All about IT Governance

For those who are not following  Symantec in the past 6 months, you may be in for a surprise that EDS was the appointed outsource vendor for taking over most portions of the IT. Although, the details are a bit sketchy, it is our belief that a minimal IT is being owned by Symantec and the rest of IT services including client services, help desk services, infrastructure management and core business applications management is handed out to EDI.

The following are the lessons we have gathered:

a. There are 3 schools of outsourcing:

1. Outsource business secondary services (helpdesk, client services etc.)

2. Outsource a portion of your growth services (Internal QA, Internal Development for IT)

3. Full-outsource: Outsource everything except Governance (including IT Security), Compliance and Vendor Management

We believe what has hapenned at Symantec is #3 above. It is a fairly large responsibility however, it is very common in Hospital environment.

b. Symantec & EDS are now in a wedlock - it is hard to even imagine looking back

c. We strongly believe, if the company is in either Run or Grow mode (not transform mode), then ‘full-outsource’ may work.

d. EDS does not come cheap however, if Symantec strongly believes in ‘Grow’ mode, maybe IT was a distraction (I know, I can’t believe I am saying it).

e. Company employees would have to be transitioned in various packages however, it may be hard to build a team ever again to build out IT - so skeletal team within above areas and ‘hard to find’ liaison people would have to be retained.

There are a few deployment strategies and would depend on the ’status’ of your IT applications and ‘posture’ of the current DR deployment (if any).

I will describe the status and posture later but lets assume that most of our Business Applications are deployed (and not being built/delivered) - then, one of the deployment strategies we have seen is: Start a DR site with ‘DR ecosystem’ and ‘Testing/Staging ecosystem’ in the same datacenter.  As you build the DR ecosystem, you can test failovers on testing/staging ecosystem with DR ecosystem. On reaching sufficient comfort, start fail-overs between primary-to-DR site. Treat this as a baseline and any additions, you first test between DR ecosystem and test ecosystem and finally test with primary site. This is a great strategy as long as your IT ecosystem is fairly stable.

The assumption made above is what I call ’status’ - term refered to status of the business applications - IT is always deploying upgrades one way or another - its hard to see a ’stable’ status.

In addition to the status, you could also have a ‘temporary’ DR site and that is what I call ‘posture’ - term used to refer to DR ecosystem status. In the above scenario, the posture of the DR site is completely new.

Now back to another example, if the status of IT is half-ready and DR posture is ‘temporary’, it might be best to do this:  Start creating a clone of the existing temporary DR site (assuming it is deployed right) at the new site - start migrating application by application to the new site. Essentially, you are doing DR-to-DR site migration and performing fail-over between these 2 sites. This would also enable you to not dump all the investments in the old DR site. While you are doing this, new deployments should be done in the new IT ecosystem and perform fail-over tests between the primary and new DR site. Once the old DR site is ready to be retired, start the fail-over testing between new DR site and the primary site.  Whoooosh!

If the IT status is pretty stable and posture is fully ready however you need a new DR site for logistics reason (which is very rare), you may want to consider building a DR site (which is almost or better clone of the old DR site) and perform Primary-to-New-DR-failovers and build this one application at a time.

If you need additional perspectives, please let me know. Maybe, I can help.

What is the bare minimal risks that your ITGC SOX controls should cover? Although there are ITGI/COBIT documents to guide you through, one can easily get lost in the maze - I thought, I will put together this checklist for anybody to apply this to get a comfort. 

a. New accounts or modifications to account privileges are approved - OS, DB, App, Network, Logical Security (VPN, ActiveDirectory, NIS, LDAP and other authentication sources),  Physical security.

b. Terminations are performed timely - applicable to same processes as above

c. Transfers are handled as combination of b. and a. above - applicable to same processes as a.

d. Controls around change management to specifically address database object level changes, data fixes, data migrations and UATs are recorded for all application level changes

e.  Shared privilege level account management - passwords are changed when team members having access to shared privileged accounts leave the company - applicable to all areas listed in a.

f. Storage backup is tested for recovery (especially for databases)

g. Password configuration complies with enterprise standards for areas listed in a.

Optional:

a.  Account reviews in each area listed in a.

That would depend on how you deploy it.

With the right set of policy and deployment standards, the cost can be zeroed right from the initial days of deployment and you need to strike a balance between the risk vs. the benefit - benefit as in agile environment, risk as in exposure to security threats.

For example one of the policy could be:

No sharing of virtualized zones between business applications. Say, ERP and HR applications.

With this, you reduce the security exposures and leverage existing set of controls (assuming existing controls are strong). However, here is where you strike a balance between risk vs. benefit. You lose a bit of agility of virtualization however it is well worth it.

Another policy could be:

All changes to the virtualized zones/clusters align with the existing change processes with appropriate approvals including access changes. Again, assuming there is a strong change process and security related changes are well addressed.

This would help increase awareness and reduce the ‘knee jerk’ reaction to make changes to virtualized zones.

Another policy could be:

Migration to production (regardless of OS, Database, Applications, interfaces, Network connectivity, storage changes) complies with Security checklist - this checklist ensures securing the base operating system (hardening), physical and logical access to console, admin access to each zones and so on for each associated layers.

Another policy could be:

Logs of changes are maintained in the system (global zone and virutal zone) and is audited periodically.

This is to ensure that potential exposures are detected and addressed.

Overall, in a typical public company, almost all of the above controls would exist (existence is one thing, execution is another) excepting the last audit of logs control.

Here is an interesting conversation that started me to write this blog by Christofer Hoff at http://rationalsecurity.typepad.com/blog/2008/05/virtualizing-se.html

He has raised several key points of which one of them is SOD - I agree with his concerns on SOD and with a strong policy and procedures, it can be addressed and is not impossible.