If your emphasis has been on acquiring the skills to put your company on the Web, you may want to ask if you have the necessary knowledge of system and network security to protect your firm once it gets there.
With the growing emphasis on connecting IBM i systems to the Internet, one has to ask if the current focus on developers' tools to get there isn't tantamount to educating a 10-year-old on alternative trails for skateboarding over to the interstate highway. If you don't know how to secure the system once you're online, is it really wise to keep accelerating up the onramp?
Despite the reputation that IBM i on Power has for security as a standalone system, when connecting to the Internet, there are so many opportunities for compromise that no system administrator who isn't a bona fide security expert can be certain of the outcome.
Management's directives to move forward with connecting company's networks, and sensitive IBM i data, to the Internet could be like lighting a long fuse leading to a stick of dynamite. The recent successful hacker intrusion and identity theft from Epsilon, an email marketing firm that contracts with Best Buy, Disney, Ralph's, and many others, is only the tip of the iceberg. Hackers today are after the corporate jewels, important proprietary company information that can be resold on the black market.
IBM i system administrators are generally a cautious group who have not eagerly turned their IBM i servers into busy hubs for Web-based services, despite the tools and capabilities of the systems to do so. Despite the added programming complexity and need to learn new skills such as HTML, JavaScript, and PHP for Web development, the real concern percolating up from the subconscious is, "Once we get there, can I keep this data secure?" Unless that same system administrator is a highly sophisticated security architect, the answer very well could be, "no."
There are several IBM Redbooks on this topic that anyone responsible for data security on an Internet-connected system should read, if they haven't already. And it may be that most administrators already have downloaded these free books, since both have been available for quite some time. The first is called AS/400 Internet Security: Protecting Your AS/400 from HARM in the Internet, and the second is titled AS/400 Internet Security Scenarios: A Practical Approach. Both have the distinction of being either authored or co-authored by Marcela Adan, a remarkable IBM trainer and writer on security-related topics. The latter title was a collaborative effort by Adan and a half a dozen other IBM security experts from around the world.
Before the books, which are 274 and 432 pages, respectively, go into detail of exactly how you can secure your system, they stress the need for a formal security policy that takes the approach that security must be layered in order to be effective. If one security barrier is breached, others are still in place to blunt an attack. The assumption is that there are skilled people with time on their hands who are actively seeking access to your data.
In AS/400 Internet Security: Protecting Your AS/400 from HARM in the Internet, the authors write, "In this book, we take a layered approach to securing your AS/400 system when attaching it to the Internet. We focus on system and application security. We strongly recommend not connecting the system to the Internet until you are 100 percent sure that you have thoroughly reviewed OS/400 security and that the TCP/IP applications you have chosen to use across the Internet are properly and securely configured."
I find it amusing that they don't say, "We strongly recommend not connecting the system to the Internet until your boss orders you to—regardless of the consequences!" While the book doesn't focus on network security per se, the authors acknowledge that network security is a part of Internet security and do treat how to configure firewalls.
One approach to reducing risk that the authors cite is to use the services of an Internet Service Provider (ISP) to store, serve, and maintain your company Web pages. The other is to contract for an email gateway with an Internet provider such as IBM Global Network (IGN). You connect the IBM i to the Internet via SNA link using SNA protocols to forward and receive email through the gateway. This way, the IBM i is protected because it's not connected directly to the Internet; with no Internet, there is no risk. They feel this is better than having a standalone server serve up Web pages since in that case, the company Web site is still vulnerable to being defaced or infected with invisible malware intended for unsuspecting visitors.
Questions to be answered before connecting to the Internet include these:
- What are your Internet applications?
- How do the users access your Internet server?
- What are the security objectives?
- What are the security measures to implement?
- Is there still a risk? If yes, what is the probability and potential damage?
Because connecting to the Internet opens not only one, but many points where a hacker could enter your system, it is important to identify those points where security is compromised and devise a layered approach, equivalent to multiple locks, to help protect it. Even if one lock is broken, the others are still in place to protect your system.
The authors point out that there are basically four separate areas where you should apply security measures:
- Network security: Controlling access to the server
- Application security: Security that is specific to an application, such as FTP or TELNET. For instance, do you want to enable only anonymous users, or will you require a user ID and password?
- Transaction security: This is where you ensure data privacy and authentication of business partners.
- System security: Take advantage of IBM i's inherent security features to lock down access to the data by using user authorities and other built-in features.
It is no secret that many IBM i shops still have their systems set to default security settings that came from the factory. Presumably, this is because the system administrator, who might be the business owner, has neither the time nor the interest in learning how the system's security tools really work. If this is a condition that isn't likely to change, then you probably don't want to connect this system to the Internet—yet. If you don't know how to adjust system-wide security values, manage user profiles and passwords, manipulate resource security, and define general TCP/IP definitions, then it's too early to connect your system to the Internet.
The catch-22 in opening up your system to the risks of the Internet is that while doing so introduces a level of complexity to your security that you may not have been familiar with in the past, keeping it simple is one way to reduce the chances of overlooking some vulnerability.
So learning the tools and languages that can seamlessly take your company to the Web is actually step two in taking advantage of the opportunities that await your company on the Internet. Step one is becoming your own in-house security expert, and the two aforementioned IBM Redbooks are a good place to start.
as/400, os/400, iseries, system i, i5/os, ibm i, power systems, 6.1, 7.1, V7, V6R1
LATEST COMMENTS
MC Press Online