Posts

Avoiding the Security Pitfalls of SD-WAN and Network Modernization

Network modernization, like any wave of innovation, is multifaceted in its good intentions. It’s about rearchitecting your network so it is better able to handle increasing traffic and high-bandwidth-consuming apps such as video, ensure availability and quality of experience, flex for the delivery of new revenue-generating service offerings, and reduce network and application maintenance and overall costs.

The much ballyhooed yet still somewhat enigmatic cloud, with its highly virtualized and outsourced infrastructure, has already delivered some of this modernization by enabling organizations to offload some traffic from today’s predominantly hair-pinned and expensive MPLS-based WANs in favor of direct user access to Internet services. The cloud ecosystem offers other network modernization enablers such as shared service economies of scale, ready-to-leverage network capabilities such as automation, and transport independence (i.e., the ability to use broadband, LTE, Carrier Ethernet and MPLS “lines”).

Software-defined WANs (SD-WANs) could occupy a complementary network management and orchestration role to relieve some of the cost of (and dependence on) today’s rigid and expensive private networks. However, the path to network modernization is not all neatly wrapped and tied in pink ribbons, and uncertainty exists from a security perspective as well. Every time a user, whether stationed at one of your branch offices or remote, accesses the Internet directly he or she is potentially opening Pandora’s Box or letting sensitive data out. MPLS schemes require this sort of risky traffic to first pass through the core network for networking protocol and security application, which is a good thing, but at what cost? Traffic over MPLS lines can be dozens of times the Mbps/month cost versus broadband and the public Internet, so you want to orchestrate traffic in a way that reserves private lines for high-priority traffic and utilizes the public Internet for lower-priority interactions. Although SD-WAN may be ideal for this role and faster enablement of branch office and mobile workers through software-as-a-service, it is not an advanced security solution.

 

Advanced Security for SD-WAN and Cloud Networks

SD-WAN, which can empower organizations to exercise centralized SaaS control over traffic to and from the cloud and the WAN as a whole, poses some vulnerability issues. Centralized security is more difficult to administer when traffic isn’t backhauled to the data center or network hub, and malicious code and hacker schemes can more easily pass through to your distributed users undetected (north-south traffic).

What’s more, without the intervention of advanced security mechanisms, infections can more easily spread laterally – from user to user, system to system, and office to office (east-west traffic).

If you’re going to capitalize on the potential efficiencies of the cloud and SD-WAN controllers, you must first secure the egressing of traffic directly between the Internet and remote sites as well as protect against lateralization attacks. This can be accomplished through an advanced security solution designed for the cloud, which includes fully integrated next-generation firewall and endpoint protection as-a-service.

 

Secure Network Modernization Webinar

These and other topics will be explored during a webinar titled, “Avoiding the Security Pitfalls of SD-WAN and Network Modernization,” moderated by Security Now, and presented by Rik Turner, Principal Analyst, Ovum, and Ken Ammon, Chief Strategy Officer, OPAQ.

By attending this webcast, you will:

  • Understand the top security vulnerabilities plaguing companies as they modernize their networks
  • Learn how critical security vulnerabilities can be easily addressed with security-as-a-service
  • Discover how cloud and automation are enabling companies to simplify their ability to modernize their networks and security

Register for the webinar.

Download the white paper.

What is a Next-Gen Host-Based Firewall and why would anybody care?

Host-Based Firewalls are a simple technology that is generally used to prevent unwanted inbound traffic by port number. They don’t play a significant role in most enterprise security programs because its too much work to manage policies for each individual host. Instead, organizations prefer to enforce policies with network firewall devices that can protect large numbers of hosts from a single location.

New technologies have recently started to change this by providing a way to manage large numbers of individual endpoint firewall policies from a central system. We call these solutions Next-Generation Host-Based Firewalls, although the term Micro-Segmentation is also sometimes applied to this space. There are two primary trends that are driving this change:

Cloud Adoption: Workload mobility combined with the absence of traditional network architecture in cloud environments has meant that in some cases, firewall policies have to be managed on an individual endpoint basis, and there need to be tools that facilitate this.

Sophisticated Targeted Attacks: These days the initial point of infection for an attacker within a network is just a foothold that is used to spread internally in search of vital information to steal or encrypt with ransomware. This fact has driven organizations to pursue a Zero Trust approach to network security, where hosts inside the perimeter are not considered inherently more trustworthy than hosts outside the perimeter. The ultimate Zero Trust model means that every host is capable of defending itself, and tools are needed to orchestrate that defense.

As various Next-Generation Host-Based Firewall solutions have come on the market, the market has begun to define itself around a few key features or characteristics that all of these products share:

Central Policy Management: Obviously a “table stakes” requirement for these solutions is the ability to create a policy for a large number of individual endpoints from a central policy management tool. These policies are managed in one place, but enforced in many places — by each individual endpoint system.

Network Visualization:Crafting a security policy for large numbers of endpoints can be challenging. Next-Generation Host-Based Firewalls typically collect logs of network traffic from each endpoint and can provide the user with the ability to see and explore their network and it’s interrelationships. This can be a powerful tool for investigating security incidents as well as building policies that can contain them.

Abstract Policy Making: Traditional Firewalls enforce policy based on IPs, ports, and protocols. This can be inadequate for dealing with the complex set of interactions that occur on an internal network where workloads and workstations can move around. Typically, Next-Generation Host-Based Firewalls allow policies to be defined based on the identity of a user or of a workload or application, regardless of what system, IP or port is involved. This makes policy definition much simpler by allowing the user to express rules in human terms.

In our view there are three main architectural approaches to building Next-Generation Host-Based Firewalls. In describing these architectures, we use the words “active” and “passive” to refer to the role that the central policy management system takes in making case-by-case enforcement decisions.

Passive: A Passive Next-Generation Host-Based Firewall system is capable of pushing traditional, static firewall policies out to endpoints, but the central policy management system takes no direct role in policy enforcement. When new connections are made or received by each endpoint, the endpoint evaluates them against the policy it has been given and chooses whether to allow or block them.

This architecture has the advantage of imposing minimal latency at connection establishment and being resilient against temporary loss of connectivity between endpoints and the central policy manager.

Active: Instead of pushing static policies out to endpoints, an Active Next-Generation Host-Based Firewall makes enforcement decisions at a central controller. When each new connection is made or received by each endpoint, the endpoint contacts the controller and the controller decides whether or not the endpoint should allow or block, on a case-by-case basis. In this sense the controller is playing an active role in making enforcement decisions.

This architecture has the advantage of being able to adapt policy enforcement decisions immediately to changing circumstances on the network, such as when a host moves to a different network segment, or when a decision has been made to quarantine a compromised host. This adaptability is necessary to enable micro-segmentation in traditional office environments where rapid changes are commonplace. While there is some cost associated with this architecture, the latency and availability impacts are comparable to those imposed by the use of DNS servers.

Hybrid: A hybrid solution combines the best of both architectural approaches, allowing for dynamic policies to be enforced in real time by a central controller, with static backups in place that can make rapid decisions when the controller cannot be reached.

The OPAQ Cloud delivers Active & Hybrid Next-Generation Host-Based Firewalls. We believe that these technologies play a key role in securing the hybrid networks of today, especially as workloads move to the cloud and networks de-perimeterize. They enable enterprises to pursue a true Zero Trust approach to network security — where hosts on the internal network are not inherently trusted. The Zero Trust model is a prerequisite for defense against sophisticated threat actors, and a step toward totally new kinds of enterprise network architectures where perimeter defenses are no longer required.

Drawing a New Map of Enterprise Networking

Earlier this year I got to hear Tim O’Reilly speak at Grand Central Tech as part of their Authors @ GCT lecture series. Mr. O’Reilly is out promoting his new book, “WTF? What’s the Future and Why It’s Up To Us.” One of themes of his book is the process of innovation – how we go about creating technologies that completely change the way that we think, work, and live.

O’Reilly writes about drawing visual maps of the different elements within a company’s business plan, in order to understand how they interrelate with each other, a process that he learned about from a strategic consulting firm called BEAM. He then proceeds to draw such a map for an on-demand transportation company like Uber or Lyft.

There was a particular way that on-demand transportation worked a decade ago – you called a cab company, and a dispatcher announced your location on a radio network, and hopefully one of the cab drivers agreed to pick you up. Over time a particular set of technologies have become available, including the Internet, smart phones, and dispatching algorithms, that have enabled a completely different way of organizing this process. However, the new map for on-demand transportation didn’t draw itself – it was the job of innovators to realize that an opportunity existed to connect each of these ingredients in a new way, and to persuade the public that this new way is, in fact, a better way.

Of course, this got me thinking about what we’re doing at OPAQ Networks. IT organizations have been building enterprise networks in the same way ever since we started connecting businesses to the Internet in the early 1990’s. I usually credit Steven Bellovin and William Cheswick for drawing the original maps of this territory in their book “Firewalls and Internet Security.” This model is often called the “perimeter security model” – “We’ve got a bunch of sensitive computer systems here in our corporate headquarters, so we connected all of our satellite offices into that headquarters and we’ve built a stack of security solutions there to protect everything.”

Over time that model has started to show signs of strain. The sensitive systems that used to collect at headquarters are gone – they’ve moved into the cloud. However, the security stack is still there, and all kinds of traffic is still getting backhauled through headquarters for the sole purpose of sending it through the stack. Despite this approach, attackers are successfully getting inside by infecting end user workstations. Once their malware is running on the other side of the firewall, they have free range over the internal network and can get right to the data they want to steal.

At OPAQ Networks we are building a new map for this territory. First, we’re moving the security stack into the cloud, where the sensitive assets now live. This solves the backhaul problem, because satellite offices and remote VPN users can connect to cloud assets through our network instead of backhauling through a corporate headquarters. OPAQ has a nationwide network of points of presence and more than 200 peering relationships with major service providers that enable us to get traffic to it’s destination as efficiently and reliably as possible. Most small and medium sized enterprises don’t have the means to build this kind of infrastructure for themselves.

Second, we’re introducing software-defined network segmentation, a completely new technology that provides enterprises with unparalleled visibility and control over their internal networks. Using this tool, it’s possible to granularly segment internal networks so that end users only have access to the resources that they need, without having to reconfigure VLANs or wrestle with NAC solutions. Our partners’ midsize customers are able to adopt a better security posture, so that a single endpoint compromise does not imperil their entire business.

We are entering a time when the traditional way of building enterprise networks is being disrupted, and other maps are being drawn. Google’s BeyondCorp is one such map, along with the idea of Zero Trust Networks that was eloquently detailed in a recent O’Reilly publication. These approaches suggest doing away with the VPN and the security stack entirely, placing internal applications directly on the Internet and connecting users to them through authenticating proxy servers.

While I believe the BeyondCorp approach has merit, and there is a great deal that we can learn from it, it’s also very difficult for small and medium sized businesses to adopt. The traditional security stack delivered from the cloud has value, particularly for businesses where consistent patch and configuration management can be a challenge. The VPN has value, because it draws a clear line between the organization’s assets and the outside world. The problem is that these assets are often hosted in the wrong place today, and better segmentation is needed behind them.

This is what we’re doing at OPAQ Networks – we’re drawing a new map for the practice of enterprise networking in the cloud computing era. By leveraging network security-as-a-service, software-define network segmentation, and a modern, global network infrastructure, we’re enabling our customers to build networks that are more efficient, reliable, and secure than they have ever been before.

Simplified Microsegmentation — From the Cloud

It is time to change the way that organizations approach network segmentation. In the past few years we have seen a mounting collection of threats target the wide open nature of most organizations’ internal computer networks. Although security pros have been harping on this for some time, most networks remain crunchy on the outside and chewy in the middle – once attackers get past the perimeter, they often have access to any and everything inside the organization.

We’ve seen repeated threats recently exploit this exposure. We’ve seen incidents where entire organizations are crippled from ransomware spreading internally within their networks. We’ve seen the return of internet worms like WannaCry and NotPetya. We’ve seen more automated attacks that pivot from an initial point of compromise within a Windows network to Domain Admin access. In fact, experts are predicting significant increases in the volume of these attacks because of developments in attack automation.

Almost every organization needs to improve their network segmentation strategy in their internal network to cut down on these threats. What is preventing organizations from taking action?

Traditional Network Segmentation is Complex and Difficult to Manage

Unfortunately, the traditional approach to implementing network segmentation poses significant challenges. Configuring and managing internal firewalls and VLANs is both labor intensive and relatively inflexible. Network architecture is usually driven by the need to provide connectivity rather than security. Organizing machines with different security requirements onto separate VLANs is complex, and as soon as the work is done, users demand changes. Deploying multi-factor authentication for internal applications and services can also be a daunting project as each application must be separately integrated.

It’s no wonder organizations — particularly midsize enterprises — continue to struggle with implementing a smart, sustainable network segmentation strategy. What are midsize enterprises — and the service providers supporting them — supposed to do?

Zero Trust Software-Defined Network Segmentation from the Cloud

The term “microsegmentation” has recently become a buzzword in the IT world. These solutions provide a manageable way to lock down east/west traffic policies for cloud workloads. However, many of the threats we’re seeing – ransomware, worms, and domain lateralization – target end user workstations instead. What organizations need is a technology that provides easy-to-deploy software-defined microsegmentation capability that is flexible enough to support the entire enterprise network.

Since the acquisition of Drawbridge Networks in May 2017, we have embarked on integrating unique intellectual property into the OPAQ Cloud that allows users to manage software-defined microsegmentation for the entire enterprise, from a single pane of glass. The OPAQ PathProtect™ capability dramatically simplifies network segmentation, enhances network visibility and control, and enforces policy locally at each device, whether it’s a cloud workload or an employee laptop.

OPAQ PathProtect™ works by connecting software agents running on endpoints with a central controller hosted in the OPAQ Cloud. This architecture provides visibility and control from the cloud into every network interaction happening on every endpoint. This capability gives you the power to investigate incidents, protect against insider and external attacks, and prevent certain devices, such as compromised endpoints, from talking to other workstations on the network.

Microsegmentation with OPAQ PathProtect™ can be used to define granular access segments for users that operate independently from the network’s hardware and physical topology. It also can be easily updated when business needs change. Segments can be defined based on user identity, group membership and job function, and they will follow users as their laptops move throughout the network. OPAQ PathProtect™ can be used to enforce multi-factor authentication for access to any resource or service on the network, without any need to integrate with individual applications. This is possible because the central controller oversees all communication within the network and can authenticate users before allowing traffic to flow.

These capabilities allow organizations to adopt a security posture that is more aligned with Zero Trust security principles, in which users only have access to the specific applications required by their job function. Cutting down on unnecessary access closes the avenues that malware and network attackers use to spread laterally within an organization.

Microsegmentation for Endpoints, Not Just Data Centers

OPAQ PathProtect™ is a microsegmentation solution that can protect the whole network, including workstations, servers, datacenters, and cloud workloads, supporting the following capabilities and use cases:

  • Network Visibility provides detailed topological views of the interactions between hosts on the internal network. It is possible to drill down into different timeframes, hosts, users, process names, ports, and protocols for complete insight into network activity.
  • Network Access Control (NAC) to assign which resources, hosts and users can access services on the network. For example, unmanaged hosts can be prevented from accessing sensitive servers, and are identified and cataloged when they send traffic.
  • Multi -Factor Authentication (MFA) integration enables step-up authentication to tighten security for VPN access and within the internal network.
  • Granular Segmentation which is completely separate from the physical network architecture or network addressing, can be used to segment specific devices, applications, and data, and can keep track of hosts as they move around the network.
  • Quarantine allows organizations to quickly isolate infected hosts from sensitive resources at the touch of a button.

To find out more, view the press announcement, sign up for our upcoming webcast and schedule a demo to see how simple microsegmentation can be from the cloud.

OPAQ CTO Tom Cross Writes on Lateralization Attacks in First Article on CSO Online

Lateralization attacks are commonly used in most sophisticated breaches today. An adversary will typically gain a foothold inside the victim’s network by installing malware on a vulnerable device.

From there, the attacker will compromise other computers within the organization by moving laterally throughout the compromised network. A number of experts are predicting an increase this year in Windows Domain lateralization attacks. Organizations are increasingly looking for a solution that can prevent and isolate lateralization attacks from spreading in their network.

OPAQ chief technology officer Tom Cross was recently invited to be a regular contributor to CSOonline, one of our industry’s most respected publications. In Tom’s first article, he discusses lateralization attacks against Windows networks, and how to defend against them. You can read the full article here.