_splitduplex "SPLITDUPLEX" (default: SHARED) | Sharing policy for the link.
435 state | no | ON\|OFF (default: ON) | Allows you to to turn this link on or off (working / not working)
436 bandwidth_file | no | string | Allows you to use a file as input for bandwidth.
437 latency_file | no | string | Allows you to use a file as input for latency.
438 state_file | no | string | Allows you to use a file as input for states.
441 #### Possible shortcuts for ``latency`` ####
443 When using the latency attribute, you can specify the latency by using the scientific
444 notation or by using common abbreviations. For instance, the following three tags
448 <link id="LINK1" bandwidth="125000000" latency="5E-6"/>
449 <link id="LINK1" bandwidth="125000000" latency="5us"/>
450 <link id="LINK1" bandwidth="125000000" latency="0.000005"/>
453 Here, the second tag uses "us", meaning "microseconds". Other shortcuts are:
455 Name | Abbreviation | Time (in seconds)
456 ---- | ------------ | -----------------
457 Week | w | 7 * 24 * 60 * 60
458 Day | d | 24 * 60 * 60
462 Millisecond | ms | 0.001 = 10^(-3)
463 Microsecond | us | 0.000001 = 10^(-6)
464 Nanosecond | ns | 0.000000001 = 10^(-9)
465 Picosecond | ps | 0.000000000001 = 10^(-12)
467 #### Sharing policy ####
469 \anchor sharing_policy_shared
470 By default a network link is \b SHARED, i.e., if two or more data flows go
471 through a link, the bandwidth is shared fairly among all data flows. This
472 is similar to the sharing policy TCP uses.
474 \anchor pf_sharing_policy_fatpipe
475 On the other hand, if a link is defined as a \b FATPIPE,
476 each flow going through this link will be provided with the complete bandwidth,
477 i.e., no sharing occurs and the bandwidth is only limiting each flow individually.
478 Please note that this is really on a per-flow basis, not only on a per-host basis!
479 The complete bandwidth provided by this link in this mode
480 is ``number_of_flows*bandwidth``, with at most ``bandwidth`` being available per flow.
482 Using the FATPIPE mode allows to model backbones that won't affect performance
485 \anchor pf_sharing_policy_splitduplex
486 The last mode available is \b SPLITDUPLEX. This means that SimGrid will
487 automatically generate two links (one carrying the suffix _UP and the other the
488 suffix _DOWN) for each ``<link>`` tag. This models situations when the direction
489 of traffic is important.
492 Transfers from one side to the other will interact similarly as
493 TCP when ACK returning packets circulate on the other direction. More
494 discussion about it is available in the description of link_ctn description.
496 In other words: The SHARED policy defines a physical limit for the bandwidth.
497 The FATPIPE mode defines a limit for each application,
498 with no upper total limit.
501 Tip: By using the FATPIPE mode, you can model big backbones that
502 won't affect performance (except latency).
507 <link id="SWITCH" bandwidth="125000000" latency="5E-5" sharing_policy="FATPIPE" />
510 #### Expressing dynamism and failures ####
512 Similar to hosts, it is possible to declare links whose state, bandwidth
513 or latency changes over time (see Section \ref pf_host_dynamism for details).
515 In the case of network links, the ``bandwidth`` and ``latency`` attributes are
516 replaced by the ``bandwidth_file`` and ``latency_file`` attributes.
517 The following XML snippet demonstrates how to use this feature in the platform
518 file. The structure of the files "link1.bw" and "link1.lat" is shown below.
521 <link id="LINK1" state_file="link1.fail" bandwidth="80000000" latency=".0001" bandwidth_file="link1.bw" latency_file="link1.lat" />
525 Even if the syntax is the same, the semantic of bandwidth and latency
526 trace files differs from that of host availability files. For bandwidth and
527 latency, the corresponding files do not
528 express availability as a fraction of the available capacity but directly in
529 bytes per seconds for the bandwidth and in seconds for the latency. This is
530 because most tools allowing to capture traces on real platforms (such as NWS)
531 express their results this way.
533 ##### Example of "link1.bw" file #####
541 In this example, the bandwidth changes repeatedly, with all changes
542 being repeated every 12 seconds.
544 At the beginning of the the simulation, the link's bandwidth is 80,000,000
545 B/s (i.e., 80 Mb/s); this value was defined in the XML snippet above.
546 After four seconds, it drops to 40 Mb/s (line 2), and climbs
547 back to 60 Mb/s after another 4 seconds (line 3). The value does not change any
548 more until the end of the period, that is, after 12 seconds have been simulated).
549 At this point, periodicity kicks in and this behavior is repeated: Seconds
550 12-16 will experience 80 Mb/s, 16-20 40 Mb/s etc.).
552 ##### Example of "link1.lat" file #####
561 In this example, the latency varies with a period of 5 seconds.
562 In the xml snippet above, the latency is initialized to be 0.0001s (100µs). This
563 value will be kept during the first second, since the latency_file contains
564 changes to this value at second one, two and three.
565 At second one, the value will be 0.001, i.e., 1ms. One second later it will
566 be adjusted to 0.01 (or 10ms) and one second later it will be set again to 1ms. The
567 value will not change until second 5, when the periodicity defined in line 1
568 kicks in. It then loops back, starting at 100µs (the initial value) for one second.
570 #### The ``<prop/>`` tag ####
572 Similar to the ``<host>`` tag, a link may also contain the ``<prop/>`` tag; see the host
573 documentation (Section \ref pf_tag_host) for an example.
576 \subsubsection pf_backbone <backbone/>
579 This tag is <b>only available</b> when the containing network zone uses the "Cluster" routing mode!
581 Using this tag, you can designate an already existing link to be a backbone.
583 Attribute name | Mandatory | Values | Description
584 --------------- | --------- | ------ | -----------
585 id | yes | string | Name of the link that is supposed to act as a backbone.
587 \subsection pf_storage Storage
590 This is a prototype version that should evolve quickly, hence this
591 is just some doc valuable only at the time of writing.
592 This section describes the storage management under SimGrid ; nowadays
593 it's only usable with MSG. It relies basically on linux-like concepts.
594 You also may want to have a look to its corresponding section in
595 @ref msg_file ; access functions are organized as a POSIX-like
598 \subsubsection pf_sto_conc Storage - Main Concepts
600 The storage facilities implemented in SimGrid help to model (and account for)
601 storage devices, such as tapes, hard-drives, CD or DVD devices etc.
602 A typical situation is depicted in the figure below:
604 \image html ./webcruft/storage_sample_scenario.png
605 \image latex ./webcruft/storage_sample_scenario.png "storage_sample_scenario" width=\textwidth
607 In this figure, two hosts called Bob and Alice are interconnected via a network
608 and each host is physically attached to a disk; it is not only possible for each host to
609 mount the disk they are attached to directly, but they can also mount disks
610 that are in a remote location. In this example, Bob mounts Alice's disk remotely
611 and accesses the storage via the network.
613 SimGrid provides 3 different entities that can be used to model setups
614 that include storage facilities:
616 Entity name | Description
617 --------------- | -----------
618 \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).
619 \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.
620 \ref pf_tag_mount "mount" | Must be wrapped by a \ref pf_tag_host tag; declares which storage(s) this host has mounted and where (i.e., the mountpoint).
623 \anchor pf_storage_content_file
624 ### %Storage Content File ###
626 In order to assess exactly how much time is spent reading from the storage,
627 SimGrid needs to know what is stored on the storage device (identified by distinct (file-)name, like in a file system)
628 and what size this content has.
631 The content file is never changed by the simulation; it is parsed once
632 per simulation and kept in memory afterwards. When the content of the
633 storage changes, only the internal SimGrid data structures change.
635 \anchor pf_storage_content_file_structure
636 #### Structure of a %Storage Content File ####
638 Here is an excerpt from two storage content file; if you want to see the whole file, check
639 the file ``examples/platforms/content/storage_content.txt`` that comes with the
642 SimGrid essentially supports two different formats: UNIX-style filepaths should
643 follow the well known format:
646 /lib/libsimgrid.so.3.6.2 12710497
650 /bin/simgrid_update_xml 5018
651 /bin/graphicator 66986
652 /bin/simgrid-colorizer 2993
657 Windows filepaths, unsurprisingly, use the windows style:
660 \Windows\avastSS.scr 41664
661 \Windows\bfsvc.exe 75264
662 \Windows\bootstat.dat 67584
663 \Windows\CoreSingleLanguage.xml 31497
665 \Windows\dchcfg64.exe 335464
666 \Windows\dcmdev64.exe 93288
670 The different file formats come at a cost; in version 3.12 (and most likely
671 in later versions, too), copying files from windows-style storages to unix-style
672 storages (and vice versa) is not supported.
674 \anchor pf_storage_content_file_create
675 #### Generate a %Storage Content File ####
677 If you want to generate a storage content file based on your own filesystem (or at least a filesystem you have access to),
678 try running this command (works only on unix systems):
681 find . -type f -exec ls -1s --block=1 {} \; 2>/dev/null | awk '{ print $2 " " $1}' > ./content.txt
684 \subsubsection pf_storage_entities The Storage Entities
686 These are the entities that you can use in your platform files to include
687 storage in your model. See also the list of our \ref pf_storage_example_files "example files";
688 these might also help you to get started.
690 \anchor pf_storage_entity_storage_type
691 #### \<storage_type\> ####
693 Attribute name | Mandatory | Values | Description
694 --------------- | --------- | ------ | -----------
695 id | yes | string | Identifier of this storage_type; used when referring to it
696 model | no | string | In the future, this will allow to change the performance model to use
697 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)
698 content | yes | string | Path to a \ref pf_storage_content_file "Storage Content File" on your system. This file must exist.
700 This tag must contain some predefined model properties, specified via the <model_prop> tag. Here is a list,
701 see below for an example:
703 Property id | Mandatory | Values | Description
704 --------------- | --------- | ------ | -----------
705 Bwrite | yes | string | Bandwidth for write access; in B/s (but you can also specify e.g. "30MBps")
706 Bread | yes | string | Bandwidth for read access; in B/s (but you can also specify e.g. "30MBps")
709 A storage_type can also contain the <b><prop></b> tag. The <prop> tag allows you
710 to associate additional information to this <storage_type> and follows the
711 attribute/value schema; see the example below. You may want to use it to give information to
712 the tool you use for rendering your simulation, for example.
714 Here is a complete example for the ``storage_type`` tag:
716 <storage_type id="single_HDD" size="4000">
717 <model_prop id="Bwrite" value="30MBps" />
718 <model_prop id="Bread" value="100MBps" />
719 <prop id="Brand" value="Western Digital" />
723 @subsubsection pf_tag_storage <storage>
725 Attributes | Mandatory | Values | Description
726 -------------- | --------- | ------ | -----------
727 id | yes | string | Identifier of this ``storage``; used when referring to it
728 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.
729 attach | yes | string | Name of a host (see Section \ref pf_tag_host) to which this storage is <i>physically</i> attached to (e.g., a hard drive in a computer)
730 content | no | string | When specified, overwrites the content attribute of \ref pf_storage_entity_storage_type "\<storage_type\>"
732 Here are two examples:
735 <storage id="Disk1" typeId="single_HDD" attach="bob" />
737 <storage id="Disk2" typeId="single_SSD"
738 content="content/win_storage_content.txt" />
741 The first example is straightforward: A disk is defined and called "Disk1"; it is
742 of type "single_HDD" (shown as an example of \ref pf_storage_entity_storage_type "\<storage_type\>" above) and attached
743 to a host called "bob" (the definition of this host is omitted here).
745 The second storage is called "Disk2", is still of the same type as Disk1 but
746 now specifies a new content file (so the contents will be different from Disk1)
747 and the filesystem uses the windows style; finally, it is attached to a second host,
748 called alice (which is again not defined here).
750 \subsubsection pf_tag_mount <mount>
752 | Attribute | Mandatory | Values | Description |
753 | ----------- | ----------- | -------- | ------------- |
754 | id | yes | string | Refers to a \ref pf_storage_entity_storage "<storage>" entity that will be mounted on that computer |
755 | name | yes | string | Path/location to/of the logical reference (mount point) of this disk
757 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)
758 is; this location is specified by the ``name`` attribute.
760 Here is a simple example, taken from the file ``examples/platform/storage.xml``:
763 <storage_type id="single_SSD" size="500GiB">
764 <model_prop id="Bwrite" value="60MBps" />
765 <model_prop id="Bread" value="200MBps" />
768 <storage id="Disk2" typeId="single_SSD"
769 content="content/win_storage_content.txt"
771 <storage id="Disk4" typeId="single_SSD"
772 content="content/small_content.txt"
775 <host id="alice" speed="1Gf">
776 <mount storageId="Disk2" name="c:"/>
779 <host id="denise" speed="1Gf">
780 <mount storageId="Disk2" name="c:"/>
781 <mount storageId="Disk4" name="/home"/>
785 This example is quite interesting, as the same device, called "Disk2", is mounted by
786 two hosts at the same time! Note, however, that the host called ``alice`` is actually
787 attached to this storage, as can be seen in the \ref pf_storage_entity_storage "<storage>"
788 tag. This means that ``denise`` must access this storage through the network, but SimGrid automatically takes
789 care of that for you.
791 Furthermore, this example shows that ``denise`` has mounted two storages with different
792 filesystem types (unix and windows). In general, a host can mount as many storage devices as
796 Again, the difference between ``attach`` and ``mount`` is simply that
797 an attached storage is always physically inside (or connected to) that machine;
798 for instance, a USB stick is attached to one and only one machine (where it's plugged-in)
799 but it can only be mounted on others, as mounted storage can also be a remote location.
801 ###### Example files #####
803 \verbinclude example_filelist_xmltag_mount
805 \subsubsection pf_storage_example_files Example files
807 Several examples were already discussed above; if you're interested in full examples,
808 check the the following platforms:
810 1. ``examples/platforms/storage.xml``
811 2. ``examples/platforms/remote_io.xml``
813 If you're looking for some examplary C code, you may find the source code
814 available in the directory ``examples/msg/io/`` useful.
816 \subsubsection pf_storage_examples_modelling Modelling different situations
818 The storage functionality of SimGrid is type-agnostic, that is, the implementation
819 does not presume any type of storage, such as HDDs/SSDs, RAM,
820 CD/DVD devices, USB sticks etc.
822 This allows the user to apply the simulator for a wide variety of scenarios; one
823 common scenario would be the access of remote RAM.
825 #### Modelling the access of remote RAM ####
827 How can this be achieved in SimGrid? Let's assume we have a setup where three hosts
828 (HostA, HostB, HostC) need to access remote RAM:
838 An easy way to model this scenario is to setup and define the RAM via the
839 \ref pf_storage_entity_storage "storage" and \ref pf_storage_entity_storage_type "storage type"
840 entities and attach it to a remote dummy host; then, every host can have their own links
841 to this host (modelling for instance certain scenarios, such as PCIe ...)
846 RAM - Dummy -- Host B
851 Now, if read from this storage, the host that mounts this storage
852 communicates to the dummy host which reads from RAM and
853 sends the information back.
856 \section pf_routing Routing
858 To achieve high performance, the routing tables used within SimGrid are
859 static. This means that routing between two nodes is calculated once
860 and will not change during execution. The SimGrid team chose to use this
861 approach as it is rare to have a real deficiency of a resource;
862 most of the time, a communication fails because the links experience too much
863 congestion and hence, your connection stops before the timeout or
864 because the computer designated to be the destination of that message
867 We also chose to use shortest paths algorithms in order to emulate
868 routing. Doing so is consistent with the reality: [RIP](https://en.wikipedia.org/wiki/Routing_Information_Protocol),
869 [OSPF](https://en.wikipedia.org/wiki/Open_Shortest_Path_First), [BGP](https://en.wikipedia.org/wiki/Border_Gateway_Protocol)
870 are all calculating shortest paths. They do require some time to converge, but
871 eventually, when the routing tables have stabilized, your packets will follow
874 \subsection pf_tag_zone <zone>
876 Before SimGrid v3.16, networking zones used to be called Autonomous
877 Systems, but this was misleading as zones may include other zones in a
878 hierarchical manner. If you find any remaining reference to network
879 zones, please report this as a bug.
881 Attribute | Value | Description
882 ----------- | ------------------------------------------------- | ----------------------------------------------
883 id | String (mandatory) | The identifier of this zone (must be unique)
884 routing | One of the existing routing algorithm (mandatory) | See Section \ref pf_rm for details.
888 <zone id="zone0" routing="Full">
889 <host id="host1" speed="1000000000"/>
890 <host id="host2" speed="1000000000"/>
891 <link id="link1" bandwidth="125000000" latency="0.000100"/>
892 <route src="host1" dst="host2"><link_ctn id="link1"/></route>
896 In this example, zone0 contains two hosts (host1 and host2). The route
897 between the hosts goes through link1.
899 \subsection pf_rm Routing models
901 For each network zone, you must define explicitly which routing model will
902 be used. There are 3 different categories for routing models:
904 1. \ref pf_routing_model_shortest_path "Shortest-path" based models: SimGrid calculates shortest
905 paths and manages them. Behaves more or less like most real life
907 2. \ref pf_routing_model_manual "Manually-entered" route models: you have to define all routes
908 manually in the platform description file; this can become
909 tedious very quickly, as it is very verbose.
910 Consistent with some manually managed real life routing.
911 3. \ref pf_routing_model_simple "Simple/fast models": those models offer fast, low memory routing
912 algorithms. You should consider to use this type of model if
913 you can make some assumptions about your network zone.
914 Routing in this case is more or less ignored.
916 \subsubsection pf_raf The router affair
918 Using routers becomes mandatory when using shortest-path based
919 models or when using the bindings to the ns-3 packet-level
920 simulator instead of the native analytical network model implemented
923 For graph-based shortest path algorithms, routers are mandatory, because these
924 algorithms require a graph as input and so we need to have source and
925 destination for each edge.
927 Routers are naturally an important concept ns-3 since the
928 way routers run the packet routing algorithms is actually simulated.
929 SimGrid's analytical models however simply aggregate the routing time
930 with the transfer time.
932 So why did we incorporate routers in SimGrid? Rebuilding a graph representation
933 only from the route information turns out to be a very difficult task, because
934 of the missing information about how routes intersect. That is why we
935 introduced routers, which are simply used to express these intersection points.
936 It is important to understand that routers are only used to provide topological
939 To express this topological information, a <b>route</b> has to be
940 defined in order to declare which link is connected to a router.
943 \subsubsection pf_routing_model_shortest_path Shortest-path based models
945 The following table shows all the models that compute routes using
946 shortest-paths algorithms are currently available in SimGrid. More detail on how
947 to choose the best routing model is given in the Section called \"\ref pf_routing_howto_choose_wisely\".
949 | Name | Description |
950 | --------------------------------------------------- | -------------------------------------------------------------------------- |
951 | \ref pf_routing_model_floyd "Floyd" | Floyd routing data. Pre-calculates all routes once |
952 | \ref pf_routing_model_dijkstra "Dijkstra" | Dijkstra routing data. Calculates routes only when needed |
953 | \ref pf_routing_model_dijkstracache "DijkstraCache" | Dijkstra routing data. Handles some cache for already calculated routes. |
955 All those shortest-path models are instanciated in the same way and are
956 completely interchangeable. Here are some examples:
958 \anchor pf_routing_model_floyd
963 <zone id="zone0" routing="Floyd">
965 <cluster id="my_cluster_1" prefix="c-" suffix=""
966 radical="0-1" speed="1000000000" bw="125000000" lat="5E-5"
967 router_id="router1"/>
969 <zone id="zone1" routing="None">
970 <host id="host1" speed="1000000000"/>
973 <link id="link1" bandwidth="100000" latency="0.01"/>
975 <zoneroute src="my_cluster_1" dst="zone1"
978 <link_ctn id="link1"/>
984 zoneroute given at the end gives a topological information: link1 is
985 between router1 and host1.
987 #### Example platform files ####
989 This is an automatically generated list of example files that use the Floyd
990 routing model (the path is given relative to SimGrid's source directory)
992 \verbinclude example_filelist_routing_floyd
994 \anchor pf_routing_model_dijkstra
997 #### Example platform files ####
999 This is an automatically generated list of example files that use the Dijkstra
1000 routing model (the path is given relative to SimGrid's source directory)
1002 \verbinclude example_filelist_routing_dijkstra
1006 <zone id="zone_2" routing="Dijkstra">
1007 <host id="zone_2_host1" speed="1000000000"/>
1008 <host id="zone_2_host2" speed="1000000000"/>
1009 <host id="zone_2_host3" speed="1000000000"/>
1010 <link id="zone_2_link1" bandwidth="1250000000" latency="5E-4"/>
1011 <link id="zone_2_link2" bandwidth="1250000000" latency="5E-4"/>
1012 <link id="zone_2_link3" bandwidth="1250000000" latency="5E-4"/>
1013 <link id="zone_2_link4" bandwidth="1250000000" latency="5E-4"/>
1014 <router id="central_router"/>
1015 <router id="zone_2_gateway"/>
1016 <!-- routes providing topological information -->
1017 <route src="central_router" dst="zone_2_host1"><link_ctn id="zone_2_link1"/></route>
1018 <route src="central_router" dst="zone_2_host2"><link_ctn id="zone_2_link2"/></route>
1019 <route src="central_router" dst="zone_2_host3"><link_ctn id="zone_2_link3"/></route>
1020 <route src="central_router" dst="zone_2_gateway"><link_ctn id="zone_2_link4"/></route>
1024 \anchor pf_routing_model_dijkstracache
1025 ### DijkstraCache ###
1027 DijkstraCache example:
1029 <zone id="zone_2" routing="DijkstraCache">
1030 <host id="zone_2_host1" speed="1000000000"/>
1032 (platform unchanged compared to upper example)
1035 #### Example platform files ####
1037 This is an automatically generated list of example files that use the DijkstraCache
1038 routing model (the path is given relative to SimGrid's source directory):
1040 Editor's note: At the time of writing, no platform file used this routing model - so
1041 if there are no example files listed here, this is likely to be correct.
1043 \verbinclude example_filelist_routing_dijkstra_cache
1045 \subsubsection pf_routing_model_manual Manually-entered route models
1047 | Name | Description |
1048 | ---------------------------------- | ------------------------------------------------------------------------------ |
1049 | \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. |
1051 \anchor pf_routing_model_full
1056 <zone id="zone0" routing="Full">
1057 <host id="host1" speed="1000000000"/>
1058 <host id="host2" speed="1000000000"/>
1059 <link id="link1" bandwidth="125000000" latency="0.000100"/>
1060 <route src="host1" dst="host2"><link_ctn id="link1"/></route>
1064 #### Example platform files ####
1066 This is an automatically generated list of example files that use the Full
1067 routing model (the path is given relative to SimGrid's source directory):
1069 \verbinclude example_filelist_routing_full
1071 \subsubsection pf_routing_model_simple Simple/fast models
1073 | Name | Description |
1074 | ---------------------------------------- | ------------------------------------------------------------------------------ |
1075 | \ref pf_routing_model_cluster "Cluster" | This is specific to the \ref pf_tag_cluster "<cluster/>" tag and should not be used by the user, as several assumptions are made. |
1076 | \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. |
1077 | \ref pf_routing_model_vivaldi "Vivaldi" | Perfect when you want to use coordinates. Also see the corresponding \ref pf_P2P_tags "P2P section" below. |
1079 \anchor pf_routing_model_cluster
1083 In this mode, the \ref pf_cabinet "<cabinet/>" tag is available.
1085 #### Example platform files ####
1087 This is an automatically generated list of example files that use the Cluster
1088 routing model (the path is given relative to SimGrid's source directory):
1090 \verbinclude example_filelist_routing_cluster
1092 \anchor pf_routing_model_none
1096 This model does exactly what it's name advertises: Nothing. There is no routing
1097 available within this model and if you try to communicate within the zone that
1098 uses this model, SimGrid will fail unless you have explicitly activated the
1099 \ref options_model_select_network_constant "Constant Network Model" (this model charges
1100 the same for every single communication). It should
1101 be noted, however, that you can still attach an \ref pf_tag_zoneroute "ZoneRoute",
1102 as is demonstrated in the example below:
1104 \verbinclude platforms/cluster_and_one_host.xml
1106 #### Example platform files ####
1108 This is an automatically generated list of example files that use the None
1109 routing model (the path is given relative to SimGrid's source directory):
1111 \verbinclude example_filelist_routing_none
1114 \anchor pf_routing_model_vivaldi
1117 For more information on how to use the [Vivaldi Coordinates](https://en.wikipedia.org/wiki/Vivaldi_coordinates),
1118 see also Section \ref pf_P2P_tags "P2P tags".
1120 Note that it is possible to combine the Vivaldi routing model with other routing models;
1121 an example can be found in the file \c examples/platforms/cloud.xml. This
1122 examples models a NetZone using Vivaldi that contains other NetZones that use different
1125 #### Example platform files ####
1127 This is an automatically generated list of example files that use the None
1128 routing model (the path is given relative to SimGrid's source directory):
1130 \verbinclude example_filelist_routing_vivaldi
1133 \subsection ps_dec Defining routes
1135 There are currently four different ways to define routes:
1137 | Name | Description |
1138 | ------------------------------------------------- | ----------------------------------------------------------------------------------- |
1139 | \ref pf_tag_route "route" | Used to define route between host/router |
1140 | \ref pf_tag_zoneroute "zoneRoute" | Used to define route between different zones |
1141 | \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. |
1142 | \ref pf_tag_bypassasroute "bypassZoneRoute" | Used in the same way as bypassRoute, but for zones |
1144 Basically all those tags will contain an (ordered) list of references
1145 to link that compose the route you want to define.
1147 Consider the example below:
1150 <route src="Alice" dst="Bob">
1151 <link_ctn id="link1"/>
1152 <link_ctn id="link2"/>
1153 <link_ctn id="link3"/>
1157 The route here from host Alice to Bob will be first link1, then link2,
1158 and finally link3. What about the reverse route? \ref pf_tag_route "Route" and
1159 \ref pf_tag_zoneroute "zoneroute" have an optional attribute \c symmetrical, that can
1160 be either \c YES or \c NO. \c YES means that the reverse route is the same
1161 route in the inverse order, and is set to \c YES by default. Note that
1162 this is not the case for bypass*Route, as it is more probable that you
1163 want to bypass only one default route.
1165 For an \ref pf_tag_zoneroute "zoneroute", things are just slightly more complicated, as you have
1166 to give the id of the gateway which is inside the zone you want to access ...
1167 So it looks like this:
1170 <zoneroute src="zone1" dst="zone2"
1171 gw_src="router1" gw_dst="router2">
1172 <link_ctn id="link1"/>
1176 gw == gateway, so when any message are trying to go from zone1 to zone2,
1177 it means that it must pass through router1 to get out of the zone, then
1178 pass through link1, and get into zone2 by being received by router2.
1179 router1 must belong to zone1 and router2 must belong to zone2.
1181 \subsubsection pf_tag_linkctn <link_ctn>
1183 This entity has only one purpose: Refer to an already existing
1184 \ref pf_tag_link "<link/>" when defining a route, i.e., it
1185 can only occur as a child of \ref pf_tag_route "<route/>"
1187 | Attribute name | Mandatory | Values | Description |
1188 | --------------- | --------- | ------ | ----------- |
1189 | id | yes | String | The identifier of the link that should be added to the route. |
1190 | direction | maybe | UP\|DOWN | If the link referenced by \c id has been declared as \ref pf_sharing_policy_splitduplex "SPLITDUPLEX", this indicates which direction the route traverses through this link: UP or DOWN. If you don't use SPLITDUPLEX, do not use this attribute or SimGrid will not find the right link.
1192 #### Example Files ####
1194 This is an automatically generated list of example files that use the \c <link_ctn/>
1195 entity (the path is given relative to SimGrid's source directory):
1197 \verbinclude example_filelist_xmltag_linkctn
1199 \subsubsection pf_tag_zoneroute <zoneRoute>
1201 The purpose of this entity is to define a route between two
1202 NetZones. Recall that all zones form a tree, so to connect two
1203 sibiling zones, you must give such a zoneRoute specifying the source
1204 and destination zones, along with the gateway in each zone (ie, the
1205 point to reach within that zone to reach the netzone), and the list of
1206 links in the ancestor zone to go from one zone to another.
1208 So, to go from an host \c src_host that is within zone \c src, to an
1209 host \c dst_host that is within \c dst, you need to:
1211 - move within zone \c src, from \c src_host to the specified \c gw_src;
1212 - traverse all links specified by the zoneRoute (they are supposed to be within the common ancestor);
1213 - move within zone \c dst, from \c gw_dst to \c dst_host.
1215 #### Attributes ####
1217 | Attribute name | Mandatory | Values | Description |
1218 | --------------- | --------- | ------ | ----------- |
1219 | src | yes | String | The identifier of the source zone |
1220 | dst | yes | String | See the \c src attribute |
1221 | gw_src | yes | String | The gateway that will be used within the src zone; this can be any \ref pf_tag_host "Host" or \ref pf_router "Router" defined within the src zone. |
1222 | gw_dst | yes | String | Same as \c gw_src, but with the dst zone instead. |
1223 | symmetrical | no | YES\|NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1228 <zone id="zone0" routing="Full">
1229 <cluster id="my_cluster_1" prefix="c-" suffix=".me"
1230 radical="0-149" speed="1000000000" bw="125000000" lat="5E-5"
1231 bb_bw="2250000000" bb_lat="5E-4"/>
1233 <cluster id="my_cluster_2" prefix="c-" suffix=".me"
1234 radical="150-299" speed="1000000000" bw="125000000" lat="5E-5"
1235 bb_bw="2250000000" bb_lat="5E-4"/>
1237 <link id="backbone" bandwidth="1250000000" latency="5E-4"/>
1239 <zoneroute src="my_cluster_1" dst="my_cluster_2"
1240 gw_src="c-my_cluster_1_router.me"
1241 gw_dst="c-my_cluster_2_router.me">
1242 <link_ctn id="backbone"/>
1244 <zoneroute src="my_cluster_2" dst="my_cluster_1"
1245 gw_src="c-my_cluster_2_router.me"
1246 gw_dst="c-my_cluster_1_router.me">
1247 <link_ctn id="backbone"/>
1252 \subsubsection pf_tag_route <route>
1254 The principle is the same as for
1255 \ref pf_tag_zoneroute "ZoneRoute": The route contains a list of links that
1256 provide a path from \c src to \c dst. Here, \c src and \c dst can both be either a
1257 \ref pf_tag_host "host" or \ref pf_router "router". This is mostly useful for the
1258 \ref pf_routing_model_full "Full routing model" as well as for the
1259 \ref pf_routing_model_shortest_path "shortest-paths" based models (as they require
1260 topological information).
1263 | Attribute name | Mandatory | Values | Description |
1264 | --------------- | --------- | ---------------------- | ----------- |
1265 | src | yes | String | The value given to the source's "id" attribute |
1266 | dst | yes | String | The value given to the destination's "id" attribute. |
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 A route in the \ref pf_routing_model_full "Full routing model" could look like this:
1274 <route src="Tremblay" dst="Bourassa">
1275 <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"/>
1279 A route in the \ref pf_routing_model_shortest_path "Shortest-Path routing model" could look like this:
1281 <route src="Tremblay" dst="Bourassa">
1286 You must only have one link in your routes when you're using them to provide
1287 topological information, as the routes here are simply the edges of the
1288 (network-)graph and the employed algorithms need to know which edge connects
1289 which pair of entities.
1291 \subsubsection pf_tag_bypassasroute bypasszoneroute
1293 As said before, once you choose
1294 a model, it (most likely; the constant network model, for example, doesn't) calculates routes for you. But maybe you want to
1295 define some of your routes, which will be specific. You may also want
1296 to bypass some routes defined in lower level zone at an upper stage:
1297 <b>bypasszoneroute</b> is the tag you're looking for. It allows to
1298 bypass routes defined between already defined between zone (if you want
1299 to bypass route for a specific host, you should just use byPassRoute).
1300 The principle is the same as zoneroute : <b>bypasszoneroute</b> contains
1301 list of links that are in the path between src and dst.
1303 #### Attributes ####
1305 | Attribute name | Mandatory | Values | Description |
1306 | --------------- | --------- | ---------------------- | ----------- |
1307 | src | yes | String | The value given to the source zone's "id" attribute |
1308 | dst | yes | String | The value given to the destination zone's "id" attribute. |
1309 | gw_src | yes | String | The value given to the source gateway's "id" attribute; this can be any host or router within the src zone |
1310 | gw_dst | yes | String | The value given to the destination gateway's "id" attribute; this can be any host or router within the dst zone|
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 <bypasszoneRoute src="my_cluster_1" dst="my_cluster_2"
1317 gw_src="my_cluster_1_router"
1318 gw_dst="my_cluster_2_router">
1319 <link_ctn id="link_tmp"/>
1323 This example shows that link \c link_tmp (definition not displayed here) directly
1324 connects the router \c my_cluster_1_router in the source cluster to the router
1325 \c my_cluster_2_router in the destination router. Additionally, as the \c symmetrical
1326 attribute was not given, this route is presumed to be symmetrical.
1328 \subsubsection pf_tag_bypassroute bypassRoute
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 zone at an upper stage :
1334 <b>bypassRoute</b> is the tag you're looking for. It allows to bypass
1335 routes defined between <b>host/router</b>. The principle is the same
1336 as route : <b>bypassRoute</b> contains list of links references of
1337 links that are in the path between src and dst.
1339 #### Attributes ####
1341 | Attribute name | Mandatory | Values | Description |
1342 | --------------- | --------- | ---------------------- | ----------- |
1343 | src | yes | String | The value given to the source zone's "id" attribute |
1344 | dst | yes | String | The value given to the destination zone's "id" attribute. |
1345 | symmetrical | no | YES \| NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1350 <bypassRoute src="host_1" dst="host_2">
1351 <link_ctn id="link_tmp"/>
1355 This example shows that link \c link_tmp (definition not displayed here) directly
1356 connects host \c host_1 to host \c host_2. Additionally, as the \c symmetrical
1357 attribute was not given, this route is presumed to be symmetrical.
1359 \subsection pb_baroex Basic Routing Example
1361 Let's say you have an zone named zone_Big that contains two other zone, zone_1
1362 and zone_2. If you want to make a host (h1) from zone_1 with another one
1363 (h2) from zone_2 then you'll have to proceed as follows:
1364 \li First, you have to ensure that a route is defined from h1 to the
1365 zone_1's exit gateway and from h2 to zone_2's exit gateway.
1366 \li Then, you'll have to define a route between zone_1 to zone_2. As those
1367 zone are both resources belonging to zone_Big, then it has to be done
1368 at zone_big level. To define such a route, you have to give the
1369 source zone (zone_1), the destination zone (zone_2), and their respective
1370 gateway (as the route is effectively defined between those two
1371 entry/exit points). Elements of this route can only be elements
1372 belonging to zone_Big, so links and routers in this route should be
1373 defined inside zone_Big. If you choose some shortest-path model,
1374 this route will be computed automatically.
1376 As said before, there are mainly 2 tags for routing :
1377 \li <b>zoneroute</b>: to define routes between two <b>zone</b>
1378 \li <b>route</b>: to define routes between two <b>host/router</b>
1380 As we are dealing with routes between zone, it means that those we'll
1381 have some definition at zone_Big level. Let consider zone_1 contains 1
1382 host, 1 link and one router and zone_2 3 hosts, 4 links and one router.
1383 There will be a central router, and a cross-like topology. At the end
1384 of the crosses arms, you'll find the 3 hosts and the router that will
1385 act as a gateway. We have to define routes inside those two zone. Let
1386 say that zone_1 contains full routes, and zone_2 contains some Floyd
1387 routing (as we don't want to bother with defining all routes). As
1388 we're using some shortest path algorithms to route into zone_2, we'll
1389 then have to define some <b>route</b> to gives some topological
1390 information to SimGrid. Here is a file doing it all :
1393 <zone id="zone_Big" routing="Dijkstra">
1394 <zone id="zone_1" routing="Full">
1395 <host id="zone_1_host1" speed="1000000000"/>
1396 <link id="zone_1_link" bandwidth="1250000000" latency="5E-4"/>
1397 <router id="zone_1_gateway"/>
1398 <route src="zone_1_host1" dst="zone_1_gateway">
1399 <link_ctn id="zone_1_link"/>
1402 <zone id="zone_2" routing="Floyd">
1403 <host id="zone_2_host1" speed="1000000000"/>
1404 <host id="zone_2_host2" speed="1000000000"/>
1405 <host id="zone_2_host3" speed="1000000000"/>
1406 <link id="zone_2_link1" bandwidth="1250000000" latency="5E-4"/>
1407 <link id="zone_2_link2" bandwidth="1250000000" latency="5E-4"/>
1408 <link id="zone_2_link3" bandwidth="1250000000" latency="5E-4"/>
1409 <link id="zone_2_link4" bandwidth="1250000000" latency="5E-4"/>
1410 <router id="central_router"/>
1411 <router id="zone_2_gateway"/>
1412 <!-- routes providing topological information -->
1413 <route src="central_router" dst="zone_2_host1"><link_ctn id="zone_2_link1"/></route>
1414 <route src="central_router" dst="zone_2_host2"><link_ctn id="zone_2_link2"/></route>
1415 <route src="central_router" dst="zone_2_host3"><link_ctn id="zone_2_link3"/></route>
1416 <route src="central_router" dst="zone_2_gateway"><link_ctn id="zone_2_link4"/></route>
1418 <link id="backbone" bandwidth="1250000000" latency="5E-4"/>
1420 <zoneroute src="zone_1" dst="zone_2"
1421 gw_src="zone_1_gateway"
1422 gw_dst="zone_2_gateway">
1423 <link_ctn id="backbone"/>
1428 \section pf_other Other tags
1430 The following tags can be used inside a \<platform\> tag even if they are not
1431 directly describing the platform:
1433 - @ref pf_tag_config passes configuration options, e.g. to change the network model;
1434 - @ref pf_tag_prop gives user-defined properties to various elements
1436 \subsection pf_tag_config <config>
1438 Adding configuration flags into the platform file is particularly
1439 useful when the described platform is best used with specific
1440 flags. For example, you could finely tune SMPI in your platform file directly.
1442 | Attribute | Values | Description |
1443 | ---------- | ------------------- | ---------------------------------------------- |
1444 | id | String (optional) | This optional identifier is ignored by SimGrid |
1446 * **Included tags:** @ref pf_tag_prop to specify a given configuration item (see @ref options).
1448 Any such configuration must be given at the very top of the platform file.
1453 <?xml version='1.0'?>
1454 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid.dtd">
1455 <platform version="4">
1457 <prop id="maxmin/precision" value="0.000010" />
1458 <prop id="cpu/optim" value="TI" />
1459 <prop id="network/model" value="SMPI" />
1460 <prop id="smpi/bw-factor" value="65472:0.940694;15424:0.697866;9376:0.58729" />
1463 <zone id="zone0" routing="Full">
1467 \subsection pf_tag_prop <prop>
1469 Defines a user-defined property, identified with a name and having a
1470 value. You can specify such properties to most kind of resources:
1471 @ref pf_tag_zone, @ref pf_tag_host, @ref pf_tag_storage,
1472 @ref pf_tag_cluster and @ref pf_tag_link. These values can be retrieved
1473 at runtime with MSG_zone_property() or simgrid::s4u::NetZone::property(),
1474 or similar functions.
1476 | Attribute | Values | Description |
1477 | --------- | ---------------------- | ----------------------------------------------------------------------------------------- |
1478 | id | String (mandatory) | Identifier of this property. Must be unique for a given property holder, eg host or link. |
1479 | value | String (mandatory) | Value of this property; The semantic is completely up to you. |
1481 * **Included tags:** none.
1486 <prop id="Operating System" value="Linux" />
1490 \subsection pf_trace trace and trace_connect
1492 Both tags are an alternate way to pass files containing information on
1493 availability, state etc. to an entity. (See also @ref howto_churn).
1494 Instead of referring to the file directly in the host, link, or
1495 cluster tag, you proceed by defining a trace with an id corresponding
1496 to a file, later a host/link/cluster, and finally using trace_connect
1497 you say that the file trace must be used by the entity.
1503 <zone id="zone0" routing="Full">
1504 <host id="bob" speed="1000000000"/>
1506 <trace id="myTrace" file="bob.trace" periodicity="1.0"/>
1507 <trace_connect trace="myTrace" element="bob" kind="POWER"/>
1511 The order here is important. \c trace_connect must come
1512 after the elements \c trace and \c host, as both the host
1513 and the trace definition must be known when \c trace_connect
1514 is parsed; the order of \c trace and \c host is arbitrary.
1517 #### \c trace attributes ####
1520 | Attribute name | Mandatory | Values | Description |
1521 | --------------- | --------- | ---------------------- | ----------- |
1522 | id | yes | String | Identifier of this trace; this is the name you pass on to \c trace_connect. |
1523 | 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. |
1524 | trace_periodicity | yes | String | This is the same as for \ref pf_tag_host "hosts" (see there for details) |
1526 Here is an example of trace when no file name is provided:
1529 <trace id="myTrace" periodicity="1.0">
1536 #### \c trace_connect attributes ####
1538 | Attribute name | Mandatory | Values | Description |
1539 | --------------- | --------- | ---------------------- | ----------- |
1540 | kind | no | HOST_AVAIL\|POWER\|<br/>LINK_AVAIL\|BANDWIDTH\|LATENCY (Default: HOST_AVAIL) | Describes the kind of trace. |
1541 | trace | yes | String | Identifier of the referenced trace (specified of the trace's \c id attribute) |
1542 | element | yes | String | The identifier of the referenced entity as given by its \c id attribute |
1544 \section pf_hints Hints, tips and frequently requested features
1546 Now you should know at least the syntax and be able to create a
1547 platform by your own. However, after having ourselves wrote some platforms, there
1548 are some best practices you should pay attention to in order to
1549 produce good platform and some choices you can make in order to have
1550 faster simulations. Here's some hints and tips, then.
1552 @subsection pf_hints_search Finding the platform example that you need
1554 Most platform files that we ship are in the @c examples/platforms
1555 folder. The good old @c grep tool can find the examples you need when
1556 wondering on a specific XML tag. Here is an example session searching
1557 for @ref pf_trace "trace_connect":
1560 % cd examples/platforms
1561 % grep -R -i -n --include="*.xml" "trace_connect" .
1562 ./two_hosts_platform_with_availability_included.xml:26:<trace_connect kind="SPEED" trace="A" element="Cpu A"/>
1563 ./two_hosts_platform_with_availability_included.xml:27:<trace_connect kind="HOST_AVAIL" trace="A_failure" element="Cpu A"/>
1564 ./two_hosts_platform_with_availability_included.xml:28:<trace_connect kind="SPEED" trace="B" element="Cpu B"/>
1565 ./two_hosts.xml:17: <trace_connect trace="Tremblay_power" element="Tremblay" kind="SPEED"/>
1568 \subsection pf_hint_generating How to generate different platform files?
1570 This is actually a good idea to search for a better platform file,
1571 that better fit the need of your study. To be honest, the provided
1572 examples are not representative of anything. They exemplify our XML
1573 syntax, but that's all. small_platform.xml for example was generated
1574 without much thought beyond that.
1576 The best thing to do when possible is to write your own platform file,
1577 that model the platform on which you run your code. For that, you
1578 could use <a href="https://gitlab.inria.fr/simgrid/platform-calibration">our
1579 calibration scripts</a>. This leads to very good fits between the
1580 platform, the model and the needs. The g5k.xml example resulted of
1581 such an effort, which also lead to <a href="https://github.com/lpouillo/topo5k/">an
1582 ongoing attempt</a> to automatically extract the SimGrid platform from
1583 the <a href="http://grid5000.fr/">Grid'5000</a> experimental platform.
1584 But it's hard to come up with generic models. Don't take these files
1585 too seriously. Actually, you should always challenge our models and
1586 your instanciation if the accuracy really matters to you (see <a
1587 href="https://hal.inria.fr/hal-00907887">this discussion</a>).
1589 But such advices only hold if you have a real platform and a real
1590 application at hand. It's moot for more abstract studies working on
1591 ideas and algorithms instead of technical artefacts. Well, in this
1592 case, there unfortunately is nothing better than this old and rusty
1593 <a href="http://pda.gforge.inria.fr/tools/download.html">simulacrum</a>.
1594 This project is dormant since over 10 years (and you will have to
1595 update the generated platforms with <tt>bin/simgrid_update_xml</tt> to
1596 use them), but that's the best we have for this right now....
1598 \subsection pf_zone_h Zone Hierarchy
1599 The network zone design allows SimGrid to go fast, because computing route is
1600 done only for the set of resources defined in the current zone. If you're using
1601 only a big zone containing all resource with no zone into it and you're
1602 using Full model, then ... you'll loose all interest into it. On the
1603 other hand, designing a binary tree of zone with, at the lower level,
1604 only one host, then you'll also loose all the good zone hierarchy can
1605 give you. Remind you should always be "reasonable" in your platform
1606 definition when choosing the hierarchy. A good choice if you try to
1607 describe a real life platform is to follow the zone described in
1608 reality, since this kind of trade-off works well for real life
1611 \subsection pf_exit_zone Exit Zone: why and how
1612 Users that have looked at some of our platforms may have notice a
1613 non-intuitive schema ... Something like that :
1617 <zone id="zone_4" routing="Full">
1618 <zone id="exitzone_4" routing="Full">
1619 <router id="router_4"/>
1621 <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"/>
1622 <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"/>
1623 <link id="4_1" bandwidth="2250000000" latency="5E-5"/>
1624 <link id="4_2" bandwidth="2250000000" latency="5E-5"/>
1625 <link id="bb_4" bandwidth="2250000000" latency="5E-4"/>
1626 <zoneroute src="cl_4_1"
1628 gw_src="c_4_1-cl_4_1_router"
1629 gw_dst="c_4_2-cl_4_2_router">
1630 <link_ctn id="4_1"/>
1631 <link_ctn id="bb_4"/>
1632 <link_ctn id="4_2"/>
1634 <zoneroute src="cl_4_1"
1636 gw_src="c_4_1-cl_4_1_router"
1638 <link_ctn id="4_1"/>
1639 <link_ctn id="bb_4"/>
1641 <zoneroute src="cl_4_2"
1643 gw_src="c_4_2-cl_4_2_router"
1645 <link_ctn id="4_2"/>
1646 <link_ctn id="bb_4"/>
1651 In the zone_4, you have an exitzone_4 defined, containing only one router,
1652 and routes defined to that zone from all other zone (as cluster is only a
1653 shortcut for an zone, see cluster description for details). If there was
1654 an upper zone, it would define routes to and from zone_4 with the gateway
1655 router_4. It's just because, as we did not allowed (for performances
1656 issues) to have routes from an zone to a single host/router, you have to
1657 enclose your gateway, when you have zone included in your zone, within an
1658 zone to define routes to it.
1660 \subsection pf_P2P_tags P2P or how to use coordinates
1661 SimGrid allows you to use some coordinated-based system, like vivaldi,
1662 to describe a platform. The main concept is that you have some peers
1663 that are located somewhere: this is the function of the
1664 <b>coordinates</b> of the \<peer\> or \<host\> tag. There's nothing
1665 complicated in using it, here is an example:
1668 <?xml version='1.0'?>
1669 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid.dtd">
1670 <platform version="4">
1672 <zone id="zone0" routing="Vivaldi">
1673 <host id="100030591" coordinates="25.5 9.4 1.4" speed="1.5Gf" />
1674 <host id="100036570" coordinates="-12.7 -9.9 2.1" speed="7.3Gf" />
1676 <host id="100429957" coordinates="17.5 6.7 18.8" speed="8.3Gf" />
1681 Coordinates are then used to calculate latency (in microseconds)
1682 between two hosts by calculating the distance between the two hosts
1683 coordinates with the following formula: distance( (x1, y1, z1), (x2,
1684 y2, z2) ) = euclidian( (x1,y1), (x2,y2) ) + abs(z1) + abs(z2)
1686 In other words, we take the euclidian distance on the two first
1687 dimensions, and then add the absolute values found on the third
1688 dimension. This may seem strange, but it was found to allow better
1689 approximations of the latency matrices (see the paper describing
1692 Note that the previous example defines a routing directly between hosts but it could be also used to define a routing between zone.
1693 That is for example what is commonly done when using peers (see Section \ref pf_peer).
1695 <?xml version='1.0'?>
1696 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid.dtd">
1697 <platform version="4">
1699 <zone id="zone0" routing="Vivaldi">
1700 <peer id="peer-0" coordinates="173.0 96.8 0.1" speed="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us"/>
1701 <peer id="peer-1" coordinates="247.0 57.3 0.6" speed="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us" />
1702 <peer id="peer-2" coordinates="243.4 58.8 1.4" speed="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us" />
1706 In such a case though, we connect the zone created by the <b>peer</b> tag with the Vivaldi routing mechanism.
1707 This means that to route between zone1 and zone2, it will use the coordinates of router_zone1 and router_zone2.
1708 This is currently a convention and we may offer to change this convention in the DTD later if needed.
1709 You may have noted that conveniently, a peer named FOO defines an zone named FOO and a router named router_FOO, which is why it works seamlessly with the <b>peer</b> tag.
1712 \subsection pf_routing_howto_choose_wisely Choosing wisely the routing model to use
1715 Choosing wisely the routing model to use can significantly fasten your
1716 simulation/save your time when writing the platform/save tremendous
1717 disk space. Here is the list of available model and their
1718 characteristics (lookup : time to resolve a route):
1720 \li <b>Full</b>: Full routing data (fast, large memory requirements,
1722 \li <b>Floyd</b>: Floyd routing data (slow initialization, fast
1723 lookup, lesser memory requirements, shortest path routing only).
1724 Calculates all routes at once at the beginning.
1725 \li <b>Dijkstra</b>: Dijkstra routing data (fast initialization, slow
1726 lookup, small memory requirements, shortest path routing only).
1727 Calculates a route when necessary.
1728 \li <b>DijkstraCache</b>: Dijkstra routing data (fast initialization,
1729 fast lookup, small memory requirements, shortest path routing
1730 only). Same as Dijkstra, except it handles a cache for latest used
1732 \li <b>None</b>: No routing (usable with Constant network only).
1733 Defines that there is no routes, so if you try to determine a
1734 route without constant network within this zone, SimGrid will raise
1736 \li <b>Vivaldi</b>: Vivaldi routing, so when you want to use coordinates
1737 \li <b>Cluster</b>: Cluster routing, specific to cluster tag, should
1740 \subsection pf_switch I want to describe a switch but there is no switch tag!
1742 Actually we did not include switch tag. But when you're trying to
1743 simulate a switch, assuming
1744 fluid bandwidth models are used (which SimGrid uses by default unless
1745 ns-3 or constant network models are activated), the limiting factor is
1746 switch backplane bandwidth. So, essentially, at least from
1747 the simulation perspective, a switch is similar to a
1748 link: some device that is traversed by flows and with some latency and
1749 so,e maximum bandwidth. Thus, you can simply simulate a switch as a
1751 can be connected to this "switch", which is then included in routes just
1755 \subsection pf_multicabinets I want to describe multi-cabinets clusters!
1757 You have several possibilities, as usual when modeling things. If your
1758 cabinets are homogeneous and the intercabinet network negligible for
1759 your study, you should just create a larger cluster with all hosts at
1762 In the rare case where your hosts are not homogeneous between the
1763 cabinets, you can create your cluster completely manually. For that,
1764 create an As using the Cluster routing, and then use one
1765 <cabinet> for each cabinet. This cabinet tag can only be used an
1766 As using the Cluster routing schema, and creating
1768 Be warned that creating a cluster manually from the XML with
1769 <cabinet>, <backbone> and friends is rather tedious. The
1770 easiest way to retrieve some control of your model without diving into
1771 the <cluster> internals is certainly to create one separate
1772 <cluster> per cabinet and interconnect them together. This is
1773 what we did in the G5K example platform for the Graphen cluster.
1775 \subsection pf_platform_multipath I want to express multipath routing in platform files!
1777 It is unfortunately impossible to express the fact that there is more
1778 than one routing path between two given hosts. Let's consider the
1779 following platform file:
1782 <route src="A" dst="B">
1785 <route src="B" dst="C">
1788 <route src="A" dst="C">
1793 Although it is perfectly valid, it does not mean that data traveling
1794 from A to C can either go directly (using link 3) or through B (using
1795 links 1 and 2). It simply means that the routing on the graph is not
1796 trivial, and that data do not following the shortest path in number of
1797 hops on this graph. Another way to say it is that there is no implicit
1798 in these routing descriptions. The system will only use the routes you
1799 declare (such as <route src="A" dst="C"><link_ctn
1800 id="3"/></route>), without trying to build new routes by aggregating
1803 You are also free to declare platform where the routing is not
1804 symmetrical. For example, add the following to the previous file:
1807 <route src="C" dst="A">
1813 This makes sure that data from C to A go through B where data from A
1814 to C go directly. Don't worry about realism of such settings since
1815 we've seen ways more weird situation in real settings (in fact, that's
1816 the realism of very regular platforms which is questionable, but
1817 that's another story).
