L3 topology diagram with a large number of networks

I am working on redesigning a network that is using a single class B network for there LAN workstations and servers. I proposed for them to separate the two and put all of their servers into one class C network and the workstation in another class C. They stated that they wanted 14 separate networks for the workstations, different for section of the building, and 28 for the servers. The reason for the 28 is due to the fact they didn't keep track of how they used their address space and they do not want to renumber servers due to firewall rules. I have 50 networks hanging off of the collapsed core switch "WS-C6513-E" and finding room for all of these network is quite difficult. Does anyone have any experience with creating network topology diagrams with a large amount of networks hanging off of one device? I have attached a doc of what I have done.

Brandon
Sorry, you did say that in the last bit of your first post.
Last time i was doing topology diagrams i believe Visio had a pop up function where you could click over it but obviously if you are printing them out that does no good.
Unless you can summarise them to one entry with a note of how many subnets are being used within the summary i can't think of a way to do it. You could match the vlan number to the subnet (which you have partly done) and then just list the vlan numbers with the network prefix and make a note the subnet is derived from the vlan number.
Jon

Similar Messages

Problem in compilation with very large number of method parameters

I have java file which I created using WSDL2Java. Since the actual WSDL has a complex type with a large number of elements(around 600) in it, Consequently the resulting java file(from WSDL2Java) has a method that takes 600 parameters of various types. When I try to compile it using javac at command prompt, it says "Too many parameters" and doesn't compile. The same is compiling successfully using JBuilder X . The only way I could compile successfully at command prompt is by reducing the number of parameters to around 250 but unfortunately that it's not a workable solution. Does Sun specify any upper bound on number of parameters that can be passed to a method?

... a method that takes 600 parameters ...Not compatible with the spec, see Method Descriptors.
When I try to compile it using javac at
command prompt, it says "Too many parameters" and
doesn't compile.As it should.
The same is compiling successfully using JBuilder X .If JBuilder produces a class file, that class file may very well be invalid.
The only way I could compile
successfully at command prompt is by reducing the
number of parameters to around 250Which is what the spec says.
but unfortunately that it's not a workable solution.Pass an array of objects - an array is just one object.
Does Sun specify
any upper bound on number of parameters that can be
passed to a method?Yes.

Old files with a large number of comments no longer work properly in either Adobe Acrobat DC or or Acrobat Reader

I made a series of interactive ebooks wiith Adobe Acrobat Pro. They worked fine in the old Adobe Reader. But now the commenting feature does not work properly either in Adobe Reader DC or Adobe Acrobat DC. When I open the comments, everything slows down or else I get blank pages. The tools are not usable. There seems to be some compatibility issue between what was done in these books with the commenting tools and the new applications. Has anybody experienced the same?

Hi Anubha,
Thanks for your reply. With the new Acrobat Reader DC, I have the same problem on two computers, one with Windows 7 and one with Windows 8.1. I downloaded a trial version of Adobe Acrobat DC on the computer with 8,1. Then I found out that I had the same problem with Acrobat DC as I did with Reader DC.
I made these e-books with Acrobat Adobe Pro 11. I never had a problem with Pro 11 or with previous Readers. Also I have sent these e-books out to other people who had Reader 11 and they had no problems. Now it seems that the books are not usable – at least in the fullest way -- for me or for anyone else using the latest version of Reader.
Any help would be appreciated. I have put a lot of work into these e-books.
De : Anubha Goel
Envoyé : April-21-15 12:47 AM
À : michael jaegermeister
Objet : You have been mentioned by Anubha Goel in Re: old files with a large number of comments no longer work properly in either Adobe Acrobat DC or or Acrobat Reader in Adobe Community
You have been mentioned
by Anubha Goel <https://forums.adobe.com/people/Anubha+Goel?et=notification.mention> in Re: old files with a large number of comments no longer work properly in either Adobe Acrobat DC or or Acrobat Reader in Adobe Community - View Anubha Goel's reference to you <https://forums.adobe.com/message/7456900?et=notification.mention#7456900>

Reading a csv file with a large number of columns

Hello
I have been attempting to read data from large csv files with 38 columns by reading a line using readline and scanning the linebuffer using scan.
The file size can be up to 100 MB.
Scan does not seem support the large number of fields.
Any suggestions on reading the 38 comma separated fields. There is one header line in the file.
Thanks
Solved!
Go to Solution.

see if strtok() is useful http://www.elook.org/programming/c/strtok.html

JDev: af:table with a large number of rows

Hi
We are developing with JDeveloper 11.1.2.1. We have a VO that returns > 2.000.000 of rows and that we display in a af:table with access mode 'scrollable' (the default) and 'in Batches of' 101. The user can select one row and do CRUD operations in the VO with popups. The application works fine but I read that scroll very large number of rows is not a good idea because can cause OutOfMemory exception if the user uses the scroll bar many times. I have tried with access mode in 'Range Paging' but the application works in strange ways. Sometimes when I select a row to edit, if the selected row is the number 430 in the popup is show it the number 512 and when I want to insert a new row throws this exception:
oracle.jbo.InvalidOperException: JBO-25053: No se puede navegar con filas no enviadas en RangePaging RowSet.
     at oracle.jbo.server.QueryCollection.get(QueryCollection.java:2132)
     at oracle.jbo.server.QueryCollection.fetchRangeAt(QueryCollection.java:5430)
     at oracle.jbo.server.ViewRowSetIteratorImpl.scrollRange(ViewRowSetIteratorImpl.java:1329)
     at oracle.jbo.server.ViewRowSetIteratorImpl.setRangeStartWithRefresh(ViewRowSetIteratorImpl.java:2730)
     at oracle.jbo.server.ViewRowSetIteratorImpl.setRangeStart(ViewRowSetIteratorImpl.java:2715)
     at oracle.jbo.server.ViewRowSetImpl.setRangeStart(ViewRowSetImpl.java:3015)
     at oracle.jbo.server.ViewObjectImpl.setRangeStart(ViewObjectImpl.java:10678)
     at oracle.adf.model.binding.DCIteratorBinding.setRangeStart(DCIteratorBinding.java:3552)
     at oracle.adfinternal.view.faces.model.binding.RowDataManager._bringInToRange(RowDataManager.java:101)
     at oracle.adfinternal.view.faces.model.binding.RowDataManager.setRowIndex(RowDataManager.java:55)
     at oracle.adfinternal.view.faces.model.binding.FacesCtrlHierBinding$FacesModel.setRowIndex(FacesCtrlHierBinding.java:800)
     at weblogic.work.ExecuteThread.execute(ExecuteThread.java:209)
     at weblogic.work.ExecuteThread.run(ExecuteThread.java:178)
<LoopDiagnostic> <dump> [8261] variableIterator variables passivated >>> TrackQueryPerformed def
<LifecycleImpl> <_handleException> ADF_FACES-60098:El ciclo de vida de Faces recibe excepciones no tratadas en la fase RENDER_RESPONSE 6
What is the best way to display this amount of data in a af:table and do CRUD operations?
Thanks
Edited by: 972255 on 05/12/2012 09:51

Hi,
honestly, the best way is to provide users with an option to filter the result set displayed in the table to reduce the result set size. No-one will query 2.00.000 rows using the table scrollbar.
So one hint for optimization would be a query form (e.g. af:query)
To answer your question "srollable" vs. "page range", see
http://docs.oracle.com/cd/E21043_01/web.1111/b31974/bcadvvo.htm#ADFFD1179
Pay attention to what is written in the context of +"The range paging access mode is typically used for paging through read-only row sets, and often is used with read-only view objects.".+
Frank

How to design Storage Spaces with a large number of drives

I am wondering how one might go about designing a storage space for a large number of drives. Specifically I've got 45 x 4TB drives. As i am not extremely familiar with storage spaces, i'm a bit confused as to how I should go about designing this. Here is
how i would do it in hardware raid and i'd like to know how to best match this setup in Storage Spaces. I've been burned twice now by poorly designed storage spaces and i don't want to get burned again. I want to make sure if a drive fails, i'm able to properly
replace it without SS tossing it's cookies.
In the hardware raid world, i would divide these 45 x 4TB drives into three separate 15 disk Raid 6's. (Thus losing 6 drives to parity) Each raid 6 would show up as a separate volume/drive to the parent
OS. If any disk failed in any of the three raids, i would simply pull it out and put a new disk back in and the raid would rebuild itself.
Here is my best guess for storage spaces. I would create 3 separate storage pools each containing 15 disks. I would then create a separate
Dual Parity Virtual Disk for each pool. (Also losing 6 drives to parity) Each virtual disk would appear as a separate volume/disk
to the parent OS. Did i miss anything?
Additionally, is there any benefit to breaking up my 45 disks into 3 separate pools? Would it be better to create one giant pool with all 45 disks and then create 3 (or however many) virtual disks on top of that one pool?

I am wondering how one might go about designing a storage space for a large number of drives. Specifically I've got 45 x 4TB drives. As i am not extremely familiar with storage spaces, i'm a bit confused as to how I should go about designing this. Here is
how i would do it in hardware raid and i'd like to know how to best match this setup in Storage Spaces. I've been burned twice now by poorly designed storage spaces and i don't want to get burned again. I want to make sure if a drive fails, i'm able to properly
replace it without SS tossing it's cookies.
In the hardware raid world, i would divide these 45 x 4TB drives into three separate 15 disk Raid 6's. (Thus losing 6 drives to parity) Each raid 6 would show up as a separate volume/drive to the parent
OS. If any disk failed in any of the three raids, i would simply pull it out and put a new disk back in and the raid would rebuild itself.
Here is my best guess for storage spaces. I would create 3 separate storage pools each containing 15 disks. I would then create a separate
Dual Parity Virtual Disk for each pool. (Also losing 6 drives to parity) Each virtual disk would appear as a separate volume/disk
to the parent OS. Did i miss anything?
Additionally, is there any benefit to breaking up my 45 disks into 3 separate pools? Would it be better to create one giant pool with all 45 disks and then create 3 (or however many) virtual disks on top of that one pool?
1) Try to avoid parity and esp. double parity RAIDs with a typical VM workload. It's dominated by small reads (OK) and small writes (not OK as whole parity stripe gets updated with any "ready-modify-write" sequence). As a result writes would be DOG slow.
Another nasty parity RAID characteristic is very long rebuild times... It's pretty easy to get second (third with double parity) drive failure during re-build process and that would render the whole RAID set useless. Solution would be to use RAID10. Much safer,
faster to work and rebuild compared to RAID5/6 but wastes half of raw capacity...
2) Creating "islands" of storage is an extremely effective way of stealing IOPS away from your config. Typical modern RAID set would run out of IOPS long before running out of capacity so unless you're planning to have a file dump of an ice cold data or
CCTV storage you'll absolutely need all IOPS from all spindles @ the same time. This again means One Big RAID10, OBR10.
Hope this helped a bit :) Good luck!
StarWind VSAN [Virtual SAN] clusters Hyper-V without SAS, Fibre Channel, SMB 3.0 or iSCSI, uses Ethernet to mirror internally mounted SATA disks between hosts.

Status and messaging for systems with a large number of classes

Very common dilemma we coders face when creating
systems involving a large number of classes:
Any standard framework to take care of the global status of the whole application and of gui subsystems
for both status handling and report messages to users?
Something light if possible...and not too much threads.

Ah, I see,
I found JPanel with CardLayout or a JTabbedPane very good for control of several GUI in an application - as alternative organization tool I use a JTree, which is used for both, selecting and organizing certain tasks or data in the application - tasks are normally done with data associated with them (that is, what an object is for), so basically a click onto a node in this JTree invokes an interface method of that object (the userObject), which is associated with this node.
Event handling should be done by the event-handling-thread only as far as possible - it is responsible for it so leave this job to it. This will give you control over the order in which the events are handled. Sometimes it needs a bit more work to obey this rule - for example communication coming from the outside (think of a chat channel for example) must be converted to an event source driven by a thread normally. As soon as it is an event source, you can leave it's event handling to the event handling thread again and problems with concurrent programming are minimized again.
It is the same with manipulating components or models of components - leave it to the event handling thread using a Runnable and SwingUtilities.invokeLater(Runnable). This way you can be sure that each manipulation is done after the other in the order you have transferred it to the event handling thread.
When you do this consequently most of your threads will idle most of the time - so give them a break using Thread.sleep(...) - not all platforms provide preemptive multitasking and this way it is garanteed, that the event handling thread will get a chance to run most of the time - which results in fast GUI update and fast event handling.
Another thing is, that you should use "divide and conquer" also within a single GUI panel - place components in subpanels and transfer the responsibility for the components in this panel to exactly these subpanels - think of a team manager which makes his employees work together. He reports up to his super manager and transfers global order from his boss into specific tasks by delegation to the components, he is managing. If you have this in mind, when you design classes, you will have less problem - each class is responsible for a certain task - define it clearly and define to whom it is reporting (its listeners) and what these listeners may be interested in.
When you design the communication structure within your hierarchy of classes (directors, managers, team managers and workers) have in mind, that the communication structure should not break the management hierarchy. A director gives global orders to a manager, which delegates several tasks to the team managers, which make their workers do what is needed. This structure makes a big company controlable by directors - the same principles can also keep control within an application.
greetings Marsian

Possible bug for emails with a large number of recipients

When receiving emails which was also sent to a large number of people (more than can fit in two lines), it then displays the text 'and 10 more..." or however more there are. Clicking on this text is supposed to then display all the email addresses the email was sent to. However clicking on this is for some reason doing nothing. Is this a bug, or am I the only one having this problem?

When receiving emails which was also sent to a large number of people (more than can fit in two lines), it then displays the text 'and 10 more..." or however more there are. Clicking on this text is supposed to then display all the email addresses the email was sent to. However clicking on this is for some reason doing nothing. Is this a bug, or am I the only one having this problem?

Plotting an array with a large number of data points

Hello All,
I am collecting data every second for 24 Hrs. How do I plot this on a XY plot? I have the data collection setup in a while loop and my initial attempt was to use a shift register and a build array to make the array and that works OK. The problem is I do not know if this is a wise option in terms of memory usage. (I end up making 3 arrays and 3*86400=259200 data points in 3 shift registers).
What is the better way to plot 86400 data points in a XY plot? I am also writing the data to a file every second.
Thanks,
KB

Allan,
Thanks for the message. I am attaching with this message a ccpy of my vi where you can see that I am collecting data from 3 counters, writing them to a file and displaying them. Can you comment on this Vi with regard to memory management?
Yes, I would like to see your examples. Thanks once again.
KB
Attachments:
PC21.vi ‏418 KB

Have you had experience with a large number of queues in XI/PI?

We currently have an application that sends work orders using quality of service EOIO. The problem with this approach is that if one message fails, then no further messages are processed until the error is resolved or the message is cancelled. Additionally, throughput is affected since there is no opportunity for parallel processing.
We are considering changing the application so that rather than having one queue for all work orders, we would specify a separate queue for each work order. This would potentially create thousands of queues but since queues are a logical concept in XI there should be a negligible resource overhead.
Has anyone tried this? Are there any problems with the qRFC QIN Scheduler in the event of a mass failure?
We intend to run some tests but are looking for some words of wisdom from anyone else who has been there before.

hi,
check these links on controlling queues so that message processing and the available workprocesses to a queue is controlled.
https://www.sdn.sap.com/irj/servlet/prt/portal/prtroot/docs/library/uuid/59e837d3-0201-0010-c096-dc1869733413
https://www.sdn.sap.com/irj/servlet/prt/portal/prtroot/docs/library/uuid/70ada5ef-0201-0010-1f8b-c935e444b0ad
http://help.sap.com/saphelp_nw04/helpdata/en/04/827440c36ed562e10000000a155106/frameset.htm
Thanks,
Vijaya.

Pivot table with very large number of columns

Hello,
here is the situation:
One table that contains raw data; from this table I feed one with extract information (3 fields); I have to turn the content in a pivot table
Ro --- Co --- Va
A A 1
A B 1
A C 2
B A 11
Turned in
A B C...
A 1 1 2
B 11 null null
To do this I do a query like:
select r, sum(decode(c,'A',Va) COLA, sum(decode(c,'B',Va) COLB , sum(decode(c,'C',Va) COLC,.... sum(decode(c,'XYZ',Va) COLXYZ from table group by r
The statement is generated by a script (cfmx) and it works until I reach a query that try to have 672 values for c; which means 672 columns...
Oracle doesn't like that: ORA-01467: sort key too long
I like this way has it is getting the result fast.
I have tried different solution a the CFMX level with for that specific query, I got timeout (query table with loop on co within loop on ro)
Is there any work around?
I am using Oracle 9i.
Tahnk you!

insert into extracted_data select c, r, v, p from full_data where <specific_clause>
The values for C are from a query: select disctinct c from extracted_data
and it is the same for R
R and C are varchar2(3999)
I suppose that I can split on the first letter of the C column as:
SELECT r, low.cola, low.colb, . . ., low.colm,
high.coln, high.colo, . . ., high.colz
FROM (SELECT r, SUM(DECODE(c, 'A', va)) cola, . . .
SUM(DECODE(c, 'M', va)) colm
FROM table
WHERE c like 'A%'
GROUP BY r) Alpha_A,
(SELECT r, SUM(DECODE(c, 'N', va)) coln, . . .
SUM(DECODE(c, 'Z', va)) colz
FROM table
WHERE c like 'B%'
GROUP BY r) Alpha_B,
(SELECT r, SUM(DECODE(c, 'N', va)) coln, . . .
SUM(DECODE(c, 'Z', va)) colz
FROM table
WHERE c like 'C%'
GROUP BY r) Alpha_C
(SELECT r, SUM(DECODE(c, 'zN', va)) coln, . . .
SUM(DECODE(c, 'zZ', va)) colz
FROM table
WHERE c like 'Z%'
GROUP BY r) Alpha_Z
WHERE alpha_A.r = alpha_B.r and apha_a.r = alpha_C.r ... and alpha_a.r = alpha_z.r
I will have 27 select statement joined... I have to check if even like that I will not reach the limit within one of the statement select
"in real life"
select GRPW.r, GRPW.W0, GRPC.C0, GRPC.C1 from
(select r, sum(decode(C, 'Wall, unspecified',cases)) W0 from tmp_maqueje where upper(C) like 'W%' group by r) GRPW,
select r,
sum(decode(C, 'Ceramic tiles, indoors',cases)) C0,
sum(decode(C, 'Cement surface, outdoors (Concrete/cement block, see Structural element, A11)',cases)) C1
from tmp_maqueje where upper(C) like 'C%' group by r) GRPC
where GRPW.r = GRPC.r
order by GRPW.r, GRPW.W0, GRPC.C0, GRPC.C1
Message was edited by:
maquejp

Best practice for handling data for a large number of indicators

I'm looking for suggestions or recommendations for how to best handle a UI with a "large" number of indicators. By large I mean enough to make the block diagram quite large and ugly after the data processing for each indicator is added. The data must be "unpacked" and then decoded, e.g., booleans, offset binary bit fields, etc. The indicators are updated once/sec. I am leanding towards a method that worked well for me previously, that is, binding network shared variables to each indicator, then using several sub-vis to process the particular piece of data and write to the appropriate variables.
I was curious what others have done in similar circumstances.
Bill
“A child of five could understand this. Send someone to fetch a child of five.”
― Groucho Marx
Solved!
Go to Solution.

I can certainly feel your pain.
Note that's really what is going on in that png You can see the Action Engine responsible for updating the display to the far right.
In my own defence: the FP concept was presented to the client's customer before they had a person familliar with LabVIEW identified. So I worked it this way from no choice of mine. I knew it would get ugly before I walked in the door and chose to meet the challenge head on anyway. Defer Panel Updates was my very good friend. The sensors these objects represent were constrained to pass info via a single ZigBee network so I had the benefit of fairly low data rates as well but even changing view (Yes there is a display mode that swaps what information is displayed for each sensor) was updated fast enough that the user still got a responsive GUI.
(the GUI did scale poorly though! That is a lot of wires! I was greateful to Jack for the Idea to make align and distribute work on wires)
Jeff

How to handle a large number of query parameters for a Browse screen

I need to implement an advanced search functionality in a browse screen for a large table. The table has 80+ columns and therefore will have a large number of possible query parameters. The screen will be built on a modeled query with all
of the parameters marked as optional. Given the large number of parameters, I am thinking that it would be better to use a separate screen to receive the parameter input from the user, rather than a Popup. Is it possible for example to have a search
button on the browse screen (screen a) open a new screen (screen b) that contains all of the search parameters, have the user enter the parameters they want, then click a button to send all of the parameters back to screen a where the query is executed and
the search results are returned to the table control? This would effectively make screen b an advanced modal window for screen a. In addition, if the user were to execute the query, then want to change a parameter, they would need to be able to
re-open screen b and have all of their original parameters still set. How would you implement this, or otherwise deal with a large number of optional query parameters in the html client? My initial thinking is to store all of the parameters in
an object and use beforeShown/afterClosed to pass them between the screens, but I'm not quite sure how to make that work. TIA

Wow Josh, thanks. I have a lot of reading to do. What I ultimately plan to do with this (my other posts relate to this too), is have a separate screen for advanced filtering that also allows the user to save their queries if desired.
There is an excellent way to get at all of the query information in the Query_Executed() method. I just put an extra Boolean parameter in the query called "SaveQuery" and when true, the Query_Executed event triggers an entry into a table with
the query name, user name, and parameter value pairs that the user entered. Upon revisiting the screen, I want the user to be able to select from their saved queries and load all the screen parameters (screen properties) from their selected query.
I almost have it working. It may be as easy as marking all of the screen properties that are query parameters as screen parameters (not required), then passing them in from the saved query data (filtered by username, queryname, and selected
item). I'll post an update once I get it. Probably will have some more questions as I go through it. Thanks again!

Large number of JSP performance

Hi,
a colleague of me made tests with a large number of JSP and identified a
performance problem.
I believe I found a solution to his problem. I tested it with WLS 5.1 SP2
and SP3 and MS jview SDK 4.
The issue was related to the duration of the initial call of the nth JSP,
our situation as we are doing site hosting.
The solution is able to perform around 14 initial invocations/s no matter if
the invocation is the first one or the
3000th one and the throughput can go up to 108 JSPs/s when the JSP are
already loaded, the JSPs being the
snoopservlet example copied 3000 times.
The ratio have more interest than the values as the test machine (client and
WLS 5.1) was a 266Mhz laptop.
I repeat the post of Marc on 2/11/2000 as it is an old one:
Hi all,
I'm wondering if any of you has experienced performance issue whendeploying
a lot of JSPs.
I'm running Weblogic 4.51SP4 with performance pack on NT4 and Jdk1.2.2.
I deployed over 3000 JSPs (identical but with a distinct name) on myserver.
I took care to precompile them off-line.
To run my tests I used a servlet selecting randomly one of them and
redirecting the request.
getServletContext().getRequestDispatcher(randomUrl).forward(request,response);
The response-time slow-down dramaticaly as the number of distinct JSPs
invoked is growing.
(up to 100 times the initial response time).
I made some additional tests.
When you set the properties:
weblogic.httpd.servlet.reloadCheckSecs=-1
weblogic.httpd.initArgs.*.jsp=..., pageCheckSeconds=-1, ...
Then the response-time for a new JSP seems linked to a "capacity increase
process" and depends on the number of previously activated JSPs. If you
invoke a previously loaded page the server answers really fast with no
delay.
If you set previous properties to any other value (0 for example) the
response-time remains bad even when you invoke a previously loaded page.SOLUTION DESCRIPTION
Intent
The package described below is design to allow
* Fast invocation even with a large number of pages (which can be the case
with Web Hosting)
* Dynamic update of compiled JSP
Implementation
The current implementation has been tested with JDK 1.1 only and works with
MS SDK 4.0.
It has been tested with WLS 5.1 with service packs 2 and 3.
It should work with most application servers, as its requirements are
limited. It requires
a JSP to be able to invoke a class loader.
Principle
For a fast invocation, it does not support dynamic compilation as described
in the JSP
model.
There is no automatic recognition of modifications. Instead a JSP is made
available to
invalidate pages which must be updated.
We assume pages managed through this package to be declared in
weblogic.properties as
weblogic.httpd.register.*.ocg=ocgLoaderPkg.ocgServlet
This definition means that, when a servlet or JSP with a .ocg extension is
requested, it is
forwarded to the package.
It implies 2 things:
* Regular JSP handling and package based handling can coexist in the same
Application Server
instance.
* It is possible to extend the implementation to support many extensions
with as many
package instances.
The package (ocgLoaderPkg) contains 2 classes:
* ocgServlet, a servlet instantiating JSP objects using a class loader.
* ocgLoader, the class loader itself.
A single class loader object is created.
Both the JSP instances and classes are cached in hashtables.
The invalidation JSP is named jspUpdate.jsp.
To invalidate an JSP, it has simply to remove object and cache entries from
the caches.
ocgServlet
* Lazily creates the class loader.
* Retrieves the target JSP instance from the cache, if possible.
* Otherwise it uses the class loader to retrieve the target JSP class,
create a target JSP
instance and stores it in the cache.
* Forwards the request to the target JSP instance.
ocgLoader
* If the requested class has not the extension ocgServlet is configured to
process, it
behaves as a regular class loader and forwards the request to the parent
or system class
loader.
* Otherwise, it retrieves the class from the cache, if possible.
* Otherwise, it loads the class.
Do you thing it is a good solution?
I believe that solution is faster than standard WLS one, because it is a
very small piece of code but too because:
- my class loader is deterministic, if the file has the right extension I
don't call the classloader hierarchy first
- I don't try supporting jars. It has been one of the hardest design
decision. We definitely need a way to
update a specific page but at the same time someone post us NT could have
problems handling
3000 files in the same directory (it seems he was wrong).
- I don't try to check if a class has been updated. I have to ask for
refresh using a JSP now but it could be an EJB.
- I don't try to check if a source has been updated.
- As I know the number of JSP I can set pretty accurately the initial
capacity of the hashtables I use as caches. I
avoid rehash.

Use a profiler to find the bottlenecks in the system. You need to determine where the performance problems (if you even have any) are happening. We can't do that for you.

Large number of JSP performance [repost for grandemange]

Hi,
a colleague of me made tests with a large number of JSP and identified a
performance problem.
I believe I found a solution to his problem. I tested it with WLS 5.1 SP2
and SP3 and MS jview SDK 4.
The issue was related to the duration of the initial call of the nth JSP,
our situation as we are doing site hosting.
The solution is able to perform around 14 initial invocations/s no matter if
the invocation is the first one or the
3000th one and the throughput can go up to 108 JSPs/s when the JSP are
already loaded, the JSPs being the
snoopservlet example copied 3000 times.
The ratio have more interest than the values as the test machine (client and
WLS 5.1) was a 266Mhz laptop.
I repeat the post of Marc on 2/11/2000 as it is an old one:
Hi all,
I'm wondering if any of you has experienced performance issue whendeploying
a lot of JSPs.
I'm running Weblogic 4.51SP4 with performance pack on NT4 and Jdk1.2.2.
I deployed over 3000 JSPs (identical but with a distinct name) on myserver.
I took care to precompile them off-line.
To run my tests I used a servlet selecting randomly one of them and
redirecting the request.
getServletContext().getRequestDispatcher(randomUrl).forward(request,response);
The response-time slow-down dramaticaly as the number of distinct JSPs
invoked is growing.
(up to 100 times the initial response time).
I made some additional tests.
When you set the properties:
weblogic.httpd.servlet.reloadCheckSecs=-1
weblogic.httpd.initArgs.*.jsp=..., pageCheckSeconds=-1, ...
Then the response-time for a new JSP seems linked to a "capacity increase
process" and depends on the number of previously activated JSPs. If you
invoke a previously loaded page the server answers really fast with no
delay.
If you set previous properties to any other value (0 for example) the
response-time remains bad even when you invoke a previously loaded page.SOLUTION DESCRIPTION
Intent
The package described below is design to allow
* Fast invocation even with a large number of pages (which can be the case
with Web Hosting)
* Dynamic update of compiled JSP
Implementation
The current implementation has been tested with JDK 1.1 only and works with
MS SDK 4.0.
It has been tested with WLS 5.1 with service packs 2 and 3.
It should work with most application servers, as its requirements are
limited. It requires
a JSP to be able to invoke a class loader.
Principle
For a fast invocation, it does not support dynamic compilation as described
in the JSP
model.
There is no automatic recognition of modifications. Instead a JSP is made
available to
invalidate pages which must be updated.
We assume pages managed through this package to be declared in
weblogic.properties as
weblogic.httpd.register.*.ocg=ocgLoaderPkg.ocgServlet
This definition means that, when a servlet or JSP with a .ocg extension is
requested, it is
forwarded to the package.
It implies 2 things:
* Regular JSP handling and package based handling can coexist in the same
Application Server
instance.
* It is possible to extend the implementation to support many extensions
with as many
package instances.
The package (ocgLoaderPkg) contains 2 classes:
* ocgServlet, a servlet instantiating JSP objects using a class loader.
* ocgLoader, the class loader itself.
A single class loader object is created.
Both the JSP instances and classes are cached in hashtables.
The invalidation JSP is named jspUpdate.jsp.
To invalidate an JSP, it has simply to remove object and cache entries from
the caches.
ocgServlet
* Lazily creates the class loader.
* Retrieves the target JSP instance from the cache, if possible.
* Otherwise it uses the class loader to retrieve the target JSP class,
create a target JSP
instance and stores it in the cache.
* Forwards the request to the target JSP instance.
ocgLoader
* If the requested class has not the extension ocgServlet is configured to
process, it
behaves as a regular class loader and forwards the request to the parent
or system class
loader.
* Otherwise, it retrieves the class from the cache, if possible.
* Otherwise, it loads the class.
Do you thing it is a good solution?
I believe that solution is faster than standard WLS one, because it is a
very small piece of code but too because:
- my class loader is deterministic, if the file has the right extension I
don't call the classloader hierarchy first
- I don't try supporting jars. It has been one of the hardest design
decision. We definitely need a way to
update a specific page but at the same time someone post us NT could have
problems handling
3000 files in the same directory (it seems he was wrong).
- I don't try to check if a class has been updated. I have to ask for
refresh using a JSP now but it could be an EJB.
- I don't try to check if a source has been updated.
- As I know the number of JSP I can set pretty accurately the initial
capacity of the hashtables I use as caches. I
avoid rehash.
Cheers - Wei

I dont know the upper limit, but I think 80 is too much. I have never used more than 15-20. For Nav attributes, a seperate tables are created which causes the Performance issue as result in new join at query run time. Just ask your business guy, if these can be reduced.One way could be to model these attributes as seperate characteristics. It will certainly help.
Thanks...
Shambhu

L3 topology diagram with a large number of networks

Similar Messages

Maybe you are looking for