Best Practices for DDoS Protection and Mitigation  on Google Cloud Platform  Last updated: April 12th, 2016 Introduction Protecting Shared Protecting Google’s Infrastructure DDoS Protection and Mitigation for your GCP Deployment Conclusion

Introduction A Denial of Service (DoS) attack is an attempt to render your service or application unavailable to your end users. With Distributed Denial of Service (DDoS) attacks, the attackers use multiple resources (often a large number of compromised hosts/instances) to orchestrate large scale attacks against targets. This document describes the best practices for protecting against and mitigating such DDoS attacks for your Google Cloud Platform (GCP) deployment.

Protecting Shared Infrastructure Google has mechanisms in place to protect its cloud infrastructure and its production services. These mechanisms are designed to ensure that no single service can overwhelm the shared infrastructure and to provide isolation among customers using the shared infrastructure. Details of these mechanisms are out of scope for this document.

DDoS Protection and Mitigation for your GCP Deployment Successfully thwarting and handling DDoS attacks for your GCP deployment is a shared responsibility between Google Cloud Platform and you. ​ DDoS defense involves deploying detection systems, implementing barriers and being able to absorb attacks by scaling in order to prevent attackers from overwhelming or disabling access to your services or applications. Google Cloud Platform provides several of these mechanisms automatically and you can follow the best practices detailed below on your end to help secure your GCP deployment:

● Reduce the attack surface for your GCE deployment ○ Provision your own isolated and secure piece of the Google Cloud with Google Cloud Virtual Network​ . View the best practice ​ here​ .



○ Isolate and secure your deployment using subnetworks and networks, firewall rules, tags and Identity and Access Management (IAM). ○ Open access to ports and protocols that you need using firewall rules and/or ​ protocol forwarding​ . ○ GCP provides anti-spoofing protection for the private network (IP addresses) by default. ○ GCP automatically provides isolation between virtual networks.

● Isolate your internal traffic from the external world ○ Deploy instances without public IPs unless necessary. ○ You can set up a NAT gateway or SSH bastion to limit the number of instances that are exposed to the internet. ○ Once available, deploy Internal Load Balancing for your internal client instances accessing internally deployed services thereby avoiding exposure to the external world. [Internal LB expected to be available in the second half of 2016.]

● DDoS Protection by enabling Proxy-based Load Balancing ○ When you enable HTTP(S) Load Balancing or SSL proxy Load Balancing, Google infrastructure mitigates and absorbs many Layer 4 and below attacks, such as SYN floods, IP fragment floods, port exhaustion, etc. ○ If you have HTTP(S) Load Balancing with instances in multiple regions, you are able to disperse your attack across instances around the globe.

● Scale to absorb the attack ○ Protection by Google Frontend infrastructure With Google Cloud Global Load Balancing, the frontend infrastructure which terminates user traffic, automatically scales to absorb certain types of attacks (e.g., SYN floods) before they reach your compute instances. ○ Anycast-based Load Balancing​ : HTTP(S) Load Balancing and SSL proxy enable a single anycast IP to front-end your deployed backend instances in all regions. Normally your user traffic is directed to the closest backend with capacity; in the event of a DDoS attack, the additional advantage of this approach is that it increases the surface area to absorb this attack by moving traffic to instances with available capacity in any region where backends are deployed. ○ Autoscaling​ : When you configure HTTP(S)or SSL Proxy Load Balancing, Google frontend infrastructure that terminates your user traffic protects your backends. You should also provision sufficient number of instances  


and/or configure autoscaling to handle spikes in traffic. In the event of a sudden traffic spike, the load balancing proxy layer will distribute the traffic across all the backends with available capacity. In parallel, the autoscaler ramps up the backends inline with traffic that needs to be handled.

● Protection with CDN Offloading ○ Google Cloud CDN acts as a proxy between your clients and your origin servers. For cacheable content, Cloud CDN caches and services this content from points-of-presence (POPs) closer to your users as opposed to sending them to backend servers (instances). In the event of DDoS attack for cacheable content, the requests are sent to POPs all over the globe as opposed to your origin servers, thereby providing a larger set of locations to absorb the attack. ○ If you use CDN Interconnect, you can leverage the additional DDoS Protection provided by our CDN Interconnect partners. You can check the partner page​ for specifics on their DDoS protection capabilities.

● Deploy third-party DDoS protection solutions ○ In order to meet your specific needs of protection for DDoS attack prevention/mitigation, consider purchasing specialized third-party DDoS protection solutions to protect against such attacks. ○ You can also deploy DDoS solutions available via ​ Google Cloud Launcher​ .

● App Engine deployment ○ App Engine is designed to be a fully multi-tenant system and implements a number of safeguards intended to ensure that a single bad application will not impact the performance or availability of other applications on the platform. ○ App Engine sits behind the Google Front End which mitigates and absorbs many Layer 4 and below attacks, such as SYN floods, IP fragment floods, port exhaustion, etc. ○ You can also specify a set of IPs/IP networks via a ​ dos.yaml file to block them from accessing your application(s).

● Google Cloud Storage ○ If you do not want to require your users to have a Google account in order to be able to access your Google Cloud Storage resources, you can control access using ​ signed URLs​ .



● API rate-limiting ○ API rate limits define the number of requests that can be made to the Google Compute Engine API​ . API rate limits apply on a per-project basis. ○ Currently, projects are ​ limited to an API rate limit​ of 20 requests/second.

● Resource Quotas ○ Compute Engine enforces ​ quotas on resource usage for a variety of reasons. For example, quotas protect the community of Google Cloud Platform users by preventing unforeseen spikes in usage. ​ Special quotas limit access for projects that are just exploring Google Cloud Platform on a free trial basis.

Conclusion Google Cloud Platform provides a number of features to defend against DDoS attacks. You can use these in conjunction with the above mentioned best practices and other measures tailored to your requirements to make your GCP deployment resilient to DDoS attacks.



Best Practices for DDoS Protection and Mitigation on Google Cloud ...

Apr 12, 2016 - A Denial of Service (DoS) attack is an attempt to render your service or ... Google Cloud Virtual Network​. View the best practice ​here​. 1 ...

197KB Sizes 1 Downloads 611 Views

Recommend Documents

Best Practices Award Template Cloud
Yet as the cloud matures, the market is splitting into two segments, with different expectations. The first segment comprises enterprise IT leaders who are looking beyond outsourced infrastructure; they now consider the cloud as the foundation for a

Best Practices Award Template Cloud
Scaling applications generally requires the customer's manual intervention. Usage-based pricing is based on the customer's expected capacity use, which requires the customer to be adept at capacity planning to obtain the best rates. As enterprises in

Best Practices Award Template Cloud
piled high with value-added services to handle their production workloads, the developers ... their virtual machines; far less than the productivity achieved by Google's own engineers. This realization spurred the .... Our awards team follows a 10-st

Best Practices Award Template Cloud Platform
service providers, including Microsoft, VMware, IBM, and RedHat, have joined the Google- led Kubernetes community to contribute code. In launching Container Engine and ... competitors support startups primarily through their standard offer set, inclu

Best Practices Award Template Cloud Platform
Best Practices Recognition: 10 Steps to Researching, Identifying, and Recognizing Best ... infrastructure and software than perhaps any other organization in the world. ... Google Cloud Platform's strategy also reflects the company's culture of ...

Best Practices for OTT Dynamic Ad Insertion Cloud
providers can deliver adverts that are more relevant to the viewer, OTT is more impactful as an overall experience, and will ... With the streaming format wars in the rear view mirror, HAS, specifically Apple's HTTP Live Streaming. (HLS) and the ...

Best Practices for OTT Dynamic Ad Insertion Cloud Platform
1. Abstract. 2. Fundamentals of Ad Stitching. 3. CSAI versus SSAI: Stitch ... reality for OTT services using a variety of data sources (GPS, postal code, IP tables).

Interference Mitigation and Capacity Enhancement based on ...
Interference Mitigation and Capacity Enhancement ba ... Dynamic Frequency Reuse for Femtocell Networks.pdf. Interference Mitigation and Capacity ...

Survey of DDOS Attack in Cloud Survey of DDOS ...
Platform as a Service (PaaS) is a delivery of a computing platform over the web. ... or Software as a Service (SaaS) refers to software delivered over a browser. ... One of the primary reasons the hybrid model is popular is that organizations.

SaaS Security Best Practices: Minimizing Risk in the Cloud - Media15
Aug 7, 2015 - OTP one-time password. SaaS software as a service. SBI security business intelligence. SCIM system for cross-domain identity management.

SaaS Security Best Practices: Minimizing Risk in the Cloud - Media15
Aug 7, 2015 - To minimize risk in the cloud, we have established the following best ..... are accessed by the appropriate users in the appropriate computing.

Deficit Mitigation, Juvenile Justice Reform and Voter Registration on ...
Mar 24, 2016 - Deficit Mitigation, Juvenile Justice Reform and Voter Registration ... follow the network at or visit CT-N's website.

Best Practices for Nonprofits Services
Every Google+ page begins with a profile. This profile is usually the administrator, the person most responsible for the content of your Google+ page. If you don't already have a profile, you'll need to create one here. Next, you can set up your orga

Networking Best Practices for Large Deployments
Nov 16, 2016 - rights relating to the Google services are and shall remain the exclusive ..... 1. Network Evaluation: This section contains information on ...

IBM-Best practices for KVM.pdf
Loading… Page 1. Whoops! There was a problem loading more pages. Retrying... IBM-Best practices for KVM.pdf. IBM-Best practices for KVM.pdf. Open. Extract.

Best Practices for RXSC ICO.pdf
Blend, Thin Blend, Beautiful Mind, Heart Smart and Fresh Breeze. Best Practices for RXSC ICO (10/06/17). This report is to inform and protect buyers, and. increase the chances of a successful token sale. Source: