Sybase NNTP forums - End Of Life (EOL)

The NNTP forums from Sybase - forums.sybase.com - are now closed.

All new questions should be directed to the appropriate forum at the SAP Community Network (SCN).

Individual products have links to the respective forums on SCN, or you can go to SCN and search for your product in the search box (upper right corner) to find your specific developer center.

Numbers for large Afaria implementations

14 posts in General Last posting was on 2008-07-10 00:30:18.0Z
Sybase Newsgroups Posted on 2008-05-05 18:04:12.0Z
From: "Sybase Newsgroups" <tkoberle@sybase.com>
Newsgroups: ianywhere.public.afaria.general
Subject: Numbers for large Afaria implementations
Lines: 15
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.3138
X-RFC2646: Format=Flowed; Original
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <481f4c1c$1@forums-1-dub>
Date: 5 May 2008 11:04:12 -0700
X-Trace: forums-1-dub 1210010652 10.22.241.152 (5 May 2008 11:04:12 -0700)
X-Original-Trace: 5 May 2008 11:04:12 -0700, vip152.sybase.com
X-Authenticated-User: techsupp
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.afaria.general:8
Article PK: 6

I am working on a very large Afaria implementation in which there will be
~2000 simultaneous connections to the server farm at one time. Currently we
are in the process of specing out the server farm and I am looking for
information about large implementations regarding the number of simultaneous
connections, load balancing and hardware configuration for the server farm
nodes.

Are there any SPS people out there that can provide me some data on this ? I
am not looking for client names or anything like that, just technical info
to guide us in choosing the best configuration for our scenario.

Best Regards
Thomaz


Christopher Heeter Posted on 2008-05-06 17:22:01.0Z
Reply-To: "Christopher Heeter" <Chris.Heeter@Sybase.com>
From: "Christopher Heeter" <Chris.Heeter@Sybase.com>
Newsgroups: ianywhere.public.afaria.general
References: <481f4c1c$1@forums-1-dub>
Subject: Re: Numbers for large Afaria implementations
Lines: 39
Organization: Sybase iAnywher
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Response
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5512
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <482093b9$1@forums-1-dub>
Date: 6 May 2008 10:22:01 -0700
X-Trace: forums-1-dub 1210094521 10.22.241.152 (6 May 2008 10:22:01 -0700)
X-Original-Trace: 6 May 2008 10:22:01 -0700, vip152.sybase.com
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.afaria.general:9
Article PK: 28

Thomaz,

In order to give help you with the hardware requirements, I will need
more information. Can you please provide more specifics about connectivity
and user habits? Are the connections on a local network or via the
Internet? Is it required that all Clients connect about the same time
(simultaneous)? OR can the connections be scheduled to be spaced further
apart, or randomized? Will SSL be involved? Will hardware virtualization
be used?

We generally do not recommend one brand/type of load balancing over
another. There is documentation available on the Technical Support website
that explains the requirements for communication. The TCP/IP Ports required
are configurable. If a Reverse Proxy will be used, there is a document that
describes how to configure ISA or Apache to allow Afaria traffic to pass.



Chris

"Sybase Newsgroups" <tkoberle@sybase.com> wrote in message
news:481f4c1c$1@forums-1-dub...
>I am working on a very large Afaria implementation in which there will be
>~2000 simultaneous connections to the server farm at one time. Currently we
>are in the process of specing out the server farm and I am looking for
>information about large implementations regarding the number of
>simultaneous connections, load balancing and hardware configuration for the
>server farm nodes.
>
> Are there any SPS people out there that can provide me some data on this ?
> I am not looking for client names or anything like that, just technical
> info to guide us in choosing the best configuration for our scenario.
>
> Best Regards
> Thomaz
>


Sybase Newsgroups Posted on 2008-05-07 14:52:22.0Z
From: "Sybase Newsgroups" <tkoberle@sybase.com>
Newsgroups: ianywhere.public.afaria.general
References: <481f4c1c$1@forums-1-dub> <482093b9$1@forums-1-dub>
Subject: Re: Numbers for large Afaria implementations
Lines: 93
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.3138
X-RFC2646: Format=Flowed; Response
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4821c226@forums-1-dub>
Date: 7 May 2008 07:52:22 -0700
X-Trace: forums-1-dub 1210171942 10.22.241.152 (7 May 2008 07:52:22 -0700)
X-Original-Trace: 7 May 2008 07:52:22 -0700, vip152.sybase.com
X-Authenticated-User: techsupp
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.afaria.general:10
Article PK: 9

Hi Chris



Thanks for your reply. I will give you a summary of the project and the
application involved. Our client is the largest beverage distribution
company in Latin America with over 4000 mobile salespeople scattered around
the country. These people will be visiting point of sales locations with a
mobile device which they will use to input sales orders. At the beginning of
each workday they all will connect to the Afaria server through the internet
using a GPRS (or Edge) cell phone connection to download the days route
which is an XML file around 2-3 MB in size. They will all do this at about
the same time, we will advise the client to randomize the connection time a
bit but I am not sure whether they will accept this. This will be the time
when the heaviest load will be put on the Afaria farm.

During the course of the day all the salespeople will upload their sales
orders at pre-specified times. Again, they will all do it at the same time
but this time the amount of data will be considerably smaller (up to 100
files of arounf 3K).



Consideration about the environment:



- Afaria's SSL support will not be used since the carrier already provides
encryption.

- All the Afaria servers will be virtualized



The client has agreed on a Cisco ACE load balancer so this issue is behind
us. Our concern at the moment is to figure out whether the server farm we
have suggested is enough to get the job done. At the moment they are
building 4 Afaria servers, each equiped with 4 cores (dual duo-core
processors) and 4GB of memory. They are considering adding 2 more to account
for any servers that become unavailable.



Any input you (or anyone else) can provise will be greatly appreciated

Thomaz

"Christopher Heeter" <Chris.Heeter@Sybase.com> wrote in message
news:482093b9$1@forums-1-dub...
> Thomaz,
>
> In order to give help you with the hardware requirements, I will need
> more information. Can you please provide more specifics about connectivity
> and user habits? Are the connections on a local network or via the
> Internet? Is it required that all Clients connect about the same time
> (simultaneous)? OR can the connections be scheduled to be spaced further
> apart, or randomized? Will SSL be involved? Will hardware virtualization
> be used?
>
> We generally do not recommend one brand/type of load balancing over
> another. There is documentation available on the Technical Support
> website that explains the requirements for communication. The TCP/IP Ports
> required are configurable. If a Reverse Proxy will be used, there is a
> document that describes how to configure ISA or Apache to allow Afaria
> traffic to pass.
>
>
>
> Chris
>
>
> "Sybase Newsgroups" <tkoberle@sybase.com> wrote in message
> news:481f4c1c$1@forums-1-dub...
>>I am working on a very large Afaria implementation in which there will be
>>~2000 simultaneous connections to the server farm at one time. Currently
>>we are in the process of specing out the server farm and I am looking for
>>information about large implementations regarding the number of
>>simultaneous connections, load balancing and hardware configuration for
>>the server farm nodes.
>>
>> Are there any SPS people out there that can provide me some data on this
>> ? I am not looking for client names or anything like that, just technical
>> info to guide us in choosing the best configuration for our scenario.
>>
>> Best Regards
>> Thomaz
>>
>
>


Kalpesh Patel Posted on 2008-06-02 10:16:32.0Z
Sender: 5ffb.4843c6c7.1804289383@sybase.com
From: Kalpesh Patel
Newsgroups: ianywhere.public.afaria.general
Subject: Re: Numbers for large Afaria implementations
X-Mailer: WebNews to Mail Gateway v1.1t
Message-ID: <4843c880.6029.1681692777@sybase.com>
References: <4821c226@forums-1-dub>
NNTP-Posting-Host: 10.22.241.41
X-Original-NNTP-Posting-Host: 10.22.241.41
Date: 2 Jun 2008 03:16:32 -0700
X-Trace: forums-1-dub 1212401792 10.22.241.41 (2 Jun 2008 03:16:32 -0700)
X-Original-Trace: 2 Jun 2008 03:16:32 -0700, 10.22.241.41
Lines: 16
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.afaria.general:20
Article PK: 21

Hi there. As iAnywhere support will inform you, the
configuration of the NICs on each of the four (or more)
servers in your farm is key.

This is so that traffic can be routed to handle the maximum
number of concurrent client connections. Given that you will
be downloading a large amount of data, I would also
investigate timeout settings and MTU sizes over the mobile
network. I've had problems with Afaria and large data in the
past with GPRS.

As to your server specification, it would seem that the
server is more than ample to handle a simple file download.
Is there anything else that the Afaria server will be doing
to these clients? (E.G. Security or Inventory Manager
channels)


Christopher Heeter Posted on 2008-06-02 12:07:50.0Z
Reply-To: "Christopher Heeter" <Chris.Heeter@Sybase.com>
From: "Christopher Heeter" <Chris.Heeter@Sybase.com>
Newsgroups: ianywhere.public.afaria.general
References: <4821c226@forums-1-dub> <4843c880.6029.1681692777@sybase.com>
Subject: Re: Numbers for large Afaria implementations
Lines: 27
Organization: Sybase iAnywhere
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Original
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5512
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4843e296$1@forums-1-dub>
Date: 2 Jun 2008 05:07:50 -0700
X-Trace: forums-1-dub 1212408470 10.22.241.152 (2 Jun 2008 05:07:50 -0700)
X-Original-Trace: 2 Jun 2008 05:07:50 -0700, vip152.sybase.com
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.afaria.general:21
Article PK: 13

For further clarification, the issue with larger files does not lie with
Afaria, but has to do with the configuration of the network that the
Wireless Carrier is providing for GPRS. Afaria by default for faster
communications, GPRS is basically about the same or slower speeds as
traditional dial-up. Afaria can be tuned. Please refer to Knowledge Base
Articles #3632 for additional information.

<Kalpesh Patel> wrote in message news:4843c880.6029.1681692777@sybase.com...
> Hi there. As iAnywhere support will inform you, the
> configuration of the NICs on each of the four (or more)
> servers in your farm is key.
>
> This is so that traffic can be routed to handle the maximum
> number of concurrent client connections. Given that you will
> be downloading a large amount of data, I would also
> investigate timeout settings and MTU sizes over the mobile
> network. I've had problems with Afaria and large data in the
> past with GPRS.
>
> As to your server specification, it would seem that the
> server is more than ample to handle a simple file download.
> Is there anything else that the Afaria server will be doing
> to these clients? (E.G. Security or Inventory Manager
> channels)


Thomaz Koberle Posted on 2008-06-18 17:58:45.0Z
From: "Thomaz Koberle" <tkoberle@sybase.com>
Newsgroups: ianywhere.public.afaria.general
References: <4821c226@forums-1-dub> <4843c880.6029.1681692777@sybase.com>
Subject: Re: Numbers for large Afaria implementations
Lines: 40
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.3138
X-RFC2646: Format=Flowed; Original
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <48594cd5@forums-1-dub>
Date: 18 Jun 2008 10:58:45 -0700
X-Trace: forums-1-dub 1213811925 10.22.241.152 (18 Jun 2008 10:58:45 -0700)
X-Original-Trace: 18 Jun 2008 10:58:45 -0700, vip152.sybase.com
X-Authenticated-User: techsupp
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.afaria.general:34
Article PK: 144

Thanks for your reply.

Our client has done the server upgrades and are now runnig 3 quad-processor
servers with 4GB of memory each. We ran some stress tests and so far they
are holding well for up to 3000 concurrent connections doing small file
transfers (~500 Bytes).We still need to push those numbers up to 4500
connections to get an OK from management.

The problem now seems to have shifted to bandwidth limitations when doing
large file downloads. We ran the same tests for file of around 700K and
couldn't get past ~ 100 concurrent connections. More than that and we
started getting timeouts and lost links. I'mnot yet sure how we can help
from the Afaria standpoint since that seems to be more of a telecomm issue
( I have already consulted note 3632).

You mentioned that you've had timeout issues in the past. Do you have any
advice on what can be done to alleviate the ?

Best Regards
Thomaz

<Kalpesh Patel> wrote in message news:4843c880.6029.1681692777@sybase.com...
> Hi there. As iAnywhere support will inform you, the
> configuration of the NICs on each of the four (or more)
> servers in your farm is key.
>
> This is so that traffic can be routed to handle the maximum
> number of concurrent client connections. Given that you will
> be downloading a large amount of data, I would also
> investigate timeout settings and MTU sizes over the mobile
> network. I've had problems with Afaria and large data in the
> past with GPRS.
>
> As to your server specification, it would seem that the
> server is more than ample to handle a simple file download.
> Is there anything else that the Afaria server will be doing
> to these clients? (E.G. Security or Inventory Manager
> channels)


Christopher Heeter Posted on 2008-06-19 13:24:11.0Z
Reply-To: "Christopher Heeter" <SybaseiAnywhere@Live.com>
From: "Christopher Heeter" <SybaseiAnywhere@Live.com>
Newsgroups: ianywhere.public.afaria.general
References: <4821c226@forums-1-dub> <4843c880.6029.1681692777@sybase.com> <48594cd5@forums-1-dub>
Subject: Re: Numbers for large Afaria implementations
Lines: 62
Organization: Sybase iAnywhere
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Response
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5512
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <485a5dfb$1@forums-1-dub>
Date: 19 Jun 2008 06:24:11 -0700
X-Trace: forums-1-dub 1213881851 10.22.241.152 (19 Jun 2008 06:24:11 -0700)
X-Original-Trace: 19 Jun 2008 06:24:11 -0700, vip152.sybase.com
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.afaria.general:37
Article PK: 103

Thomaz,
If you are starting to see timeouts, that generally means that you
are either hitting the limits in Afaria and the OS AND/OR you have bandwidth
constraints on the network pipe coming in that Afaria cannot overcome. In
order to achieve the goal of 4500 concurrent connections, you will have to
add additional servers to the Afaria Farm AND ensure you have enough total
network bandwidth via the pipe that the traffic is coming in on to get to
the Afaria Farm. Have you tried setting up Bandwidth Throttling (BWT)? It
will reduce the overall amount of throughput per Session and allow for more
concurrent sessions, with the cost of longer session duration due to reduced
throughput...

Thanks,
Chris

"Thomaz Koberle" <tkoberle@sybase.com> wrote in message
news:48594cd5@forums-1-dub...
> Thanks for your reply.
>
> Our client has done the server upgrades and are now runnig 3
> quad-processor servers with 4GB of memory each. We ran some stress tests
> and so far they are holding well for up to 3000 concurrent connections
> doing small file transfers (~500 Bytes).We still need to push those
> numbers up to 4500 connections to get an OK from management.
>
> The problem now seems to have shifted to bandwidth limitations when doing
> large file downloads. We ran the same tests for file of around 700K and
> couldn't get past ~ 100 concurrent connections. More than that and we
> started getting timeouts and lost links. I'mnot yet sure how we can help
> from the Afaria standpoint since that seems to be more of a telecomm issue
> ( I have already consulted note 3632).
>
> You mentioned that you've had timeout issues in the past. Do you have any
> advice on what can be done to alleviate the ?
>
> Best Regards
> Thomaz
>
> <Kalpesh Patel> wrote in message
> news:4843c880.6029.1681692777@sybase.com...
>> Hi there. As iAnywhere support will inform you, the
>> configuration of the NICs on each of the four (or more)
>> servers in your farm is key.
>>
>> This is so that traffic can be routed to handle the maximum
>> number of concurrent client connections. Given that you will
>> be downloading a large amount of data, I would also
>> investigate timeout settings and MTU sizes over the mobile
>> network. I've had problems with Afaria and large data in the
>> past with GPRS.
>>
>> As to your server specification, it would seem that the
>> server is more than ample to handle a simple file download.
>> Is there anything else that the Afaria server will be doing
>> to these clients? (E.G. Security or Inventory Manager
>> channels)
>
>


Thomaz Koberle Posted on 2008-07-04 12:48:42.0Z
From: "Thomaz Koberle" <tkoberle@sybase.com>
Newsgroups: ianywhere.public.afaria.general
References: <4821c226@forums-1-dub> <4843c880.6029.1681692777@sybase.com> <48594cd5@forums-1-dub> <485a5dfb$1@forums-1-dub>
Subject: Re: Numbers for large Afaria implementations
Lines: 80
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.3138
X-RFC2646: Format=Flowed; Response
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <486e1c2a@forums-1-dub>
Date: 4 Jul 2008 05:48:42 -0700
X-Trace: forums-1-dub 1215175722 10.22.241.152 (4 Jul 2008 05:48:42 -0700)
X-Original-Trace: 4 Jul 2008 05:48:42 -0700, vip152.sybase.com
X-Authenticated-User: techsupp
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.afaria.general:72
Article PK: 181

Hi Chris

I have one last question regarding this issue. We have been experiencing
some timeouts/lost links probably due to constraints on the link (as you
mentioned). Theres one thing that's not yet clear to me; at which point does
the timeout gets triggered ? Does the Afaria server trigger the timeout and
marks the sessiosn as timed out or is it the client that drops the
connection ? In other words, at which point does the timeout gets triggered
? Server side or client side ?

Best Regards
Thomaz

"Christopher Heeter" <SybaseiAnywhere@Live.com> wrote in message
news:485a5dfb$1@forums-1-dub...
> Thomaz,
> If you are starting to see timeouts, that generally means that you
> are either hitting the limits in Afaria and the OS AND/OR you have
> bandwidth constraints on the network pipe coming in that Afaria cannot
> overcome. In order to achieve the goal of 4500 concurrent connections,
> you will have to add additional servers to the Afaria Farm AND ensure you
> have enough total network bandwidth via the pipe that the traffic is
> coming in on to get to the Afaria Farm. Have you tried setting up
> Bandwidth Throttling (BWT)? It will reduce the overall amount of
> throughput per Session and allow for more concurrent sessions, with the
> cost of longer session duration due to reduced throughput...
>
> Thanks,
> Chris
>
>
>
> "Thomaz Koberle" <tkoberle@sybase.com> wrote in message
> news:48594cd5@forums-1-dub...
>> Thanks for your reply.
>>
>> Our client has done the server upgrades and are now runnig 3
>> quad-processor servers with 4GB of memory each. We ran some stress tests
>> and so far they are holding well for up to 3000 concurrent connections
>> doing small file transfers (~500 Bytes).We still need to push those
>> numbers up to 4500 connections to get an OK from management.
>>
>> The problem now seems to have shifted to bandwidth limitations when doing
>> large file downloads. We ran the same tests for file of around 700K and
>> couldn't get past ~ 100 concurrent connections. More than that and we
>> started getting timeouts and lost links. I'mnot yet sure how we can help
>> from the Afaria standpoint since that seems to be more of a telecomm
>> issue ( I have already consulted note 3632).
>>
>> You mentioned that you've had timeout issues in the past. Do you have any
>> advice on what can be done to alleviate the ?
>>
>> Best Regards
>> Thomaz
>>
>> <Kalpesh Patel> wrote in message
>> news:4843c880.6029.1681692777@sybase.com...
>>> Hi there. As iAnywhere support will inform you, the
>>> configuration of the NICs on each of the four (or more)
>>> servers in your farm is key.
>>>
>>> This is so that traffic can be routed to handle the maximum
>>> number of concurrent client connections. Given that you will
>>> be downloading a large amount of data, I would also
>>> investigate timeout settings and MTU sizes over the mobile
>>> network. I've had problems with Afaria and large data in the
>>> past with GPRS.
>>>
>>> As to your server specification, it would seem that the
>>> server is more than ample to handle a simple file download.
>>> Is there anything else that the Afaria server will be doing
>>> to these clients? (E.G. Security or Inventory Manager
>>> channels)
>>
>>
>
>


Christopher Heeter Posted on 2008-07-07 14:24:41.0Z
Reply-To: "Christopher Heeter" <SybaseiAnywhere@Live.com>
From: "Christopher Heeter" <SybaseiAnywhere@Live.com>
Newsgroups: ianywhere.public.afaria.general
References: <4821c226@forums-1-dub> <4843c880.6029.1681692777@sybase.com> <48594cd5@forums-1-dub> <485a5dfb$1@forums-1-dub> <486e1c2a@forums-1-dub>
Subject: Re: Numbers for large Afaria implementations
Lines: 93
Organization: Sybase iAnywhere
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Response
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5512
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <48722729$1@forums-1-dub>
Date: 7 Jul 2008 07:24:41 -0700
X-Trace: forums-1-dub 1215440681 10.22.241.152 (7 Jul 2008 07:24:41 -0700)
X-Original-Trace: 7 Jul 2008 07:24:41 -0700, vip152.sybase.com
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.afaria.general:74
Article PK: 186

Thomaz,

KB Article #1813 does a pretty good job of explaining the differences
between the session failed statuses


Thanks,
Chris

"Thomaz Koberle" <tkoberle@sybase.com> wrote in message
news:486e1c2a@forums-1-dub...
> Hi Chris
>
> I have one last question regarding this issue. We have been experiencing
> some timeouts/lost links probably due to constraints on the link (as you
> mentioned). Theres one thing that's not yet clear to me; at which point
> does the timeout gets triggered ? Does the Afaria server trigger the
> timeout and marks the sessiosn as timed out or is it the client that drops
> the connection ? In other words, at which point does the timeout gets
> triggered ? Server side or client side ?
>
> Best Regards
> Thomaz
>
>
> "Christopher Heeter" <SybaseiAnywhere@Live.com> wrote in message
> news:485a5dfb$1@forums-1-dub...
>> Thomaz,
>> If you are starting to see timeouts, that generally means that you
>> are either hitting the limits in Afaria and the OS AND/OR you have
>> bandwidth constraints on the network pipe coming in that Afaria cannot
>> overcome. In order to achieve the goal of 4500 concurrent connections,
>> you will have to add additional servers to the Afaria Farm AND ensure you
>> have enough total network bandwidth via the pipe that the traffic is
>> coming in on to get to the Afaria Farm. Have you tried setting up
>> Bandwidth Throttling (BWT)? It will reduce the overall amount of
>> throughput per Session and allow for more concurrent sessions, with the
>> cost of longer session duration due to reduced throughput...
>>
>> Thanks,
>> Chris
>>
>>
>>
>> "Thomaz Koberle" <tkoberle@sybase.com> wrote in message
>> news:48594cd5@forums-1-dub...
>>> Thanks for your reply.
>>>
>>> Our client has done the server upgrades and are now runnig 3
>>> quad-processor servers with 4GB of memory each. We ran some stress tests
>>> and so far they are holding well for up to 3000 concurrent connections
>>> doing small file transfers (~500 Bytes).We still need to push those
>>> numbers up to 4500 connections to get an OK from management.
>>>
>>> The problem now seems to have shifted to bandwidth limitations when
>>> doing large file downloads. We ran the same tests for file of around
>>> 700K and couldn't get past ~ 100 concurrent connections. More than that
>>> and we started getting timeouts and lost links. I'mnot yet sure how we
>>> can help from the Afaria standpoint since that seems to be more of a
>>> telecomm issue ( I have already consulted note 3632).
>>>
>>> You mentioned that you've had timeout issues in the past. Do you have
>>> any advice on what can be done to alleviate the ?
>>>
>>> Best Regards
>>> Thomaz
>>>
>>> <Kalpesh Patel> wrote in message
>>> news:4843c880.6029.1681692777@sybase.com...
>>>> Hi there. As iAnywhere support will inform you, the
>>>> configuration of the NICs on each of the four (or more)
>>>> servers in your farm is key.
>>>>
>>>> This is so that traffic can be routed to handle the maximum
>>>> number of concurrent client connections. Given that you will
>>>> be downloading a large amount of data, I would also
>>>> investigate timeout settings and MTU sizes over the mobile
>>>> network. I've had problems with Afaria and large data in the
>>>> past with GPRS.
>>>>
>>>> As to your server specification, it would seem that the
>>>> server is more than ample to handle a simple file download.
>>>> Is there anything else that the Afaria server will be doing
>>>> to these clients? (E.G. Security or Inventory Manager
>>>> channels)
>>>
>>>
>>
>>
>
>


Thomaz Koberle Posted on 2008-07-09 19:00:01.0Z
From: "Thomaz Koberle" <tkoberle@sybase.com>
Newsgroups: ianywhere.public.afaria.general
References: <4821c226@forums-1-dub> <4843c880.6029.1681692777@sybase.com> <48594cd5@forums-1-dub> <485a5dfb$1@forums-1-dub> <486e1c2a@forums-1-dub> <48722729$1@forums-1-dub>
Subject: Re: Numbers for large Afaria implementations
Lines: 113
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.3138
X-RFC2646: Format=Flowed; Response
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <48750ab1$1@forums-1-dub>
Date: 9 Jul 2008 12:00:01 -0700
X-Trace: forums-1-dub 1215630001 10.22.241.152 (9 Jul 2008 12:00:01 -0700)
X-Original-Trace: 9 Jul 2008 12:00:01 -0700, vip152.sybase.com
X-Authenticated-User: techsupp
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.afaria.general:84
Article PK: 208

Thanks Chris. The article was helful

One of our clients is experiencing a rather high percentage of timeouts /
lost links which could be related to instabilities in their link. They are
getting close to 30% timeouts/lost links through GPRS connections. Since
they are using Afaria as a means to transmit data collected by a handheld
application, it has an impact on the application's usability since the user
has to restart the transmission manually.

From your experience is that a common occurrence? Do other clients
experience these timeouts ? Do you have a suggestion for how to minimize the
impact, for instance by implementing retries in the application ?

Regards
Thomaz

"Christopher Heeter" <SybaseiAnywhere@Live.com> wrote in message
news:48722729$1@forums-1-dub...
> Thomaz,
>
> KB Article #1813 does a pretty good job of explaining the differences
> between the session failed statuses
>
>
> Thanks,
> Chris
>
> "Thomaz Koberle" <tkoberle@sybase.com> wrote in message
> news:486e1c2a@forums-1-dub...
>> Hi Chris
>>
>> I have one last question regarding this issue. We have been experiencing
>> some timeouts/lost links probably due to constraints on the link (as you
>> mentioned). Theres one thing that's not yet clear to me; at which point
>> does the timeout gets triggered ? Does the Afaria server trigger the
>> timeout and marks the sessiosn as timed out or is it the client that
>> drops the connection ? In other words, at which point does the timeout
>> gets triggered ? Server side or client side ?
>>
>> Best Regards
>> Thomaz
>>
>>
>> "Christopher Heeter" <SybaseiAnywhere@Live.com> wrote in message
>> news:485a5dfb$1@forums-1-dub...
>>> Thomaz,
>>> If you are starting to see timeouts, that generally means that
>>> you are either hitting the limits in Afaria and the OS AND/OR you have
>>> bandwidth constraints on the network pipe coming in that Afaria cannot
>>> overcome. In order to achieve the goal of 4500 concurrent connections,
>>> you will have to add additional servers to the Afaria Farm AND ensure
>>> you have enough total network bandwidth via the pipe that the traffic is
>>> coming in on to get to the Afaria Farm. Have you tried setting up
>>> Bandwidth Throttling (BWT)? It will reduce the overall amount of
>>> throughput per Session and allow for more concurrent sessions, with the
>>> cost of longer session duration due to reduced throughput...
>>>
>>> Thanks,
>>> Chris
>>>
>>>
>>>
>>> "Thomaz Koberle" <tkoberle@sybase.com> wrote in message
>>> news:48594cd5@forums-1-dub...
>>>> Thanks for your reply.
>>>>
>>>> Our client has done the server upgrades and are now runnig 3
>>>> quad-processor servers with 4GB of memory each. We ran some stress
>>>> tests and so far they are holding well for up to 3000 concurrent
>>>> connections doing small file transfers (~500 Bytes).We still need to
>>>> push those numbers up to 4500 connections to get an OK from management.
>>>>
>>>> The problem now seems to have shifted to bandwidth limitations when
>>>> doing large file downloads. We ran the same tests for file of around
>>>> 700K and couldn't get past ~ 100 concurrent connections. More than that
>>>> and we started getting timeouts and lost links. I'mnot yet sure how we
>>>> can help from the Afaria standpoint since that seems to be more of a
>>>> telecomm issue ( I have already consulted note 3632).
>>>>
>>>> You mentioned that you've had timeout issues in the past. Do you have
>>>> any advice on what can be done to alleviate the ?
>>>>
>>>> Best Regards
>>>> Thomaz
>>>>
>>>> <Kalpesh Patel> wrote in message
>>>> news:4843c880.6029.1681692777@sybase.com...
>>>>> Hi there. As iAnywhere support will inform you, the
>>>>> configuration of the NICs on each of the four (or more)
>>>>> servers in your farm is key.
>>>>>
>>>>> This is so that traffic can be routed to handle the maximum
>>>>> number of concurrent client connections. Given that you will
>>>>> be downloading a large amount of data, I would also
>>>>> investigate timeout settings and MTU sizes over the mobile
>>>>> network. I've had problems with Afaria and large data in the
>>>>> past with GPRS.
>>>>>
>>>>> As to your server specification, it would seem that the
>>>>> server is more than ample to handle a simple file download.
>>>>> Is there anything else that the Afaria server will be doing
>>>>> to these clients? (E.G. Security or Inventory Manager
>>>>> channels)
>>>>
>>>>
>>>
>>>
>>
>>
>
>


forums.sybase.com Posted on 2008-07-10 00:30:18.0Z
From: "forums.sybase.com" <SybaseiAnywhere@Live.com>
Newsgroups: ianywhere.public.afaria.general
References: <4821c226@forums-1-dub> <4843c880.6029.1681692777@sybase.com> <48594cd5@forums-1-dub> <485a5dfb$1@forums-1-dub> <486e1c2a@forums-1-dub> <48722729$1@forums-1-dub> <48750ab1$1@forums-1-dub>
In-Reply-To: <48750ab1$1@forums-1-dub>
Subject: Re: Numbers for large Afaria implementations
Lines: 130
MIME-Version: 1.0
Content-Type: text/plain; format=flowed; charset="iso-8859-1"; reply-type=response
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Windows Mail 6.0.6001.18000
X-MimeOLE: Produced By Microsoft MimeOLE V6.0.6001.18000
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4875581a$1@forums-1-dub>
Date: 9 Jul 2008 17:30:18 -0700
X-Trace: forums-1-dub 1215649818 10.22.241.152 (9 Jul 2008 17:30:18 -0700)
X-Original-Trace: 9 Jul 2008 17:30:18 -0700, vip152.sybase.com
X-Authenticated-User: techsupp
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.afaria.general:85
Article PK: 202

Thomaz,
Did you attempt to determine the optimum packet size for the carrieris
MTU? (Referring to KB 3632 again)... Did you implement the changes described
in hte MS article?

Retries are a good idea provided that the customer's application will
wait to get the appropriate inidacation back that the transfer is
complete... otherwise they will need to build that logic into the
application collecting the data/invoking Afaria to transmit the data...

Thanks,
Chris

"Thomaz Koberle" <tkoberle@sybase.com> wrote in message
news:48750ab1$1@forums-1-dub...
> Thanks Chris. The article was helful
>
> One of our clients is experiencing a rather high percentage of timeouts /
> lost links which could be related to instabilities in their link. They are
> getting close to 30% timeouts/lost links through GPRS connections. Since
> they are using Afaria as a means to transmit data collected by a handheld
> application, it has an impact on the application's usability since the
> user has to restart the transmission manually.
>
> From your experience is that a common occurrence? Do other clients
> experience these timeouts ? Do you have a suggestion for how to minimize
> the impact, for instance by implementing retries in the application ?
>
> Regards
> Thomaz
>
> "Christopher Heeter" <SybaseiAnywhere@Live.com> wrote in message
> news:48722729$1@forums-1-dub...
>> Thomaz,
>>
>> KB Article #1813 does a pretty good job of explaining the differences
>> between the session failed statuses
>>
>>
>> Thanks,
>> Chris
>>
>> "Thomaz Koberle" <tkoberle@sybase.com> wrote in message
>> news:486e1c2a@forums-1-dub...
>>> Hi Chris
>>>
>>> I have one last question regarding this issue. We have been experiencing
>>> some timeouts/lost links probably due to constraints on the link (as you
>>> mentioned). Theres one thing that's not yet clear to me; at which point
>>> does the timeout gets triggered ? Does the Afaria server trigger the
>>> timeout and marks the sessiosn as timed out or is it the client that
>>> drops the connection ? In other words, at which point does the timeout
>>> gets triggered ? Server side or client side ?
>>>
>>> Best Regards
>>> Thomaz
>>>
>>>
>>> "Christopher Heeter" <SybaseiAnywhere@Live.com> wrote in message
>>> news:485a5dfb$1@forums-1-dub...
>>>> Thomaz,
>>>> If you are starting to see timeouts, that generally means that
>>>> you are either hitting the limits in Afaria and the OS AND/OR you have
>>>> bandwidth constraints on the network pipe coming in that Afaria cannot
>>>> overcome. In order to achieve the goal of 4500 concurrent connections,
>>>> you will have to add additional servers to the Afaria Farm AND ensure
>>>> you have enough total network bandwidth via the pipe that the traffic
>>>> is coming in on to get to the Afaria Farm. Have you tried setting up
>>>> Bandwidth Throttling (BWT)? It will reduce the overall amount of
>>>> throughput per Session and allow for more concurrent sessions, with the
>>>> cost of longer session duration due to reduced throughput...
>>>>
>>>> Thanks,
>>>> Chris
>>>>
>>>>
>>>>
>>>> "Thomaz Koberle" <tkoberle@sybase.com> wrote in message
>>>> news:48594cd5@forums-1-dub...
>>>>> Thanks for your reply.
>>>>>
>>>>> Our client has done the server upgrades and are now runnig 3
>>>>> quad-processor servers with 4GB of memory each. We ran some stress
>>>>> tests and so far they are holding well for up to 3000 concurrent
>>>>> connections doing small file transfers (~500 Bytes).We still need to
>>>>> push those numbers up to 4500 connections to get an OK from
>>>>> management.
>>>>>
>>>>> The problem now seems to have shifted to bandwidth limitations when
>>>>> doing large file downloads. We ran the same tests for file of around
>>>>> 700K and couldn't get past ~ 100 concurrent connections. More than
>>>>> that and we started getting timeouts and lost links. I'mnot yet sure
>>>>> how we can help from the Afaria standpoint since that seems to be more
>>>>> of a telecomm issue ( I have already consulted note 3632).
>>>>>
>>>>> You mentioned that you've had timeout issues in the past. Do you have
>>>>> any advice on what can be done to alleviate the ?
>>>>>
>>>>> Best Regards
>>>>> Thomaz
>>>>>
>>>>> <Kalpesh Patel> wrote in message
>>>>> news:4843c880.6029.1681692777@sybase.com...
>>>>>> Hi there. As iAnywhere support will inform you, the
>>>>>> configuration of the NICs on each of the four (or more)
>>>>>> servers in your farm is key.
>>>>>>
>>>>>> This is so that traffic can be routed to handle the maximum
>>>>>> number of concurrent client connections. Given that you will
>>>>>> be downloading a large amount of data, I would also
>>>>>> investigate timeout settings and MTU sizes over the mobile
>>>>>> network. I've had problems with Afaria and large data in the
>>>>>> past with GPRS.
>>>>>>
>>>>>> As to your server specification, it would seem that the
>>>>>> server is more than ample to handle a simple file download.
>>>>>> Is there anything else that the Afaria server will be doing
>>>>>> to these clients? (E.G. Security or Inventory Manager
>>>>>> channels)
>>>>>
>>>>>
>>>>
>>>>
>>>
>>>
>>
>>
>
>


Steve Moss Posted on 2008-06-16 16:51:55.0Z
Sender: 6b6c.48569881.1804289383@sybase.com
From: Steve Moss
Newsgroups: ianywhere.public.afaria.general
Subject: Re: Numbers for large Afaria implementations
X-Mailer: WebNews to Mail Gateway v1.1t
Message-ID: <48569a2b.6bb9.1681692777@sybase.com>
References: <481f4c1c$1@forums-1-dub>
NNTP-Posting-Host: 10.22.241.41
X-Original-NNTP-Posting-Host: 10.22.241.41
Date: 16 Jun 2008 09:51:55 -0700
X-Trace: forums-1-dub 1213635115 10.22.241.41 (16 Jun 2008 09:51:55 -0700)
X-Original-Trace: 16 Jun 2008 09:51:55 -0700, 10.22.241.41
Lines: 21
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.afaria.general:29
Article PK: 124


> I am working on a very large Afaria implementation in
> which there will be ~2000 simultaneous connections to the
> server farm at one time. Currently we are in the process
> of specing out the server farm and I am looking for
> information about large implementations regarding the
> number of simultaneous connections, load balancing and
> hardware configuration for the server farm nodes.
>
> Are there any SPS people out there that can provide me
> some data on this ? I am not looking for client names or
> anything like that, just technical info to guide us in
> choosing the best configuration for our scenario.
>
> Best Regards
> Thomaz
>
>

I would advise no more than 150 concurrent connections per
server farm member. Another note unless you want a rather
large and useless SQL database only log what you need and
only lg detail when you are looking at a problem.


Thomaz Koberle Posted on 2008-06-18 18:05:46.0Z
From: "Thomaz Koberle" <tkoberle@sybase.com>
Newsgroups: ianywhere.public.afaria.general
References: <481f4c1c$1@forums-1-dub> <48569a2b.6bb9.1681692777@sybase.com>
Subject: Re: Numbers for large Afaria implementations
Lines: 37
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.3138
X-RFC2646: Format=Flowed; Original
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <48594e7a$1@forums-1-dub>
Date: 18 Jun 2008 11:05:46 -0700
X-Trace: forums-1-dub 1213812346 10.22.241.152 (18 Jun 2008 11:05:46 -0700)
X-Original-Trace: 18 Jun 2008 11:05:46 -0700, vip152.sybase.com
X-Authenticated-User: techsupp
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.afaria.general:35
Article PK: 120

Hi Steve

We have done stress testing on a 3 server farm and found that it help up
well with up to 3000 concurrent connections. The number is actually somewhat
less than that since by the time the last connection comes in some of the
others have already finished. Each sonnection is only executing a single
session manager channel so no encryption overhead is involved.

But from my experience 150 concurrent connections sounds like a rather low
number. Any specific reason for that number ?

Thomaz

<Steve Moss> wrote in message news:48569a2b.6bb9.1681692777@sybase.com...
>> I am working on a very large Afaria implementation in
>> which there will be ~2000 simultaneous connections to the
>> server farm at one time. Currently we are in the process
>> of specing out the server farm and I am looking for
>> information about large implementations regarding the
>> number of simultaneous connections, load balancing and
>> hardware configuration for the server farm nodes.
>>
>> Are there any SPS people out there that can provide me
>> some data on this ? I am not looking for client names or
>> anything like that, just technical info to guide us in
>> choosing the best configuration for our scenario.
>>
>> Best Regards
>> Thomaz
>>
>>
> I would advise no more than 150 concurrent connections per
> server farm member. Another note unless you want a rather
> large and useless SQL database only log what you need and
> only lg detail when you are looking at a problem.


Christopher Heeter Posted on 2008-06-17 11:56:13.0Z
Reply-To: "Christopher Heeter" <Chris.Heeter@Sybase.com>
From: "Christopher Heeter" <Chris.Heeter@Sybase.com>
Newsgroups: ianywhere.public.afaria.general
References: <481f4c1c$1@forums-1-dub> <48569a2b.6bb9.1681692777@sybase.com>
Subject: Re: Numbers for large Afaria implementations
Lines: 33
Organization: Sybase iAnywhere
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Original
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5512
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4857a65d$1@forums-1-dub>
Date: 17 Jun 2008 04:56:13 -0700
X-Trace: forums-1-dub 1213703773 10.22.241.152 (17 Jun 2008 04:56:13 -0700)
X-Original-Trace: 17 Jun 2008 04:56:13 -0700, vip152.sybase.com
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.afaria.general:32
Article PK: 156

An additional factor to consider is "how much" history do you want to keep,
not just what to log. If 2-3 days is suffice for troubleshooting sessions,
then the size of the database can be reigned in somewhat. Bottom line: I
would recommend that you only log what you need AND for how long you need
it.

Thanks,
Chris

<Steve Moss> wrote in message news:48569a2b.6bb9.1681692777@sybase.com...
>> I am working on a very large Afaria implementation in
>> which there will be ~2000 simultaneous connections to the
>> server farm at one time. Currently we are in the process
>> of specing out the server farm and I am looking for
>> information about large implementations regarding the
>> number of simultaneous connections, load balancing and
>> hardware configuration for the server farm nodes.
>>
>> Are there any SPS people out there that can provide me
>> some data on this ? I am not looking for client names or
>> anything like that, just technical info to guide us in
>> choosing the best configuration for our scenario.
>>
>> Best Regards
>> Thomaz
>>
>>
> I would advise no more than 150 concurrent connections per
> server farm member. Another note unless you want a rather
> large and useless SQL database only log what you need and
> only lg detail when you are looking at a problem.