This document describes how to configure a Natural Development Server for z/OS (Batch).

The following topics are covered:

• Configuration Requirements

• NDV Configuration File

• NDV Configuration Parameters

• NDV Configuration File Example

• NDV Server Datasets

• NDV User Exits (Coded in Natural)

• Enabling NDV User Exits

• Other NDV User Exits

• Encrypted Communication

• Configuring Port Sharing

Configuration Requirements

A Natural Development Server for z/OS (Batch) requires the following z/OS language environment parameter configuration:

ENVAR(TZ=CET-1DST) CET

You can set the z/OS language environment parameters:

• With the PARM parameter specified in the EXEC card of the Natural Development Server startup job. The length of the options is limited by the maximum length of the PARM parameter.

• Assemble an LE/370 runtime option module CEEUOPT and link it to the Natural Development Server load module.

External Security Configuration

If you configure the NDV server to impersonate the NDV clients in the NDV server (NDV configuration parameter SECURITY_MODE set to IMPERSONATE or IMPERSONATE_LOCAL), the NDV server must run "program-controlled". Under RACF, the following definitions are required for the NDV server:

• The resource BPX.SERVER must be defined and the NDV server account must have READ access to this resource.

• The LOAD datasets defined in the NDV startup job definition must be defined to the program class "**".

  ralt program ** addmem('natural load library') uacc(read)
  ralt program ** addmem('NDV load library'//NOPADCHK) uacc(read)
  ralt program ** addmem('user load library'//NOPADCHK) uacc(read)

• SETR WHEN(PROGRAM) REFRESH

Additionally, each client connecting to the server must be defined in RACF and must be granted to use the z/OS Linux System Services.

NDV Configuration File

A configuration file is allocated to the name <serverid>C (for example, NDVS1C) or STGCONFG alternatively.

The configuration file is a text file located on a dataset or on an HFS file under z/OS.

The configuration file contains the server configuration parameters in the form of a keyword=value syntax. In addition, it may contain comments whose beginning is marked with a hash symbol (#).

See also the NDV Configuration File Example shown below.

NDV Configuration Parameters

The following NDV configuration parameters are available:

• CPU_TIME_TRACE

• CPU_TIME_HISTORY

• DBG_CODEPAGE

• DEFAULT_PROFILE

• FORCE_IPV4

• FRONTEND_NAME

• FRONTEND_OPTIONS

• FRONTEND_PARAMETER

• HANDLE_ABEND

• HOST_NAME

• HTPMON_ADMIN_PSW

• HTPMON_PORT

• IGNORE_PRESENT_SERVER

• INITIAL_USERID

• KEEP_TCB

• MINIMUM_STUDIO_VERSION

• PARMCHECK

• PASSWORD_MIXEDCASE

• PORT_NUMBER

• SECURITY_MODE

• SECURITY_CACHING

• SECURITY_SYSLOG

• SECURITY_TIMEOUT

• SESSION_PARAMETER

• SESSION_PARAMETER_MIXED_CASE

• SESSION_TIMEOUT

• TERMINAL_EMULATION

• THREAD_NUMBER

• THREAD_SIZE

• TRACE_FILTER

• TRACE_LEVEL

• UNICODE_SOURCE

• UPPERCASE_SYSTEMMESSAGES

CPU_TIME_TRACE

This optional configuration parameter specifies a time limit of central processing unit (CPU) time of a transaction. Transactions with a higher CPU time usage than the specified time limit are traced in the trace file. Trace level 30 is required to enable the trace. There are two types of traces: one that shows the CPU time of a single transaction and one that shows the CPU time of an entire session.

Example of traces of session DEBE5E3816036616:

04 14:21:34 00000052 Execute cpu time of session DEBE5E3816036616: 0.001295 s
04 14:25:51 00000052 Total cpu time of session DEBE5E3816036616: 0.063536 s

The value of CPU_TIME_TRACE applies also to the CPU_TIME_HISTORY parameter if specified.

The parameter can be changed dynamically in the configuration menu of the HTML Monitor.

Example:

CPU_TIME_TRACE=1,0

The setting in the example traces transactions and sessions with higher CPU time usage than 1,0 seconds.

CPU_TIME_HISTORY

This optional configuration parameter specifies a buffer that stores the specified number of transactions with the highest central processing unit (CPU) time values.

The entries are sorted by CPU time in descending order. Each stored entry contains the CPU time, the session ID, the day and time, the user ID and type of transaction executed.

The list can be displayed with the HTML Monitor Client in a browser window or with Monitor Client NATMOPI. The number of stored entries can be filtered by setting a CPU time limit with CPU_TIME_TRACE. The parameter can be changed dynamically in the configuration menu of the HTML Monitor.

Example:

CPU_TIME_HISTORY=32

The setting in the example stores the 32 highest CPU time values sorted by cpu time.

DBG_CODEPAGE

This optional configuration parameter specifies the translation table to be used by the remote debugger. By default, the remote debugger uses the code page IBM-1047, whereas the Natural Development Server uses TABA1/2.

No default value is provided.

Example:

DBG_CODEPAGE=USER

DEFAULT_PROFILE

This optional configuration parameter defines a default profile.

profile-name,dbid,fnr,password,cipher-code

No default value is provided.

Example:

DEFAULT_PROFILE=RDEVS,10,930

The setting in the example defines that, if no parameters are defined in the Map Environment dialog box of Natural Studio, the session is started with the Natural profile parameter PROFILE=(RDEVS,10,930).

Related parameter: SESSION_PARAMETER.

FORCE_IPV4

This configuration parameter applies to z/OS.

This parameter allows you to enforce the use of communication method IPV4.

FRONTEND_NAME

This configuration parameter specifies the name of the Natural front-end to be used to start a Natural session. The front-end resides on a PDS member.

No default value is provided.

Example:

FRONTEND_NAME=NATvrsSV

- where vrs stands for the version, release, system maintenance number.

FRONTEND_OPTIONS

The values of this configuration parameter may be used to specify additional options for the Natural front-end.

You may combine the above options as desired in that you add their values and set the result as shown in the example below.

Example:

FRONTEND_OPTIONS=07

The setting in this example enables the Options 01, 02 and 04.

FRONTEND_PARAMETER

This optional configuration parameter contains additional Natural front-end parameters as specified in the Startup Parameter Area.

FRONTEND_PARAMETER = 'MSGCLASSX       '

No default value is provided.

For further information, refer to the section Natural in Batch Mode in the Natural Operations documentation for z/OS.

Example:

FRONTEND_PARAMETER='MSGCLASSX       '

The setting in this example specifies that the default output class for CMPRINT is X.

HANDLE_ABEND

If an abend occurs in the server processing outside the Natural processing the abend is not trapped by the Natural abend handling. For this reason the NDV server has its own abend recovery.

It is recommended that you leave this parameter on its default value in order to limit the impact of an abend to a single user. If you set the value of this parameter to NO, any abend in the server processing terminates the complete server processing. That is, it affects all users running on that server.

Example:

HANDLE_ABEND=NO

HOST_NAME

This optional configuration parameter is necessary only if the server host supports multiple TCP/IP stacks.

No default value is provided.

Example:

HOST_NAME=node1

or

HOST_NAME=157.189.160.55

HTPMON_ADMIN_PSW

This configuration parameter defines the password required for some monitor activities (for example, Terminate Server) performed by the HTML Monitor Client.

No default value is provided.

Example:

HTPMON_ADMIN_PSW=GHAU129B

HTPMON_PORT

An NDV server can be configured to host an HTTP monitor task which serves the HTML Monitor Client running in a web browser. It is not required to run this monitor task on each server. A single task allows you to monitor all servers running at one node.

This configuration parameter defines the TCP/IP port number under which the server monitor task can be connected from a web browser.

No default value is provided.

Example:

HTPMON_PORT=3141

IGNORE_PRESENT_SERVER

An NDV server allocates a so-called "server environment" which contains the server dependent common resources. This environment is unique for each server and relates to the NDV server name.

If an NDV server with NDV CICS Adapter ends abnormally, it might leave a stuck NDV server environment within the CICS region. This causes that a restart of the server fails with error message NAT9913.

If you start an NDV server with IGNORE_PRESENT_SERVER=YES, it might damage an already running server which is using the same server name and the same CICS region.

Example:

IGNORE_PRESENT_SERVER=YES

INITIAL_USERID

At server initialization, the Natural Development Server creates a temporary Natural session to obtain the properties of the installed Natural environment.

This configuration parameter specifies the user ID to be used for this Natural session.

Example:

INITIAL_USERID=NDVINITU

See also NDV Clients must be defined to Natural Security (in the Natural Development Server Installation documentation).

KEEP_TCB

By default, the remote Natural session of a mapped environment terminates its TCB whenever you switch the focus within Natural Studio to a different mapped environment. If you toggle the focus back, the remote session is dispatched using a different TCB.

The maximum number of active TCBs is equal to the number of connected clients.

The configuration parameter KEEP_TCB specifies whether the remote Natural session should use the same TCB during its entire lifetime. This is required if you use Adabas and the Adabas parameter ADANAME is set to ADAUSER or if you want to access DB2. It could also be required if you access 3GL programs which need to be executed under the same TCB for successive calls.

Example:

KEEP_TCB=YES

MINIMUM_STUDIO_VERSION

This parameter defines a minimum version of Natural Studio which is required to operate with the NDV server. This parameter assists in performing a preliminary validation if all clients use a minimum Natural Studio version. This can be useful to smoothly upgrade to a NDV version that does not support clients whose version is below the minimum Natural Studio version.

Example:

MINIMUM_STUDIO_VERSION=62100

PARMCHECK

This parameter allows you to perform a parameter check of Natural session parameters after the server is up. The check is done by mapping to Natural. If a problem occurs, a message is written to the NDV server trace, and the server is terminated immediately. In such cases, session parameters (SESSION_PARAMETER) can be corrected. To avoid termination of the server in case of invalid Natural parameters, the parameter check can be switched off.

This parameter does not apply with SECURITY_MODE=IMPERSONATE, IMPERSONATE_LOCAL, or IMPERSONATE_REMOTE.

Example:

PARMCHECK=NO

PASSWORD_MIXEDCASE

This parameter allows you to define whether passwords specified in the Map Environment dialog are translated into upper case or not.

This parameter does only apply with SECURITY_MODE=IMPERSONATE, IMPERSONATE_LOCAL or IMPERSONATE_REMOTE.

Example:

PASSWORD_MIXEDCASE=YES

PORT_NUMBER

This configuration parameter defines the TCP/IP port number under which the server can be connected.

No default value is provided.

Example:

PORT_NUMBER=3140

SECURITY_MODE

The Natural Development Server offers a security concept that also covers the operating system resources. The client credentials are validated at the operating-system-depending security system and the client request is executed under the client's account data.

Using the SECURITY_MODE parameter, you can specify at which rank (Batch or CICS) you want to impersonate the activities of an NDV client.

Note concerning Natural for DB2 (Batch Server only): In order to be able to run the Natural Development Server with impersonation enabled, you must have linked the DB2 interface module DSNRLI (instead of DSNALI) to the Natural nucleus.

SECURITY_MODE requires Natural Security (NSC).

No default value is provided.

Example:

SECURITY_MODE=IMPERSONATE

SECURITY_CACHING

SECURITY_CACHING applies only to a Natural Development Server for z/OS (Batch), and pertains to SECURITY_MODE being set to IMPERSONATE_LOCAL.

SECURITY_CACHING is intended for server environments where password changes occur frequently. Commonly, each client action in NaturalONE starts a new server session and the password is checked each time. At worst, this requires the client to enter a new password for each client action.

In order to allow for the client to experience a smooth workflow, information about the client is stored in a client list. Based on the stored information, access is therefore possible without having to repeat entering new passwords. The duration of validity of a client entry can be set using the parameter SECURITY_TIMEOUT.

The client list stores the user Id and Host (IP Address), the time of login and the time of last access, together with a hash value of the password. The hash value of the password is generated by the sha1 algorithm. During the first login, the user is authorized and an ACEE is generated, which is also saved in the client list. For each subsequent accesses of the same user, only the hash value of the saved password is compared to the hash value of the sent password. If both passwords match, the session is started with the saved ACEE. If the hash values of the passwords are not the same, the user is prompted to enter a password.

To create an ACEE, SECURITY_CACHING requires the Natural Authorized Services Manager for the execution of authorized services of the z/OS Security Server RACF.

Example:

SECURITY_CACHING=YES

SECURITY_SYSLOG

Write a console message in case of security problems with external security. One line is written to the console, for example 'Aug-23 16:22:46.77 S0305217 BPXM023I (INITUSER) SERVERID - Invalid userid specified.' More information about the reason can be found in the appropriate trace or error log of the server.

The parameter applies to z/OS Batch servers only.

Example:

SECURITY_SYSLOG=YES

SECURITY_TIMEOUT

SECURITY_TIMEOUT only applies, if SECURITY_CACHING is set.

Cancels active client when the SECURITY_TIMEOUT parameter is met. Checks for clients active longer than n minutes once a day at HH:MM (24 hours) or every n minutes.

The server will not start if an invalid SECURITY_TIMEOUT parameter is given.

Examples:

SECURITY_TIMEOUT=19:30,480

Every day at 19:30 cancel clients logged on for more than 480 minutes.

SECURITY_TIMEOUT=19:30,0

Every day at 19:30 cancel all clients.

SECURITY_TIMEOUT=360,480

Every 360 minutes cancel clients logged on for more than 480 minutes.

SESSION_PARAMETER

This optional configuration parameter defines session parameters that precede the parameter string either specified in the Map Environment dialog of Natural Studio or defined by default by the configuration parameter DEFAULT_PROFILE.

No default value is provided.

Example 1:

SESSION_PARAMETER='NUCNAME=NATNUCvr' +
  'PROFILE=(NDVPARM,18006,48),ADAMODE=0,' +
  'BPI=(TYPE=NAT,SIZE=6044),BPI=(TYPE=EDIT,SIZE=2048)', + 
  'BPI=(TYPE=SORT,SIZE=1024)'

- where vr stands for the version and release number.

Example 2:

SESSION_PARAMETER=FNAT=(10,930)

The setting in the second example defines that every session on this Natural Development Server is started with the session parameter FNAT=(10,930) appended to the user-specified parameters or the definitions in the configuration parameter DEFAULT_PROFILE.

SESSION_PARAMETER_MIXED_CASE

This optional configuration parameter can be used to allow session parameters and URL specifications in mixed case.

SESSION_TIMEOUT

Cancel inactive sessions when the SESSION_TIMEOUT parameter is met. Check for sessions inactive longer then n minutes once a day at HH:MM (24 hours) or every n minutes.

The server will not start if an invalid SESSION_TIMEOUT parameter is given.

Examples:

SESSION_TIMEOUT=19:30,480

Every day at 19:30 cancel sessions more than 480 minutes inactive.

SESSION_TIMEOUT=360,480

Every 360 minutes cancel sessions more than 480 minutes inactive.

TERMINAL_EMULATION

This configuration parameter defines the terminal emulation to be used for processing the Natural I/O. This definition applies to all clients using that server.

Example:

TERMINAL_EMULATION=WEBIO

THREAD_NUMBER

This configuration parameter specifies the number of physical storage threads to be allocated by the Natural front-end, that is, the number of sessions that can be executed in parallel.

Example:

THREAD_NUMBER=5

THREAD_SIZE

This configuration parameter specifies the size of each physical storage thread which contains the Natural session data at execution time.

Example:

THREAD_SIZE=800

TRACE_FILTER

This optional configuration parameter enables you to restrict the trace by a logical filter in order to reduce the volume of the server trace output, for example:

TRACE_FILTER="Client=(KSP P*)"

Each request of the user ID KSP and each request of the user IDs starting with a P are traced.

See Trace Filter in the section Operating the Natural Development Server.

TRACE_LEVEL

Example:

TRACE_LEVEL=0x00000011

or alternatively

TRACE_LEVEL=31+27

The setting in the example switches on Bits 31 and 27.

UNICODE_SOURCE

This configuration parameter is used to define whether the NDV server accepts source files in Unicode or not.

Sources transmitted in Unicode are not converted using the Natural ASCII-to-EBCDIC translation tables TABA1/TABA2. All characters in the source file are supported without maintaining the Natural translation tables.

A transmission in Unicode, however, increases the CPU consumption of the server significantly.

Example:

UNICODE_SOURCE=YES

UPPERCASE_SYSTEMMESSAGES

This configuration parameter is used to enable or disable the translation of all NDV error messages and trace outputs to uppercase. This feature is for customers who are using character sets with no lowercase characters defined.

NDV Configuration File Example

 # This is a comment
SESSION_PARAMETER=profile=(stgqa,10,930) fuser=(10,32)
DEFAULT_PROFILE=DEFPROF
THREAD_NUMBER=2
THREAD_SIZE=700
FRONTEND_NAME=NATOS31L       # and another comment
PORT_NUMBER=4711

NDV Server Datasets

The Natural Development Server requires the following datasets:

Alternatively, you can qualify each dataset name by the server ID.

NDV User Exits (Coded in Natural)

Natural Single Point of Development provides the following user exits.

• NDV-UX01

• NDV-UX02

• NDV-UX03

NDV-UX01

The NDV-UX01 user exit is invoked before a Natural source object or a DDM is edited. It can be used to reject editing certain sources.

The name of the source code of this exit in the SYSLIB Natural system library is NDV-SX01.

To use this user exit, you must first enable it. For instructions, see Enabling NDV User Exits.

You can specify the following parameters for NDV-UX01:

NDV-UX02

The NDV-UX02 user exit is invoked before a Natural object or DDM, or when a user error message is deleted, copied, or moved (including the context menu functions Cut, Copy, and Paste). NDV-UX02 enables the rejection of further processing of this object, similarly to the MAINEX01 user exit of SYSMAIN in Natural for z/OS.

The name of the source code of this exit in the SYSLIB Natural system library is NDV-SX02.

To use this user exit, you must first enable it. For instructions, see Enabling NDV User Exits.

You can specify the following parameters for NDV-UX02:

NDV-UX03

The NDV-UX03 user exit provides flags for special settings within the Natural Development Server. It specifies whether resources are displayed and in which format data areas are written on the server.

The resources are not displayed by default for performance reasons, neither in the NaturalONE server view nor in the Natural Studio. If resources should be listed, set the DISPLAY-RESOURCES parameter to Y.

For higher performance, consider implementing the hyperdescriptor H1 as described in NaturalONE > NaturalONE in a Nutshell > Performance Aspects > Improve Mainframe Access.

For historical reasons, Natural supports three data area formats. Originally, the old internal format was used. Later, the new internal format was introduced. With Natural 9.1.2 (all platforms), the DEFINE DATA format was introduced. When saving a data area with NDV on the server, the DEFINE DATA format for NaturalONE or the new internal format for Natural Studio is used by default. When saving a data area directly on the server, the new internal format is used. You can override the default values for NaturalONE and Natural Studio with the NDV-UX03 user exit.

The name of the source code of this exit in the SYSLIB Natural system library is NDV-SX03.

To use this user exit, you must first enable it. For instructions, see Enabling NDV User Exits.

You can specify the following parameters for NDV-UX03:

Enabling NDV User Exits

The NDV user exits are delivered in the SYSLIB Natural system library.

The sources of these user exit routines are named NDV-SXnn, where nn denotes the number of the user exit routine.

To make an NDV user exit routine available:

1. Copy the source code of the user exit NDV-SXnn from the SYSLIB Natural system library into a user library and save it under the name NDV-UXnn.

   Renaming the user exit guarantees that the object is not affected if the user exit source is overwritten by an installation update.

2. Adapt the values of the parameters as required and stow the user exit.

3. Copy the cataloged object of NDV-UXnn to the SYSLIB Natural system library.

Other NDV User Exits

Apart from the NDV user exits that are coded in Natural, the following user exit exists.

User Exit NSECUX01

This user exit is applicable only when the parameter SECURITY_MODE is set to IMPERSONATE_LOCAL or IMPERSONATE.

This user exit allows you to adapt the user ID used for the RACF login. It is useful if the RACF user IDs and the user IDs used in Natural differ according to a standardized rule. For example, each RACF user ID is the corresponding Natural user ID preceded by two dollar signs ($$).

If the exit (the loadmodule NSECUX01) is found in the NDV load library concatenation, it is called using standard linkage conventions (direct branch using BASR instructions) before the user is validated against RACF.

You can specify the following parameters for this exit:

The exit is called using standard linkage conventions.

Sample user exit implemented in C:

#include <string.h>
#include <stdio.h>
#  pragma linkage (NSECUX01, FETCHABLE)
void NSECUX01(char sUid[64])
{
char sUidTemp[64];
  printf("Uex got usid:%s\n", sUid);
  strcpy(sUidTemp, sUid);
  sprintf(sUid, "$$%s", sUidTemp);
  printf("Uex ret usid:%s\n", sUid);
  return;
}

The exit above extends each user ID by two preceding dollar signs ($$) when it is used for RACF login.

Encrypted Communication

Communication between NaturalONE and NDV servers can be encrypted by using an SSL or (preferably) a TLS protocol.

On the server side, encryption is enabled by using the AT-TLS component of the z/OS Communications Server. AT-TLS allows you to encrypt and secure the TCP/IP communication between client and server without any changes on the server side. If you want to use encrypted communication mode, you must define a set of rules concerning the NDV server job for the policy agent (PAGENT) configuration file of AT-TLS. Encrypted communication requires a certificate associated with encryption on the server side. This encryption certificate must either be included in a RACF keyring or stored in a key database located on the z/OS USS POSIX file system.

Use the ZOSMF utility for maintaining the PAGENT.

For detailed information on AT-TLS, refer to the z/OS Communications Server documentation from IBM.

SSL/TLS Support

• SSL/TLS over AT-TLS
• Maintenance of Certificates under z/OS
• Using RACF Key Rings
• Using Key Databases
• How to configure TCP/IP for AT-TLS?
• How to Verify AT-TLS Configuration?
• Frequently Asked Questions
• Generation of a Natural development Server Certificate

SSL/TLS over AT-TLS

SSL/TLS support for the Natural development server is based on the z/OS Communication Server component AT-TLS (Application Transparent-Transport Layer Security).

AT-TLS provides SSL/TLS encryption as a configurable service for sockets applications. It is realized as an additional layer on top of the TCP/IP protocol stack, which exploits the SSL/TLS functionality in nearly or even fully transparent mode to sockets applications. AT-TLS offers three modes of operation. See z/OS Communications Server, IP Programmer’s Guide and Reference. Chapter: Application Transparent Transport Layer Security (AT-TLS).

These modes are:

• Basic

  The sockets application runs without modification in transparent mode, unaware of performing encrypted communication via AT-TLS. Thus legacy applications can run in secured mode without source code modification.

• Aware

  The application is aware of running in secured mode and is able to query TLS status information.

• Controlling

  The sockets application is aware of AT-TLS and controls the use of AT-TLS encryption services itself. This means, the application is able to switch between secured and non secured communication.

Natural development server uses the Basic mode for its SSL/TLS implementation. That is, a server configured as SSL/TLS server rejects requests from non-secured clients.

Maintenance of Certificates under z/OS

Certificates, which are to be used with AT-TLS, can be maintained in two ways under z/OS. They are stored either in RACF key rings or in key databases, which are located in the z/OS Linux file system. Which of these proceedings actually applies is defined in the AT-TLS Policy Agent Configuration file for the z/OS TCP/IP stack, which is used by the Natural HTTPS client.

IBM delivers a set of commonly used CA root certificates with each z/OS system delivery. If key rings are going to be used to hold server certificates, those root certificates must be manually imported into the key rings by the system administrator. If IBM delivers newer replacements for expired root certificates, all affected key rings have to be updated accordingly.

Unlike key rings, key databases contain the current set of root certificates automatically after they have been newly created. However, the need for maintaining always the latest set of root certificates applies to the key database alternative as well.

Using RACF Key Rings

In RACF, digital certificates are stored in so-called key rings. The RACF command RACDCERT is used to create and maintain key rings and certificates, which are contained in those key rings.

See z/OS Security Server RACF Security Administrator’s Guide, Chapter: RACF and digital certificates - RACF and key rings, and z/OS Security Server RACF, Command Language Reference, Chapter: RACF command syntax - RACDCERT.

Using Key Databases

Alternatively to RACF, certificates can be kept in key databases, which reside in the z/OS Linux services file system. For the creation and maintenance of key databases, the GSKKYMAN utility has to be used.

See z/OS Cryptographic Services PKI Services Guide and Reference, Chapter: Appendix B. Using a gskkyman key database for your certificate store.

How to configure TCP/IP for AT-TLS?

Proceed as follows:

1. In the TCP/IP configuration file, set the option TTLS in the TCPCONFIG statement.

2. Configure and start the AT-TLS Policy Agent. This agent is called by TCP/IP on each new TCP connection to check if the connection is SSL/TLS.

3. Create the Policy Agent file containing the AT-TLS rules. The Policy Agent file contains the rules to stipulate which connection is SSL/TLS.

See also z/OS Communications Server: IP Configuration Guide, Chapter: Application Transparent Transport Layer Security data protection.

The Sample Policy Agent file defines the server with the job name starting with NDVDEV and listening at port 4843 to use SSL/TLS.

The sample expects the certificate database on the HFS file /u/admin/CERT.kdb.

TTLSRule                          ConnRule01~1
{
  LocalAddrSetRef                 addr1
  RemoteAddrSetRef                addr1
  LocalPortRangeRef               portR1
  RemotePortRangeRef              portR2
  Jobname                         NDVDEV*
  Direction                       Inbound
  Priority                        255
  TTLSGroupActionRef              gAct1~NDV_Server
  TTLSEnvironmentActionRef        eAct1~NDV_Server
  TTLSConnectionActionRef         cAct1~NDV_Server
} 
TTLSGroupAction                   gAct1~NDV_Server
{ 
  TTLSEnabled                     On
} 
TTLSEnvironmentAction             eAct1~NDV_Server
{ 
  HandshakeRole                   Server
  EnvironmentUserInstance         0
  TTLSKeyringParmsRef             keyR1
} 
TTLSConnectionAction              cAct1~NDV_Server
{ 
  HandshakeRole                   Server
  TTLSCipherParmsRef              cipher1~AT-TLS__Silver
  TTLSConnectionAdvancedParmsRef  cAdv1~NDV_Server
} 
TTLSConnectionAdvancedParms       cAdv1~NDV_Server
{ 
  CertificateLabel                NDV_TEST_CERT
} 
TTLSKeyringParms                  keyR1
{ 
  Keyring                         /u/admin/CERT.kdb
  KeyringStashFile                /u/admin/CERT.sth
} 
TTLSCipherParms                   cipher1~AT-TLS__Silver
{
  V3CipherSuites                  TLS_RSA_WITH_DES_CBC_SHA
  V3CipherSuites                  TLS_RSA_WITH_3DES_EDE_CBC_SHA
  V3CipherSuites                  TLS_RSA_WITH_AES_128_CBC_SHA
} 
IpAddrSet                         addr1
{ 
  Prefix                          0.0.0.0/0
} 
PortRange                         portR1
{ 
  Port                            4843
} 
PortRange                         portR2
{ 
  Port                            1024-65535
}

How to Verify AT-TLS Configuration?

Check Policy-Agent job output JESMSGLG for:

EZZ8771I PAGENT CONFIG POLICY PROCESSING COMPLETE FOR <your TCP/IP address space>: TTLS

This message indicates a successful initialization.

Check Policy-Agent job output JESMSGLG for:

EZZ8438I PAGENT POLICY DEFINITIONS CONTAIN ERRORS FOR <your TCP/IP address space>: TTLS

This message indicates errors in the configuration file. Check the syslog.log file for further information.

Does the configuration rule cover the server?

Try to connect the server and check syslog.log for:

EZD1281I TTLS Map   CONNID: 00018BED LOCAL: 10.20.91.61..4843 REMOTE: 10.20.160.47..4889 JOBNAME: NDVDEVvr USERID: NDVSRV TYPE: InBound STATUS: Enabled RULE: ConnRule01~1 ACTIONS: gAct1 eAct1~NDV_Server cAct1~NDV_Server

The above entry indicates that the connection to Port 4843 is SSL/TLS enabled.

Frequently Asked Questions

Is there more information about problem determination?

See z/OS Communications Server, IP Diagnosis Guide, Chapter: Diagnosing Application Transparent Transport Layer Security (AT-TLS)

How to switch on P-agent trace?

See z/OS Communications Server, IP Configuration Reference, Chapter: Syslog deamon, and z/OS Communications Server, IP Configuration Guide, Chapter: Base TCP/IP system – TCP/IP Customization - Configuring the syslog daemon (syslogd).

Error at connection establishment

Find return code RC and corresponding GSK_ function name in P-agent trace.

Get description of the return code from z/OS Cryptographic Services, System Security Sockets Layer Programming, Chapter: : messages and codes – SSL function return codes.

Sample trace with trace=255:

EZD1281I TTLS Map   CONNID: 00002909 LOCAL: 10.20.91.61..1751 REMOTE: 10.20.91.117..443 JOBNAME: KSP USERID: KSP TYPE: OutBound STATUS: A
EZD1283I TTLS Event GRPID: 00000003 ENVID: 00000000 CONNID: 00002909  RC:    0 Connection Init 
EZD1282I TTLS Start GRPID: 00000003 ENVID: 00000002 CONNID: 00002909 Initial Handshake ACTIONS: gAct1 eAct1 AllUsersAsClient HS-Client   
EZD1284I TTLS Flow  GRPID: 00000003 ENVID: 00000002 CONNID: 00002909  RC:    0 Call GSK_SECURE_SOCKET_OPEN - 7EE4F718 
EZD1284I TTLS Flow  GRPID: 00000003 ENVID: 00000002 CONNID: 00002909  RC:    0 Set GSK_SESSION_TYPE -  CLIENT 
EZD1284I TTLS Flow  GRPID: 00000003 ENVID: 00000002 CONNID: 00002909  RC:    0 Set GSK_V3_CIPHER_SPECS -  090A2F 
EZD1284I TTLS Flow  GRPID: 00000003 ENVID: 00000002 CONNID: 00002909  RC:    0 Set GSK_FD - 00002909 
EZD1284I TTLS Flow  GRPID: 00000003 ENVID: 00000002 CONNID: 00002909  RC:    0 Set GSK_USER_DATA - 7EEE9B50 
EZD1284I TTLS Flow  GRPID: 00000003 ENVID: 00000002 CONNID: 00002909  RC:  435 Call GSK_SECURE_SOCKET_INIT - 7EE4F718 
EZD1283I TTLS Event GRPID: 00000003 ENVID: 00000002 CONNID: 00002909  RC:  435 Initial Handshake 00000000 7EEE8118  
EZD1286I TTLS Error GRPID: 00000003 ENVID: 00000002 CONNID: 00002909 JOBNAME: KSP USERID: KSP RULE: ConnRule01  RC:  435 Initial Handshake
EZD1283I TTLS Event GRPID: 00000003 ENVID: 00000002 CONNID: 00002909  RC:    0 Connection Close 00000000 7EEE8118 

Generation of a Natural development Server Certificate

Under z/OS, SSL/TLS certificates can be produced with the Linux System Services utility GSKKYMAN. The following steps have to be executed for the production of a new certificate, which is to be used for the SSL/TLS secured communication between Natural Web I/O Interface client and server:

1. Start a shell session out of TSO or connect via telnet to the z/OS Linux shell.

2. Start GSKKYMAN.

3. Create a new – or open an existing key database.

4. Create a self-signed certificate, for example, of type "User or server certificate with 2048-bit RSA key".

5. Export the certificate to a (HFS) file. Choose Base64 ASN.1 DER as export format.

This generated file can now be copied to the Natural development server client(s) using FTP with ASCII transfer format. On the client side, the received file should be stored with the file name suffix .CER. The certificate can now be used by the Natural development server client.

For more information on how to import the certificate, see the documentation of NaturalONE, Natural for Ajax > Configuration and Administration Configuring SSL.

If certificates are kept in a RACF key ring, the generated certificate has to be imported into the appropriate key ring using the RADCERT command.

Certificates, which are produced on a different platform, for example, on a Windows PC, can be imported into a RACF key ring or into a key database as well.

Detailed information about the use of the GSKKYMAN utility can be found in the IBM Communications server documentation, e.g. in the following manuals:

z/OS Communications Server, IP Configuration Guide

or

z/OS Cryptographic Services, System Secure Sockets Layer Programming

For the generation of certificates under Windows, a free downloadable utility named Ikeyman is available on several websites. Ikeyman is an IBM product as well and maps the functionality of GSKKYMAN to the Windows platform.

Configuring Port Sharing

The TCP/IP port sharing component of the z/OS TCP/IP Communications subsystem allows multiple listeners to listen on the same combination of port and IP address. To enable this in the TCP/IP stack, you must add the SHAREPORT or SHAREPORTWLM keyword to the PORT profile statement that reserves the TCP port for the server instances. You can find detailed information in section PORT Statement, chapter TCP/IP profile (PROFILE.TCPIP) and Configuration Statements of the IBM manual z/OS Communications Server: IP Configuration Reference found on their official website.

You can start the NDV server as explained in section Starting the Natural Development Server. You must define the used port with the PORT_NUMBER parameter in the configuration file of the NDV server. If port sharing is enabled, the started task can be executed multiple times to start multiple instances of an NDV server that listens on the same port.

• Server ID

• Reliability

• Scheduling

• TRACE, STDOUT, and STDERR

Server ID

Since the server ID is used to identify a server in the server directory, you must set a unique server ID. The server ID is passed as a parameter to the NDV server. You must set the parameters in the PARM card of the started task. For example:

//NDVSRV  EXEC PGM=NATRDEVS,REGION=0M,TIME=20,
//  PARM=('POSIX(ON) STACK(64K,32K,ANY,FREE) TRAP(ON,NOSPIE) ',
//    'TERMTHDACT(UADUMP),ENVAR(TZ=CET+0DST) /SERVERID')

Reliability

If multiple NDV servers are listening on a port and one server fails, the workload can still be done by the remaining servers. From a user point of view, the service offered by the NDV server remains available.

Scheduling

If a request for a new connection arrives, it is in the responsibility of the z/OS operating system to pass the request to one of the listening servers. The connections are distributed across the available servers using a weighted round-robin distribution based on the Servers' accept Efficiency Fractions (SEFs). Once a session is started, it is executed on this server, until it is finished.

You can identify a session in the trace file with the following piece of code:

Start of Session:
26 10:16:45 0000000E Worker task active, SID=DED9D087D842B606, client: USER

End of session:
26 10:21:53 0000000E Worker task disconnect, client: USER
26 10:21:53 0000000E Free LTCB...and finish

TRACE, STDOUT, and STDERR

Each server has its own STGTRACE, STGSTDO, and STDSTGE, which are part of the job log of each server. It is possible to redirect the output of the files to a sequential dataset, but you must ensure that each server is assigned its own sequential dataset. Otherwise, the output of the files cannot be assigned to a particular server.