Partitioning Cache Data
Why Partition
Partitioned Cost Reductions
Quorum in the network is 3. (This is down from the previous value of 4... find out why).
Any sort of partitioning strategy would need to be implemented at the titledb.xml level. Title AUs can be assigned to peers centrally, and each peer should receive a custom titledb.xml file. If LOCKSS is unwilling to support that then there is alternative using local parameters.
org.lockss.titleDbs = http://bpldb.bplonline.org/etc/adpn/titledb-local.xml
Implementing a 1 + 6 partitioning strategy can save 12% on average for each network node. 1 + 6 indicates AU owner + 6 additional network nodes. Adding 2 additional nodes to the network can decrease per node storage by an average of 30%. Adding 4 additional nodes and partitioning cache data can save per node storage on average of 41%. This means we could store up to 18 TB of data on 10.6 TB nodes.
Implementing a 1 + 5 which is 6 discrete nodes in the network (double quorum), base storage decrease is 25% with no additional nodes. 1 + 5 with 4 additional nodes achieves a staggering 50% on average per node storage reduction.
All nodes (default configuration)
au_host AUCount ContentSize in TB DiskCost in TB ADAH 778 5.11 5.13 AUB 778 5.11 5.13 BPL 778 5.11 5.13 SHC 778 5.11 5.13 TROY 778 5.11 5.13 UAB 778 5.11 5.13 UAT 778 5.11 5.13 UNA 778 5.11 5.13 40.88 41.04
Does not include vacated publisher AUs (which is between 500 and 600 GB).
1 + 6 no additional nodes
au_host AUCount ContentSize DiskCost count size cost ADAH 667 4.60 4.62 -14.27% -9.99% -9.97% AUB 676 3.83 3.85 -13.11% -25.00% -24.94% BPL 669 4.49 4.51 -14.01% -12.19% -12.15% SHC 666 4.48 4.49 -14.40% -12.41% -12.41% TROY 666 4.29 4.30 -14.40% -16.13% -16.09% UAB 667 4.58 4.60 -14.27% -10.29% -10.26% UAT 767 5.06 5.09 -1.41% -0.89% -0.86% UNA 667 4.43 4.45 -14.27% -13.27% -13.23% 35.76 35.91 -12.52% -12.49%
1 + 6 with 2 additional nodes
au_host AUCount ContentSize DiskCost count size cost ADAH 519 3.93 3.95 -33.29% -23.00% -22.98% AUB 540 3.21 3.23 -30.59% -37.15% -37.09% BPL 522 3.09 3.11 -32.90% -39.49% -39.44% SHC 519 3.22 3.23 -33.29% -36.97% -36.96% TROY 519 3.26 3.28 -33.29% -36.15% -36.09% UAB 519 3.02 3.03 -33.29% -40.93% -40.89% UAT 751 4.89 4.91 -3.47% -4.40% -4.38% UNA 519 3.26 3.28 -33.29% -36.15% -36.09% ADAH2 518 3.94 3.95 BPL2 519 3.94 3.95 35.76 35.91 -12.52% -12.49%
1 + 6 with 4 additional nodes
au_host AUCount ContentSize in TB DiskCost in TB ADAH 503 2.60 2.61 -35.35% -49.11% AUB 425 3.16 3.17 -45.37% -38.13% BPL 405 2.47 2.48 -47.94% -51.57% SHC 408 2.83 2.84 -47.56% -44.66% TROY 401 3.34 3.35 -48.46% -34.68% UAB 413 2.50 2.51 -46.92% -51.13% UAT 740 4.79 4.81 -4.88% -6.23% UNA 357 2.97 2.98 -54.11% -41.96% ADAH2 407 2.52 2.54 AUB2 501 3.25 3.26 BPL2 457 2.86 2.87 UAT2 428 2.48 2.49 35.76 35.91 -41.32% -39.68%
1 + 5 with no additional nodes
au_host AUCount ContentSize in TB DiskCost in TB ADAH 581 3.83 3.85 -25.32% -24.87% AUB 539 3.97 3.99 -30.72% -22.12% BPL 537 3.72 3.73 -30.98% -27.13% SHC 590 3.36 3.38 -24.16% -34.07% TROY 555 3.79 3.81 -28.66% -25.71% UAB 572 3.68 3.70 -26.48% -27.87% UAT 757 4.92 4.93 -2.70% -3.68% UNA 536 3.34 3.35 -31.11% -34.56% 30.65 30.78 -25.02% -25.00%
1 + 5 with 4 additional nodes
au_host AUCount ContentSize in TB DiskCost in TB ADAH 395 1.95 1.96 -49.23% -61.80% AUB 349 2.90 2.91 -55.14% -43.30% BPL 463 2.77 2.78 -40.49% -45.74% SHC 350 2.94 2.95 -55.01% -42.46% TROY 327 1.83 1.84 -57.97% -64.18% UAB 331 2.10 2.11 -57.46% -58.88% UAT 736 4.85 4.87 -5.40% -5.02% UNA 312 2.37 2.38 -59.90% -53.55% ADAH2 332 2.06 2.07 AUB2 328 2.52 2.53 BPL2 367 2.53 2.54 UAT2 377 1.83 1.84 30.65 30.78 -47.57% -46.87%
Distribution Algorithms
TBD...
Sample data was run using a static array of nodes, "randomized" with a Fisher-Yates shuffle, and least used node put in first array position.
string[] _nodes = { "ADAH", "AUB", "UAT", "UAB", "BPL", "UNA", "TROY", "SHC" };
Shuffle(_nodes); // Fisher-Yates Shuffle
//string _leastUsedNode = GetLeastUsedNodeByAUCount();
//string _leastUsedNode = GetLeastUsedNodeByContentSize();
string _leastUsedNode = GetLeastUsedNodeByDiskCost();
// put at beginning of array
Swap(_nodes, _leastUsedNode);
int _counter = 0;
// insert 6 5
foreach (string _node in _nodes)
{
if (_counter == 5) break;
if (_node.Equals(_row["au_owner"].ToString())) continue;
// process
counter++;
}
Storage Calculator
http://www.ibeast.com/content/tools/RaidCalc/RaidCalc.asp RAID Calculator
8 Disks * 3072 GB + RAID 5 = 20027.16 GB
Partition Implementation
Title List
The title list is the most crucial component for any partitioning. With close to 1000 AUs management of partition responsibilities would be cumbersome at the node level.
<lockss-config>
<property name="org.lockss.titleSet">
<property name="Birmingham Public Library">
<property name="name" value="All Birmingham Public Library AUs" />
<property name="class" value="xpath" />
<property name="xpath" value="[attributes/publisher='Birmingham Public Library']" />
</property>
</property>
<property name="org.lockss.title">
<property name="BirminghamPublicLibraryBasePluginBirminghamPublicLibraryCartographyCollectionMaps000400000599">
<property name="attributes.publisher" value="Birmingham Public Library" />
<property name="journalTitle" value="Birmingham Public Library Cartography Collection" />
<property name="type" value="journal" />
<property name="title" value="Birmingham Public Library Cartography Collection: Maps (000400-000599)" />
<property name="plugin" value="org.bplonline.adpn.BirminghamPublicLibraryBasePlugin" />
<property name="param.1">
<property name="key" value="base_url" />
<property name="value" value="http://bpldb.bplonline.org/adpn/load/" />
</property>
<property name="param.2">
<property name="key" value="group" />
<property name="value" value="Cartography" />
</property>
<property name="param.3">
<property name="key" value="collection" />
<property name="value" value="000400-000599" />
</property>
</property>
</property>
</lockss-config>
Comprehending LOCKSS Title List
Nested same name elements with different levels of nesting depth causes some difficulty in comprehension using standard tools. Standard deserialization techniques won't work because group 1 property element collection (org.lockss.titleSet) has different depth than group 2 property element collection (org.lockss.title).
protected void DeserializeXml() {
using (FileStream _f = new FileStream(Server.MapPath(@"titledb.xml"), FileMode.Open, FileAccess.Read))
{
XmlReaderSettings _sets = new XmlReaderSettings();
_sets.IgnoreWhitespace = true;
_sets.ProhibitDtd = false;
XmlReader _xml = XmlReader.Create(_f, _sets);
XmlSerializer _xs = new XmlSerializer(typeof(LockssTitleDb));
LockssTitleDb _titles = (LockssTitleDb)_xs.Deserialize(_xml);
for (int i = 0; i < _titles.LockssTitleSet.Length; i++)
{
if (i % 2 == 0)
{
// first group property (has attribute name=org.lockss.titleSet)
// titleSet displays publisher detail
// this assumes group of 2 for each publisher and AU list
}
else
{
// second group property (has attribute name=org.lockss.title)
// each iteration is a new AU for group 1 publisher
for (int j = 0; j < _titles.LockssTitleSet[i].ChildNodes.Count; j++)
{
// property name = normalized AU string
// outer AU definition
foreach (XmlNode _node in _titles.LockssTitleSet[i].ChildNodes[j].ChildNodes)
{
// loop through property tags
}
}
}
}
_f.close();
}
...
[XmlRoot("lockss-config")]
public class LockssTitleDb
{
// cannot create a single interface for all nesting depths with a single name
[XmlAnyElement()]
public XmlElement[] LockssTitleSet;
}
Local Data Store
http://bpldb.bplonline.org/images/adpn/datastore.png
XML Generation
ExpertConfig option org.lockss.titlesDb does not examine content type. It only looks at the URL ending and string matches .xml, else it assumes it is a .txt configuration file. See BaseConfigFile.java constructor.
Title Distribution
Base Distribution
Using the local data store, au_ids are auto-incremented. After the initial distribution, new AU releases can be isolated by selecting from last known Id. Row timestamp could also be used.
string[] _peers = GetPeerArray();
foreach (DataRow _row in _titles.Rows)
{
Shuffle(_peers);
int _counter = 0;
bool _isPeer = IsPeer(_row["au_pub_id"].ToString());
int _maxCount = _isPeer ? 5 : 6;
// not every publisher is a peer
// insert 6 5
foreach (string _peer in _peers)
{
if (_counter == _maxCount) break;
if (_peer.Equals(_row["au_pub_id"].ToString())) continue;
_connect.Command.Parameters.Clear();
_connect.Command.CommandText = "INSERT INTO `adpn_peer_titles` (`peer_id`, `au_id`) VALUES (?,?)";
_connect.Command.Parameters.AddWithValue("?", _peer);
_connect.Command.Parameters.AddWithValue("?", _row["au_id"].ToString());
_connect.Command.ExecuteNonQuery();
_counter++;
}
if (_isPeer)
{
// insert 1
_connect.Command.Parameters.Clear();
_connect.Command.CommandText = "INSERT INTO `adpn_peer_titles` (`peer_id`, `au_id`) VALUES (?,?)";
_connect.Command.Parameters.AddWithValue("?", _row["au_pub_id"].ToString());
_connect.Command.Parameters.AddWithValue("?", _row["au_id"].ToString());
_connect.Command.ExecuteNonQuery();
}
}
A resultant table set could look like the following:
peer_id count(*) % burden ADAH 499 55% AUB 549 61% BPL 520 57% SHC 504 56% SHC1 490 54% TROY 492 54% UAB 532 59% UAT 852 94% UAT1 482 53% UNA 510 56% 905 Titles
New Node Reshuffle
The original distribution algorithm can be used with a slight modification, only update rows in the adpn_peer_titles table when new node is up after shuffle. Assuming the new node is not considered a AU owner for existing AUs. Will reduce current AU burden on non-owner peers by maintaining a consistent node count per AU but replacing extant with new node.
string[] _peers = GetPeerArray();
foreach (DataRow _row in _titles.Rows)
{
Shuffle(_peers);
int _counter = 0;
bool _isPeer = IsPeer(_row["au_pub_id"].ToString());
int _maxCount = _isPeer ? 5 : 6;
// not every publisher is a peer
// insert 6 5
foreach (string _peer in _peers)
{
if (_counter == _maxCount) break;
if (_peer.Equals(_row["au_pub_id"].ToString())) continue;
// ONLY modify existing AU map when new node is up
if (_peer.Equals(newNodePeerId))
{
_connect.Command.Parameters.Clear();
_connect.Command.CommandText = @"
UPDATE `adpn_peer_titles` AS `a`,
(SELECT `b`.`peer_id`
FROM `adpn_peer_titles` AS `b`
WHERE `b`.`au_id` = ?
AND `b`.`peer_id` != ?
ORDER BY RAND() LIMIT 1) AS `b`
SET `a`.`peer_id` = ?
WHERE `a`.`peer_id` = `b`.`peer_id`
AND `a`.`au_id` = ? ";
_connect.Command.Parameters.AddWithValue("?", _row["au_id"].ToString());
_connect.Command.Parameters.AddWithValue("?", _row["au_pub_id"].ToString());
_connect.Command.Parameters.AddWithValue("?", _peer);
_connect.Command.Parameters.AddWithValue("?", _row["au_id"].ToString());
_connect.Command.ExecuteNonQuery();
}
}
}
Using the original distribution and reshuffling for a new node resulted in :
peer_id count(*) % burden ADAH 456 50% AUB 492 54% BPL 467 52% SHC 456 50% SHC1 438 48% TROY 445 49% UAB 466 51% UAT 845 93% UAT1 441 49% UNA 467 52% BPL1 457 50% 905 titles
Dead Node Reshuffle
When a node is no longer a part of the network:
- Delete peer_id from apdn_peer_titles
- Select a list of peer_ids and au_ids where count is less than maxcount
- Insert a shuffled peer id where not in peer_ids select list