Research

We will close out this series by examining key decision criteria to help you select an API gateway. We offer a set of questions to determine which vendor solutions support your API technically, as well as the features your developers and administrators need. These criteria can be used to check solutions against your design goals and help you walk through the evaluation process. Nota bene: use cases first It is tempting to leap to a solution. After all, API development is a major trend, and security teams want to help solve API security problems. API gateways have been designed to enable developers to jump in quickly and easily. But there is no generic API security model good enough for all APIs. APIs are a glue layer, so the priorities and drivers are found by analyzing your API use cases: from what components you are gluing together, from what environment (enterprise, B2B, legacy, etc.), to what environment (mobile, Internet of Things, third-party developers, etc). This analysis provides crucial weighting for your priorities. Product Architecture Describe the API gateway’s deployment model (software, hardware only, hardware + software, cloud, or something else). Describe the scalability model. Does the API gateway scale horizontally or vertically? What connectors and adapters, to other software and cloud services, are included? How are new versions and updates handled? What key features do you believe your product has that distinguishes it from competitors? Access Provisioning and Developer Power Tools What credentials and tokens does the API gateway support for developers and API consumers? How is access governed? What monitoring, management, and metrics features does the gateway offer? Does the product offer client-side helper SDKs (iOS, Android, JavaScript, etc.) to simplify API consumer development? Describe a typical “day in the life” of a developer, from registering a new API to production operationalization. Describe out-of-the-box self-service features for registering new APIs. Describe out-of-the-box self-service features for acquiring API keys and tokens. Describe out-of-the-box self-service features for testing APIs. Describe out-of-the-box self-service features for versioning APIs. Describe how your API catalog help developers understand the available APIs and how to use them. Development What integration is available for source code and configuration management? For extending the product, what languages and tools are required to develop wrappers, adapters, and extensions? What continuous integration tools (e.g., Jenkins) does your product work with? Access Control How are API consumers authenticated? How are API calls from API consumers authorized? What level of authorization granularity is checked? Please describe where role, group, and attribute level authorization can be enforced. What out-of-the-box features does the API gateway have for access key issuance, distribution, and verification? What out-of-the-box features does the API gateway have for access key lifecycle management? What tools are used to define technical security policy? Describe support for delegated authorization. What identity server functionality is available in the API gateway? e.g., OAuth Authorization Server, OAuth Resource server, SAML Identity Provider, SAML Relying Party, XACML PEP, XACML PDP, … What identity protocol flows are supported, and what role does the API gateway play in them? Interoperability What identity protocols and versions are supported (OAuth, SAML, etc.)? What directories are supported (Active Directory, LDAP, etc.)? What application servers are supported (WebSphere, IIS, Tomcat, SAP, etc.)? What Service and Security gateways are supported (DataPower, Intel, Vordel, Layer7, etc.)? Which cloud applications are supported? Which mobile platforms supported? Security Describe support for TLS/SSL. Is client-side TLS/SSL (“2-way mutual authentication”) supported? How. Please describe the API gateway’s support for whitelisting URLs. What out-of-the-box functionality is in place to deal with injection attacks such as SQL injection? How does the product defend against malicious JavaScript? How does the gateway defend against URL redirect attacks? How does the gateway defend against replay attacks? What is the product’s internal security model? Is Role-Based Access Control supported? Where? How is access audited? Cost Model How is the product licensed? Does cost scale based on number of users, number of servers, or another criterion? What is the charge for adapters and extensions? This checklist offers a starting point for analyzing API gateway options. Review product capabilities to identify the best candidate, keeping in mind that integration is often the most important criterion for successful deployment. It is not as simple as picking the ‘best’ product – you need to find one that fits your architecture, and is amenable to development and operation by your team. Share:

When thinking about endpoint security it is important to decide what you consider an endpoint. We define an endpoint as any computing device that can access corporate data. This deliberately broad definition includes not just PCs, but also mobile devices (smartphones and tablets). We don’t think it is too broad – employees today expect to access the data they need, on the device they are using, from wherever they are, at any time. And regardless of the details, the data needs to be protected. Of course the buzzword du jour is Bring Your Own Device (BYOD), which means you need to support employee-owned devices, just as you support corporate-owned devices today. These folks go to the local big box retailer and come home with the shiny new iDevice or Android thingy, then show up the next working day expecting their email and access to the systems they need to do their job on the shiny new device. For a while you said no because you couldn’t enforce policies on that device, nor could you assume the employee’s children or friends wouldn’t get into email and check out the draft quarterly financials. Then you were summoned to the CIOs office and told about the new BYOD policy put in place by the CFO to move some of these expensive devices off the corporate balance sheet. At that point, ‘no’ was no longer an option, so welcome to the club of everyone who has to support BYOD – without putting corporate data at risk. The first step is to define the rules of engagement – which means policies. The reality is you probably have policies in place already, so it is a case of going back and revisiting them to ensure they reflect the differences in supporting both mobile devices and the fact thats you may not own said devices. This is a Buyer’s Guide and not a policy guide, so we won’t focus on specific policies, but we will point out that without an updated set of policies to determine what employees can and cannot do – covering both mobile devices and BYOD – you have no shot at controlling anything. BYOD First let’s blow up the misconception that BYOD = mobile devices. Employees may decide they want to run their office applications in a virtual window on their new Mac, not the 4-year-old Windows XP laptop they were assigned. Which means you need to support it, even though you don’t own the device. This changes how you need to provision and protect the device, particularly in terms of enforcement granularity. For devices you don’t own, you need the ability to selectively enforce policies. You cannot dictate what applications employees run on their own machines. You cannot whitelist the websites they visit. You cannot arbitrarily decide nuke a device from orbit if it shows indicators of possible malware. Actually, if your policy says so, you probably can legally control and wipe the device. But it would make you very unpopular if you decided to blow away a device and lost a bunch of personal pictures and videos in the process. So the key with BYOD is granularity. It is reasonable to do a periodic vulnerability scan on the device to ensure it’s patched effectively. It is also reasonable to require the device be encrypted so the corporate data on it is protected. It is fair to block access to corporate networks if the device isn’t configured properly or seems to be compromised. BYOD has several implications for security. Let’s examine the impact of BYOD in terms of the aspects we have discussed already: Anti-malware: If you require anti-malware on corporate owned computers, you probably want to require it on employee-owned machines as well. It also may be required by compliance mandates for devices which access protected information. The question is whether you require each employee to use the corporate standard anti-malware solution. If so, you would use your existing anti-malware solution’s enterprise management console. If not you need the capability to confirm whether anti-malware protection is running on each device on connection. You also need to decide whether you will mandate anti-malware protection for mobile devices, given the lack of malware attacks on most mobile platforms. Hygiene: Under our definition (patch management, configuration management, and device control), the key change for BYOD is reassessment of the security posture of employee-owned device on each connection to the network. Then it comes down to a policy decision on whether you allow insecurely configured or unpatched devices on the network, or you patch and update the device using enterprise management tools. Keep in mind there may be a software licensing cost to use enterprise tools on BYOD devices. The ability to deal with BYOD really comes down to adding another dimension to policy enforcement. You need to look at each policy and figure out whether it needs to change for employee-owned devices. It is also a good idea to make sure you can both visualize and report on employee-owned devices because there will be sensitivity around ensuring they comply with BYOD policies. Mobility We just explained why mobile devices are endpoints, so we need to provide guidance on protecting them. As with most newish technology, the worst initial problem is more than security. The good news is that mobile devices are inherently better protected from attack due to better underlying operating system architectures. That means makes hygiene – including patching, configuration, and determining which applications can and should run on the devices – the key security requirement. That doesn’t mean there is no mobile malware threat. Or that rooting devices, having employees jailbreak them, dealing with new technologies which extend the attack surface such as NFC (Near Field Communications), and attackers exploiting advanced device capabilities, aren’t all real issues. But none of these is currently the most pressing issue. That can and probably will change, as attackers get better and management issues are addressed. But for now we will focus on managing mobile devices. The technologies that enable us to manage mobile devices fall into a

This is all over the news, but Wired was the first I saw to put things in the right context: Four Russians and one Ukrainian have been charged with masterminding a massive hacking spree that was responsible for stealing more than 160 million bank card numbers from companies in the U.S. over a seven-year period. The alleged hackers were behind some of the most notorious breaches for which hacker Albert Gonzalez was convicted in 2010 and is currently serving multiple 20-year sentences simultaneously. The indictments clear up a years-long mystery about two hackers involved in those attacks who were known previously only as Grig and Annex and were listed in indictments against Gonzalez as working with him to breach several large U.S. businesses, but who have not been identified until now. The hackers continued their activities long after Gonzalez was convicted, however. According to the indictment, filed in New Jersey, their spree ran from 2005 to July 2012, penetrating the networks of several of the largest payment processing companies in the world, as well as national retail outlets and financial institutions in the U.S. and elsewhere, resulting in losses exceeding $300 million to the companies. And this tidbit: A second indictment filed in New York charges one of defendants with also breaching NASDAQ computers and affecting the trading system. This is a very big win for law enforcement. There aren’t many crews working at that level any more. It also shows the long memory of the law – most of the indictments are for crimes committed around five years ago. Share: