Software Development on the SAP HANA Platform — Save 50%
Unlock the true potential of the SAP HANA platform with this book and ebook
In this article by Mark Walker, author of Software Development on the SAP HANA Platform, we will examined the fact that some of your users might not be allowed to have access to all the data in the system, or might have specific rights, maybe to create data but not to see the data that is currently in the system. We'll see concrete example of access restrictions on both objects and their content in this article.
SAP HANA provides several different layers of security from which you can benefit, and in this article we'll be looking at each of them. We'll examine how we can create each type of object, how they interact with each other, and how we can use them.
The authorization concept in SAP HANA is comprised of different notions:
We will take a look at each of these concepts in turn.
(For more resources related to this topic, see here.)
In SAP HANA, as in most of SAP's software, authorizations are grouped into roles.
A role is a collection of authorization objects, with their associated privileges. It allows us, as developers, to define self-contained units of authorization. In the same way that at the start of this book we created an attribute view allowing us to have a coherent view of our customer data which we could reuse at will in more advanced developments, authorization roles allow us to create coherent developments of authorization data which we can then assign to users at will, making sure that users who are supposed to have the same rights always have the same rights. If we had to assign individual authorization objects to users, we could be fairly sure that sooner or later, we would forget someone in a department, and they would not be able to access the data they needed to do their everyday work. Worse, we might not give quite the same authorizations to one person, and have to spend valuable time correcting our error when they couldn't see the data they needed (or worse, more dangerous and less obvious to us as developers, if the user could see more data than was intended).
It is always a much better idea to group authorizations into a role and then assign the role to users, than assign authorizations directly to users.
Assigning a role to a user means that when the user changes jobs and needs a new set of privileges; we can just remove the first role, and assign a second one.
Since, we're just starting out using authorizations in SAP HANA, let's get into this good habit right from the start. It really will make our lives easier later on.
Creating a role
Role creation is done, like all other SAP HANA development, in the Studio. If your Studio is currently closed, please open it, and then select the Modeler perspective.
In order to create roles, privileges, and users, you will yourself need privileges. Your SAP HANA user will need the ROLE ADMIN, USER ADMIN, and CREATE STRUCTURED PRIVILEGE system privileges in order to do the development work in this article.
You will see in the Navigator panel we have a Security folder, as we can see here:
Please find the Security folder and then expand this folder. You will see a subfolder called Roles. Right-click on the Roles folder and select New Role to start creating a role.
On the screen which will open, you will see a number of tabs representing the different authorization objects we can create, as we can see here:
We'll be looking at each of these in turn, in the following sections, so for the moment just give your role Name (BOOKUSER might be appropriate, if not very original).
Like many other object types in SAP HANA, once you have created a role, you can then use it inside another role. This onion-like arrangement makes authorizations a lot easier to manage.
If we had, for example, a company with two teams:
And two countries, say:
We could create a role giving access to sales analytic views, one giving purchasing analytic views, one giving access to data for France, and one giving access to data for Germany. We could then create new roles, say Sales-France, which don't actually contain any authorization objects themselves, but contain only the Sales and the France roles.
The role definition is much simpler to understand and to maintain than if we had directly created the Sales-France role and a Sales-Germany role with all the underlying objects.
Once again, as with other development objects, creating small self-contained roles and reusing them when possible will make your (maintenance) life easier.
In the Granted Roles tab we can see the list of subroles this main role contains. Note that this list is only a pointer, you cannot modify the actual authorizations and the other roles given here, you would need to open the individual role and make changes there.
Part of roles
The Part of Roles tab in the role definition screen is exactly the opposite of the Granted Roles tab.
This tab lists all other roles of which this role is a subrole. It is very useful to track authorizations, especially when you find yourself in a situation where a user seems to have too many authorizations and can see data they shouldn't be able to see.
You cannot manipulate this list as such, it exists for information only. If you want to make changes, you need to modify the main role of which this role is a subrole.
An SQL privilege is the lowest level at which we can define restrictions for using database objects.
SQL privileges apply to the simplest objects in the database such as schemas, tables and so on. No attribute, analytical, or calculation view can be seen by SQL privileges.
This is not strictly true, though you can consider it so. What we have seen as an analytical view, for example, the graphical definition, the drag and drop, the checkboxes, has been transformed into a real database object in the _SYS_BIC schema upon activation. We could therefore define SQL privileges on this database object if we wanted, but this is not recommended and indeed limits the control we can have over the view. We'll see a little later that SAP HANA has much finer-grained authorizations for views than this.
An important thing to note about SQL privileges is that they apply to the object on which they are defined. They restrict access to a given object itself, but do not at any point have any impact on the object's contents. For example, we can decide that one of our users can have access to the CUSTOMER table, but we couldn't restrict their access to only CUSTOMER values from the COUNTRY USA.
SQL privileges can control access to any object under the Catalog node in the Navigator panel.
Let's add some authorizations to our BOOK schema and its contents. At the top of the SQL Privileges tab is a green plus sign button. Now click on this button to get the Select Catalog Object dialog, shown here:
As you can see in the screenshot, we have entered the two letters bo into the filter box at the top of the dialog. As soon as you enter at least two letters into this box, the Studio will attempt to find and then list all database objects whose name contains the two letters you typed. If you continue to type, the search will be refined further.
The first thing to notice is the warning icon on the SQL Privileges tab itself:
This means that your role definition is incomplete, and the role cannot be activated and used as yet.
On the right of the screen, a list of checkbox options has appeared. These are the individual authorizations appropriate to the SQL object you have selected. In order to grant rights to a user via a role, you need to decide which of these options to include in the role.
The individual authorization names are self-explicit. For example, the CREATE ANY authorization allows creation of new objects inside a schema.
The INSERT or SELECT authorization might at first seem unusual for a schema, as it's not an object which can support such instructions. However, the usage is actually quite elegant. If a user has INSERT rights on the schema BOOK, then they have INSERT rights on all objects inside the schema BOOK.
Granting rights on the schema itself avoids having to specify the names of all objects inside the schema. It also future-proofs your authorization concept, since new objects created in the schema will automatically inherit from the existing authorizations you have defined.
On the far right of the screen, alongside each authorization is a radio button which gives an additional privilege, the possibility for a given user to, in turn, give the rights to a second user.
This is an option which should not be given to all users, and so should not be present in all roles you create; the right to attribute privileges to users should be limited to your administrators. If you give just any user the right to pass on their authorizations further, you will soon find that you are no longer able to determine who can do what in your database.
For the moment we are creating a simple role to show the working of the authorization concept in SAP HANA, so we will check all the checkboxes, and leave the radio buttons at No :
There are some SQL privileges which are necessary for any user to be able to do work in SAP HANA. These are listed below. They give access to the system objects describing the development models we create in SAP HANA, and if a user does not have these privileges, nothing will work at all, the user will not be authorized to do anything.
The SQL privileges you will need to add to the role in order to give access to basic SAP HANA system objects are:
The SELECT privilege on the _SYS_BI schema
The SELECT privilege on the _SYS_REPO schema
The EXECUTE privilege on the REPOSITORY_REST procedure
Please add these SQL privileges to your role now, in order to obtain the following result:
As you can see with the configuration we have just done, SQL privileges allow a user to access a given object and allow specific actions on the object. They do not however allow us to specify particular authorizations to the contents of the object.
In order to use such fine-grained rights, we need to create an analytic privilege, and then add it to our role, so let's do that now.
eBook Price: $35.99
Book Price: $59.99
An analytic privilege is an artifact unique to SAP HANA, it is not part of the standard SQL authorization concept.
Analytic privileges allow us to restrict access to certain values of a given attribute, analytic, or calculation view. This means that we can create one view, which by default shows all available data, and then restrict what is actually visible to different users. We could restrict visible data by company code, by country, or by region. For example, our users in Europe would be allowed to see and work with data from our customers in Europe, but not those in the USA.
An analytic privilege is created through the Quick Launch panel of Modeler , so please open that view now (or switch to the Quick Launch tab if it's already open).
You don't need to close the role definition tab that's already open, we can leave it for now, create our analytic privilege, and then come back to the role definition later.
From the Quick Launch panel, select Analytic Privilege , and then Create .
As usual with SAP HANA, we are asked to give Name , Description , and select a package for our object. We'll call it AP_EU (for analytic privilege, Europe), use the name as the description, and put it into our book package alongside our other developments. As is common in SAP HANA, we have the option of creating an analytic privilege from scratch (Create New ) or copying an existing privilege (Copy From ). We don't currently have any other analytic privileges in our development, so leave Create New selected, then click on Next to go to the second screen of the wizard, shown here:
On this page of the dialog, we are prompted to add development models to the analytic privilege. This will then allow us to restrict access to given values of these models. In the previous screenshot, we have added the CUST_REV analytic view to the analytic privilege. This will allow us to restrict access to any value we specify of any of the fields visible in the view.
To add a view to the analytic privilege, just find it in the left panel, click on its name and then click on the Add button. Once you have added the views you require for your authorizations, click on the Finish button at the bottom of the window to go to the next step.
You will be presented with the analytic privilege development panel, reproduced here:
This page allows us to define our analytic privilege completely. On the left we have the list of database views we have included in the analytic privilege. We can add more, or remove one, using the Add and Remove buttons. To the right, we can see the Associated Attributes Restrictions and Assign Restrictions boxes. These are where we define the restrictions to individual values, or sets of values. In the top box, Associated Attributes Restrictions , we define on which attributes we want to restrict access (country code or region, maybe). In the bottom box, Assign Restrictions , we define the individual values on which to restrict (for example, for company code, we could restrict to value 0001, or US22; for region, we could limit access to EU or USA).
Let's add a restriction to the REGION field of our CUST_REV view now. Click on the Add button next to the Associated Attributes Restrictions box, to see the Select Object dialog:
As can be expected, this dialog lists all the attributes in our analytic view. We just need to select the appropriate attribute and then click on OK to add it to the analytic privilege.
Measures in the view are not listed in the dialog. We cannot restrict access to a view according to numeric values. We cannot therefore, make restrictions to customers with a revenue over 1 million Euros, for example.
Please add the REGION field to the analytic privilege now.
Once the appropriate fields have been added, we can define the restrictions to be applied to them. Click on the REGION field in the Associated Attributes Restrictions box, then on the Add button next to the Assign Restrictions box, to define the restrictions we want to apply.
As we can see, restrictions can be defined according to the usual list of comparison operators. These are the same operators we used earlier to define a restricted column in our analytic views.
In our example, we'll be restricting access to those lines with a REGION column equal to EU, so we'll select Equal .
In the Value column, we can either type the appropriate value directly, or use the value help button, and the familiar Value Help Dialog which will appear, to select the value from those available in the view. Please add the EU value, either by typing it or by having SAP HANA find it for us, now.
There is one more field which needs to be added to our analytic privilege, and the reason behind might seem at first a little strange.
This point is valid for SAP HANA SP5, up to and including (at least) release 50 of the software. If this point turns out to be a bug, then it might not be necessary in later versions of the software.
The field on which we want to restrict user actions (REGION) is not actually part of the analytic view itself. REGION, if you recall, is a field which is present in CUST_REV , thanks to the included attribute view CUST_ATTR . In its current state, the analytic privilege will not work, because no fields from the analytic view are actually present in the analytic privilege.
We therefore need to add at least one of the native fields of the analytic view to the analytic privilege. We don't need to do any restriction on the field; however it needs to be in the privilege for everything to work as expected.
This is hinted at in SAP Note 1809199, SAP HANA DB: debugging user authorization errors.
Only if a view is included in one of the cube restrictions and at least one of its attribute is employed by one of the dimension restrictions, access to the view is granted by this analytical privilege.
Not an explicit description of the workings of the authorization concept, but close.
Our analytic view CUST_REV contains two native fields, CURRENCY and YEAR. You can add either of these to the analytic privilege. You do not need to assign any restrictions to the field; it just needs to be in the privilege.
Here is the state of the analytic privilege when development work on it is finished:
The Count column lists the number of restrictions in effect for the associated field. For the CURRENCY field, no restrictions are defined.
We just need (as always) to activate our analytic privilege in order to be able to use it. The activation button is the same one as we have used up until now to activate the modeling views, the round green button with the right-facing white arrow at the top-right of the panel, which you can see on the preceding screenshot.
Please activate the analytic privilege now.
Once that has been done, we can add it to our role. Return to the Role tab (if you left it open) or reopen the role now.
If you closed the role definition tab earlier, you can get back to our role by opening the Security node in the Navigator panel, then opening Roles, and double-clicking on the BOOKUSER role.
In the Analytic Privileges tab of the role definition screen, click on the green plus sign at the top, to add an analytic privilege to our role. The analytic privilege we have just created is called AP_EU, so type ap_eu into the search box at the top of the dialog window which will open. As soon as you have typed at least two characters, SAP HANA will start searching for matching analytic privileges, and your AP_EU privilege will be listed, as we can see here:
Click on OK to add the privilege to the role. We will see in a minute the effect our analytic privilege has on the rights of a particular user, but for the moment we can take a look at the second-to-last tab in the role definition screen, System Privileges .
As its name suggests, system privileges gives to a particular user the right to perform specific actions on the SAP HANA system itself, not just on a given table or view.
These are particular rights which should not be given to just any user, but should be reserved to those users who need to perform a particular task. We'll not be adding any of these privileges to our role, however we'll take a look at the available options and what they are used for.
Click on the green plus-sign button at the top of the System Privileges tab to see a list of the available privileges. By default the dialog will do a search on all available values; there are only fifteen or so, but you can as usual filter them down if you require using the filter box at the top of the dialog:
The last tab in the role definition screen concerns Package Privileges . These allow a given user to access those objects in a package.
In our example, the package is called book, so if we add the book package to our role in the Package Privileges tab, we will see the following result:
Assigning package privileges is similar to assigning SQL privileges we saw earlier. We first add the required object (here our book package), then we need to indicate exactly which rights we give to the role. As we can see in the preceding screenshot, we have a series of checkboxes on the right-hand side of the window. At least one of these checkboxes must be checked in order to save the role.
The individual rights have names which are fairly self-explanatory. REPO.READ gives access to read the package, whereas REPO.EDIT_NATIVE_OBJECTS allows modification of objects, for example.
The role we are creating is destined for an end user who will need to see the data in a role, but should not need to modify the data models in any way (and in fact we really don't want them to modify our data models, do we?). We'll just add the REPO.READ privilege, on our book package, to our role. Again we can decide whether the end user can in turn assign this privilege to others. And again, we don't need this feature in our role.
At this point, our role is finished. We have given access to the SQL objects in the BOOK schema, created an analytic privilege which limits access to the Europe region in our CUST_REV model, and given read-only access to our book package. After activation (always) we'll be able to assign our role to a test user, and then see the effect our authorizations have on what the user can do and see. Please activate the role now.
Users are probably the most important part of the authorization concept. They are where all our problems begin, and their attempts to do and see things they shouldn't are the main reason we have to spend valuable time defining authorizations in the first place.
In technical terms, a user is just another database object. They are created, modified, and deleted in the same way a modeling view is. They have properties (their name and password, for example), and it is by modifying these properties that we influence the actions that the person who connects using the user can perform.
Up until now we have been using the SYSTEM user (or the user that your database administrator assigned to you). This user is defined by SAP, and has basically the authorizations to do anything with the database.
Use of this user is discouraged by SAP, and the author really would like to insist that you don't use it for your developments. Accidents happen, and one of the great things about authorizations is that they help to prevent accidents. If you try to delete an important object with the SYSTEM user, you will delete it, and getting it back might involve a database restore. If however you use a development user with less authorization, then you wouldn't have been allowed to do the deletion, saving a lot of tears.
Of course, the question then arises, why have you been using the SYSTEM user for the last couple of hundred pages of development. The answer is simple: if the author had started the book with the authorizations article, not many readers would have gotten past page 10.
Let's create a new user now, and assign the role we have just created.
From the Navigator panel, open the Security node, right-click on User , and select New User from the menu to obtain the user creation screen as shown in the following screenshot:
Defining a user requires remarkably little information:
User Name : The login that the user will use. Your company might have a naming convention for users. Users might even already have a standard login they use to connect to other systems in your enterprise. In our example, we'll create a user with the (once again rather unimaginative) name of BOOKU.
Authentication : How will SAP HANA know that the user connecting with the name of ANNE really is Anne? There are three (currently) ways of authenticating a user with SAP HANA.
Password : This is the most common authentication system, SAP HANA will ask Anne for her password when she connects to the system. Since Anne is the only person who knows her password, we can be sure that Anne really is ANNE, and let her connect and do anything the user ANNE is allowed to do. Passwords in SAP HANA have to respect a certain format. By default this format is one capital, one lowercase, one number, and at least eight characters.
You can see and change the password policy in the system configuration. Double-click on the system name in the Navigator panel, click on the Configuration tab, type the word pass into the filter box at the top of the tab, and scroll down to indexserver.ini and then password policy . The password format in force on your system is listed as password_layout . By default this is A1a, meaning capitals, numbers, and lowercase letters are allowed. The value can also contain the # character, meaning that special characters must also be contained in the password. The only special characters allowed by SAP HANA are currently the underscore, dollar sign, and the hash character. Other password policy defaults are also listed on this screen, such as maximum_password_lifetime (the time after which SAP HANA will force you to change your password).
Kerberos and SAML : These authentication systems need to be set up by your network administrator and allow single sign-on in your enterprise. This means that SAP HANA will be able to see the Windows username that is connecting to the system. The database will assume that the authentication part (deciding whether Anne really is ANNE) has already been done by Windows, and let the user connect.
Session Client : As we saw when we created attribute and analytic views back at the start of the book, SAP HANA understands the notion of client, referring to a partition system of the SAP ERP database. In the SAP ERP, different users can work in different Clients. In our development, we filtered on Client 100. A much better way of handling filtering is to define the default client for a user when we define their account. The Session Client field can be filled with the ERP Client in which the user works. In this way we do not need to filter on the analytic models, we can leave their client value at Dynamic in the view, and the actual value to use will be taken from the user record. Once again this means maintenance of our developments is a lot simpler.
If you like, you can take a few minutes at the end of this article to create a user with a session client value of 100, then go back and reset our attribute and analytic views' default client value to Dynamic, reactivate everything, and then do a data preview with your test user. The result should be identical to that obtained when the view was filtered on client 100. However, if you then create a second user with a session client of 200, this second user will see different data.
We'll create a user with a password login, so type a password for your user now. Remember to adhere to the password policy in force on your system. Also note that the user will be required to change their password on first login.
At the bottom of the user definition screen, as we can see from the preceding screenshot, we have a series of tabs corresponding to the different authorizations we can assign to our user. These are the same tabs we saw earlier when defining a role.
As explained at the beginning of this article, it is considered best practice to assign authorizations to a role and then the role to a user, rather than assign authorizations directly to a user; this makes maintenance easier. For this reason we will not be looking at the different tabs for assigning authorizations to our user, other than the first one, Granted Roles .
The Granted Roles tab lists, and allows adding and removing roles from the list assigned to the user. By default when we create a user, they have no roles assigned, and hence have no authorizations at all in the system. They will be able to log in to SAP HANA but will be able to do no development work, and will see no data from the system.
Please click on the green plus sign button in the Granted Roles tab of the user definition screen, to add a role to the user account. You will be provided with the Select Role dialog, shown in part here:
This dialog has the familiar search box at the top, so typing the first few letters of a role name will bring up a list of matching roles. Here our role was called BOOKUSER, so please do a search for it, then select it in the list and click on OK to add it to the user account.
Once that is done, we can test our user to verify that we can perform the necessary actions with the role and user we have just created. We just need, as with all objects in SAP HANA, to activate the user object first. As usual, this is done with the round green button with the right-facing white arrow at the top-right of the screen. Please do this now.
Testing our user and role
The only real way to check if the authorizations we have defined are appropriate to the business requirements is to create a user and then try out the role to see what the user can and cannot see and do in the system.
The first thing to do is to add our new user to the Studio so we can connect to SAP HANA using this new user.
To do this, in the Navigator panel, right click on the SAP HANA system name, and select Add Additional User from the menu which appears.
This will give you the Add additional user dialog, shown in the following screenshot:
Enter the name of the user you just created (BOOKU) and the password you assigned to the user. You will be required to change the password immediately:
Click on Finish to add the user to the Studio.
You will see immediately in the Navigator panel that we can now work with either our SYSTEM user, or our BOOKU user:
We can also see straight away that BOOKU is missing the privileges to perform or manage data backups; the Backup node is missing from the list for the BOOKU user.
Let's try to do something with our BOOKU user and see how the system reacts. The way the Studio lets you handle multiple users is very elegant, since the tree structure of database objects is duplicated, one per user, you can see immediately how the different authorization profiles affect the different users. Additionally, if you request a data preview from the CUST_REV analytic view in the book package under the BOOKU user's node in the Navigator panel, you will see the data according to the BOOKU user's authorizations. Requesting the same data preview from the SYSTEM user's node will see the data according to SYSTEM's authorizations.
Let's do a data preview on the CUST_REV view with the SYSTEM user, for reference:
As we can see, there are 12 rows of data retrieved, and we have data from the EU and NAR regions.
If we ask for the same data preview using our BOOKU user, we can see much less data:
BOOKU can only see nine of the 12 data rows in our view, as no data from the NAR region is visible to the BOOKU user. This is exactly the result we aimed to achieve using our analytic privilege, in our role, assigned to our user.
In this article, we have taken a look at the different aspects of the authorization concept in SAP HANA. We examined the different authorization levels available in the system, from SQL privileges, analytic privileges, system privileges, and package privileges. We saw how to add these different authorization concepts to a role, a reusable group of authorizations.
We went on to create a new user in our SAP HANA system, examining the different types of authentications available, and the assignment of roles to users.
Finally, we logged into the Studio with our new user account, and found out the first-hand effect our authorizations had on what the user could see and do.
Resources for Article :
- SAP Netweaver: Accessing the MDM System [Article]
- SAP HANA integration with Microsoft Excel [Article]
- Exporting SAP BusinessObjects Dashboards into Different Environments [Article]
eBook Price: $35.99
Book Price: $59.99
About the Author :
Mark Walker has been working with computers since a very early age, starting out with programming the Sinclair ZX Spectrum. After graduating from Cardiff University, he moved to France, starting a 17-year (so far) career in SAP development, first on the SAP ERP, then moving to SAP BW in 2004. He specializes in process optimization, especially in the SAP BPS and BI-IP modules. He's been following SAP HANA since its release in 2011, and is a certified HANA Application Associate since early 2012.
Mark has been working for the last six years for Censio, a consultancy in Paris, France, which has expertise in SAP, Microstrategy, Business Objects, and Microsoft BI technologies. During that time he's worked on several end-to-end projects for some prestigious clients, and done some missions in SAP's name.
In November 2012, he published SAP HANA Starter, Packt Publishing, a small book about getting a foothold in the SAP HANA world.