indeXus.Net Shared Cache - high-performance, distributed caching

Commented Issue: RegExRemove exit to early

ronischuetz — Wed, 20 Aug 2008 09:28:37 GMT

public override bool RegexRemove(string regularExpression)
{
foreach(Configuration.Client.IndexusServerSetting setting in this.ServersList)
{
using (IndexusMessage msg = new IndexusMessage())
{
msg.Hostname = setting.IpAddress;
msg.Key = RegexRemoveKey;
msg.Action = IndexusMessage.ActionValue.RegexRemove;
msg.Payload = Encoding.UTF8.GetBytes(regularExpression);
return CacheUtil.Remove(msg);
}
}
return false;
}
Comments: ** Comment from web user: ronischuetz **

same happens here:

msg.Key = MultiDelKey;
msg.Payload = Formatters.Serialization.BinarySerialize(splitter[server]);
if (CacheUtil.Remove(msg))
{
return;
}
else
{
// write something to log ??
}

Created Issue: RegExRemove exit to early

ronischuetz — Wed, 20 Aug 2008 09:22:29 GMT

New Post: Some details needed on Distributed and Replicated mode

ronischuetz — Sun, 17 Aug 2008 15:24:14 GMT

hi alex,

1) please check the following link for further information:
>> In Distributed mode - let's say we have three cache nodes - A, B & C.
Disptributed Caching: http://www.sharedcache.com/cms/distributed_caching.aspx
Replicated Caching: http://www.sharedcache.com/cms/replicated_caching.aspx
>> When Add is invoked - how cache defines to which node data would be added?
>> How cache 'knows' to retrieve it from this specific node afterward and not from two others which don't have it?
>> Probably cache node is chosen in both cases using some unique hash key for specific object being manipulated - is that right?
The way we work upon distributed caching is to calculate hashcode based on your key: http://sharedcachedocu.indexus.net/html/T_MergeSystem_Indexus_WinServiceCommon_Hashing_Hash.htm
Lets say we have 3 servers then ["YourKey".GetHashCode() % 3] thats the easiest way to find your key.

2) currently you are not able to manage this within configuration but its quite easy to implement this, you add it to your close server node receive the result that data has been added and
we distribute it on the background.

hope this answers all your questions.

regards,
roni

New Post: Is Shared Cache used in any real/production big scale systems?

ronischuetz — Sun, 17 Aug 2008 15:11:41 GMT

hi alex,

there are already various systems which are using shared cache in production. At my current working place we also use shared cache in various backend systems without any problem and soon we gone use it also on our frontend. checkout also the following link, I don't know if there is a english version for this but somebody else then me who described in which use case they use Shared Cache.
http://www.kreuzhofer.de/wordpress/2008/06/24/xbap-wcf-and-debugging-with-visual-studio-2008/

may I ask for what kind of system you plan to use shared cache.

regrads
roni

New Post: Some details needed on Distributed and Replicated mode

AlexPinsker — Sun, 17 Aug 2008 13:59:45 GMT

Two questions:
1) In Distributed mode - let's say we have three cache nodes - A, B & C.
    When Add is invoked - how cache defines to which node data would be added?
    How cache 'knows' to retrieve it from this specific node afterward and not from two others which don't have it?
     Probably cache node is chosen in both cases using some unique hash key for specific object being manipulated - is that right?

2) In Replicated mode - is it possible to specify that Add function should lock until data is replicated over all the nodes (or is it returns immediately and replication continues in the background)?

Best regards,
      Alex Pinsker
      http://alexpinsker.blogspot.com

New Post: Is Shared Cache used in any real/production big scale systems?

AlexPinsker — Sun, 17 Aug 2008 13:48:22 GMT

I found some info on it in previous posts:
http://www.codeplex.com/SharedCache/Thread/View.aspx?ThreadId=28947

Regards,
Alex Pinsker
http://alexpinsker.blogspot.com

New Post: Is Shared Cache used in any real/production big scale systems?

AlexPinsker — Sun, 17 Aug 2008 09:39:11 GMT

Can any one reference a real life big-scale system that uses Shard Cache?
Like livejournal.com that use memcached?

Regards,
Alex Pinsker
http://alexpinsker.blogspot.com

New Post: Session Management with Failover... can replicated mode do it?

jsight — Fri, 15 Aug 2008 14:43:05 GMT

Hi Roni,

Thank you for your reply. I don't believe that I'd have time to write the support, but I'll keep an eye on future versions of sharedcache for any further developments. Thanks for the tip on Project Velocity... it looks like it will work nicely once they have implemented failover support. I believe that is not in the current CTP1 release.

Thanks,
Jess

New Post: Session Management with Failover... can replicated mode do it?

ronischuetz — Wed, 13 Aug 2008 19:45:03 GMT

Hi Jess,

thanks for your post. currently we have created only an empty session state provider. we do not have any project which is using shared cache for session state. maybe you would like to have a look at velocity for this issue then it will take some time until we able to provide shared cache for session state.

may you are interested to implement it?

regards,
roni

New Post: Session Management with Failover... can replicated mode do it?

jsight — Wed, 13 Aug 2008 15:31:56 GMT

We have been investigating indexus.net for high performance session state management with failover. I've seen relatively little documentation on Indexus.net's session state support, as well as managing failover with the replicated cache.

Is this something that Indexus.net is capable of? What kind of common setups are people using to manage this?

Thanks,
Jess

Updated Wiki: Home

ronischuetz — Fri, 08 Aug 2008 06:38:55 GMT

Intro

high-performance, distributed memory object caching system, generic in nature, but intended to speeding up dynamic web and / or win applications by alleviating database load.

checkout our website: http://www.SharedCache.com - indeXus.Net Shared Cache

indeXus.Net Shared Cache goes NHibernate Ver. 1.2.1

Release: release 2.0.3.270

General Overview

indeXus.Net Shared Cache is high performance distributed and replication cache system build for .Net cache and enterprise application running in server farms.

indeXus.Net Shared Cache provides distributed replicated cache to minimize the load factor. It consists the usage of two or more servers in a farm. It's replicated all data within the cluster. The big plus is simple, you have all your cache nodes on all different servers. In case one of your servers get restarted, it will receive all nodes automatically from its parent.

indeXus.Net Shared Cache uses 100% managed code which is written .Net C#.

Why you should consider to use shared cache?

There is no more efficient way to increase the scalable performance of applications then the use caching to unload deeper layers.

indeXus.Net Shared Cache Topologies

indeXus.Net Shared Cache provides a rich set of different scenarios to let you pick the option which suits your requirements. Connectivity between nodes happens on IP Layer 4 - Transport.

Distributed Caching - partitioned

your requirement:

extreme scalability

our solution:

shared-nothing architecture, automatically partition data across all cluster members

your result:

linear scalability. by partitioning the data evenly, the port-to-port throughput (the maximum amount of work that can be performed by each server) remains constant as servers are added, up to the extent of the switched fabric.

your benefits:

Partitioned: the size of the cache and the processing power available grow linearly with the size of the cluster.
Load Balanced: the responsibility for managing the data is automatically load-balanced across the cluster.
Point-to-Point: the communication for the partitioned cache is all point-to-point enabling linear scalability.

the limits:

as the amount of members in your cluster ram.

your uses:

any ram size caches, scaling with the size of the cluster.

Replicated Caching

your requirement:

Extreme Performance

our solution:

All data will be fully replicated to all cluster members.
Through the usage of fully replicated data to all members of the cluster will achive your zero latency access and extreme performance requirements.

your result:

Through zero latency access. Since the data is replicated to each member, it is available in each member without network latency and without waiting time.

your benefits:

This provides the best / possible speed for accessing data in your cache

the limits:

Cost per update: Updating a replicated cache requires pushing the new version of the data to all other cluster members, which will limit scalability if there are a high frequency of updated per member
Cost per entry: the data is replicated to every cluster member, so CLR heap space is used on each member, which will impact performance for large caches

Single Instance Caching

your requirement:

build a scalable system architecture

our solution:

There are no limitations to grow as your system grows

your result:

Linear scalability by partitioning the data on your indeXus.Net Shared Cache instance, the port-to-port throughput (the maximum amount of work that can be performed by each server) remains constant as servers are added, up to the extent of the switched fabric.

your benefits:

This provides the best / possible speed for accessing data in your cache.
While you grow, it will not require any refactoring since everything happens over configuration

the limits:

The size of available RAM

General Information

The usage of indeXus.Net Shared Cache does NOT prevent you to go on and use System.Cache.

link

Created Issue: Unhardcode job_install batch files

jafin — Tue, 29 Jul 2008 05:26:19 GMT

The batch files on my system failed as my windows and program files existed on another drive.
Attached find replacements.

Commented Issue: Graceful Service Shutdown

bklippel — Tue, 15 Jul 2008 15:25:11 GMT

When I stop the service using the Services snap-in, I get this message:
---------------------------
Services
---------------------------
Windows could not stop the IndeXus.Net service on Local Computer.
The service did not return an error. This could be an internal Windows error or an internal service error.
If the problem persists, contact your system administrator.

In the event viewer, I see this error:

Failed to stop service. System.Threading.ThreadAbortException: Thread was being aborted.
at System.Threading.Thread.AbortInternal()
at System.Threading.Thread.Abort()
at MergeSystem.Indexus.WinService.CacheExpire.Dispose() in C:\Visual Studio 2008\Projects\SharedCache\MergeSystem.Indexus.WinService\CacheExpire.cs:line 152
at MergeSystem.Indexus.WinService.ServiceLogic.ShutDown() in C:\Visual Studio 2008\Projects\SharedCache\MergeSystem.Indexus.WinService\ServiceLogic.cs:line 145
at MergeSystem.Indexus.WinService.Indexus.StopService() in C:\Visual Studio 2008\Projects\SharedCache\MergeSystem.Indexus.WinService\Indexus.cs:line 246
at MergeSystem.Indexus.WinService.Indexus.OnStop() in C:\Visual Studio 2008\Projects\SharedCache\MergeSystem.Indexus.WinService\Indexus.cs:line 296
at System.ServiceProcess.ServiceBase.DeferredStop()
Comments: ** Comment from web user: bklippel **

I should also mention, for what it is worth, i'm user x64 Server 2003.

Commented Issue: Graceful Service Shutdown

bklippel — Tue, 15 Jul 2008 15:21:08 GMT

Also, when this happens, if you fail to notice and start the service again it can merely continue running, in an incomplete state I suspect because the log will start showing "Object reference not set to an instance of an object." line items and cache objects will no longer be expired or purged until the service process is forcibly terminated.

Commented Issue: Cache dependencies between keys

ronischuetz — Tue, 15 Jul 2008 11:21:20 GMT

I wonder if it would be possible to add cache-dependencies between keys.
Comments: ** Comment from web user: ronischuetz **

currently we only can suggest to work with regular expressions - hope this helps.

New Post: Cache dependencies rules between keys

ronischuetz — Tue, 15 Jul 2008 11:18:33 GMT

hi erwin,

thanks for your post.

We have this issue also tested to create dependencies and we ended up with regular expressions. One of the people told me that he
uses it quite extensivly now. Have a look at the test program how you can work with regular expressions - its quite easy and
very fast.

regards,
roni

New Post: Performance counters?

ronischuetz — Tue, 15 Jul 2008 11:15:20 GMT

Hi Erwin,

thanks for your post. we have implemented a fully set of counters, but since we figured out that we lost always 5% per request we have take them out at all. therefore we do not plan to add them again.

most of your requested data you able to receive over stats - http://www.sharedcache.com/cms/shared_cache_stats.aspx
regards,
roni

New Post: Cache dependencies rules between keys

EBeckers — Sat, 12 Jul 2008 10:46:01 GMT

Would it be possible to add rules for dependencies between keys
Suppose i add the following items to the cache
- item#1
- item#2
- item #3
And suppose that item#3 is depended on item#1 and item#2,
This means that whenever item#1 or item#2 is updated/deleted that item#3 should automaticly be removed from the cache

If SharedCache would support something like this then it would save a lot of client-actions
since atm the clients needs to implement this behaviour and remove item#3 whenever item#1 or item#2 is updated/deleted

Erwin

New Post: Performance counters?

EBeckers — Sat, 12 Jul 2008 10:43:22 GMT

Could you please add performance counters for things like
- total # objects
- total cache size
- cache hitratio
- items added/sec
- items removed/sec
- item lookups/sec
- # of tcp connects /sec
- total # of tcp connects

btw a mmc (Microsoft Management Console) plugin would be nice too!

Erwin

Commented Issue: Cache dependencies between keys

EBeckers — Sat, 12 Jul 2008 09:04:07 GMT

I wonder if it would be possible to add cache-dependencies between keys.
Comments: ** Comment from web user: EBeckers **

Example suppose I add 3 items to the cache named item#1, item#2 and item#3. Now item#3 is depended on item#1 and item#2
so if item#1 OR item#2 gets changed/updated/deleted then item#3 should expire (removed from the cache) automaticly.

The possibility to define dependencies between cached items like this one would certainly be helpfull