Tuesday, December 31, 2013

Monday, December 23, 2013

Cisco Nexus in Action


Cisco Nexus 1000V Layer 2 Switching Configuration Guide, Release 4.0(4)SV1(1)
Overview

Table Of Contents


Overview


The Cisco Nexus 1000V Layer 2 Switching Configuration Guide, Release 4.0(4)SV1(1) provides an overview of the available Layer 2 features and how to configure them.
This chapter includes the following sections:
VLANs

Information about the Cisco Nexus 1000V

This section includes the following topics:

VEM Port Model

The Cisco Nexus 1000V differentiates the following Virtual Ethernet Module (VEM) ports:
Figure 1-1 shows how VEM ports are bound to physical and virtual VMware ports.
Figure 1-1 VEM Port View

VEM Virtual Ports

The virtual side of the VEM maps together the following three layers of ports:

Virtual NICs

There are three types of Virtual NICs in VMware. The virtual NIC (vnic) is part of the VM, and represents the physical port of the host which is plugged into the switch. The virtual kernel NIC (vmknic) is used by the hypervisor for management, VMotion, iSCSI, NFS and other network access needed by the kernel. This interface would carry the IP address of the hypervisor itself, and is also bound to a virtual Ethernet port. The vswif (not shown) appears only in COS-based systems, and is used as the VMware management port. Each of these types maps to a veth port within Nexus1000V.

Virtual Ethernet Ports

A virtual Ethernet port (vEth) represents a port on the Cisco Nexus 1000V Distributed Virtual Switch. Cisco Nexus 1000V has a flat space of vEth ports, 0...n. These vEth ports are what the virtual "cable" plugs into, and are moved to the host that the VM is running on.
Virtual Ethernet ports are assigned to port groups.

Local Virtual Ethernet Ports

Each host has a number of local vEth (lvEth) ports. These ports are dynamically selected for vEth ports needed on the host.
Local vEths do not move, and are addressable by the convention, module/port number.

VEM Physical Ports

The physical side of the VEM includes the following from top to bottom:

VMware NIC

Each physical NIC in VMware is represented by an interface called a VMNIC. The VMNIC number is allocated during VMware installation, or when a new physical NIC is installed, and remains the same for the life of the host.

Uplink Ports

Each uplink port on the host represents a physical interface. It acts a lot like an lvEth port, but since physical ports do not move between hosts, the mapping is 1:1 between an uplink port and a VMNIC.

Ethernet Ports

Each physical port added to Cisco Nexus 1000V appears as a physical Ethernet port, just as it would on a hardware-based switch.

Note The uplink ports are handled entirely by VMware, and are used to associate port configuration with VMNICs. There is no fixed relationship between the uplink number and VMNIC number, and these can be different on different hosts, and can change throughout the life of the host. On the VSM, the ethernet interface number, for example, ethernet 2/4, is derived from the VMNIC number, not the uplink number.

VSM Port Model

Figure 1-2 shows the VSM view of the network.
Figure 1-2 VSM View
The Virtual Supervisor Module (VEM) has the following ports or interfaces:

Virtual Ethernet Interfaces

Virtual Ethernet interfaces (vEths) can be associated with any of the following:
A virtual machine VNIC on the ESX host
A virtual machine kernel NIC on the ESX host
A virtual switch interface on an ESX COS host

Physical Ethernet Interfaces

Physical Ethernet interfaces (Eths) correspond to the physical NICs on the ESX host.

Port Channel Interfaces

The physical NICs of an ESX host can be bundled into a logical interface called a port channel interface.

Switching Architecture in Cisco Nexus 1000V

Each VEM attached to the VSM forwards traffic to and from the ESX server as an independent and intelligent line card. Each VLAN uses its forwarding table to learn and store MAC addresses for ports connected to the VEM.
Figure 1-3 shows the traffic flow between two VMs on different VEMs.
Figure 1-3 Traffic Flow Between VEMs

Layer 2 Ethernet Switching

The congestion related to high bandwidth and large numbers of users can be solved by assigning each device (for example, a server) to its own 10-, 100-, 1000-Mbps, or 10-Gigabit collision domain. Because each LAN port connects to a separate Ethernet collision domain, servers in a switched environment realize full bandwidth access.
Full duplex allows two stations to transmit and receive at the same time. This is unlike 10/100-Mbps Ethernet, which usually operates in half-duplex mode, so that stations can either receive or transmit but not both. When packets can flow in both directions simultaneously, the effective Ethernet bandwidth doubles. 1/10-Gigabit Ethernet operates in full-duplex only.
Each LAN port can connect to a single workstation or server or to another device through which workstations or servers connect to the network.
To reduce signal degradation, each LAN port is considered to be an individual segment. When stations connected to different LAN ports need to communicate, frames are forwarded from one LAN port to the other at wire speed to ensure full bandwidth for each session.

MAC Address Tables

To switch frames between LAN ports efficiently, a MAC addres table is maintained. The MAC address of the sending network is associated with the LAN port on which it was received. For more information about MAC address tables, see Chapter 2, "MAC Address Table."

VLANs

A VLAN is a switched network that is logically segmented by function, project team, or application, without regard to the physical locations of the users. VLANs have the same attributes of physical LANs, but you can group end stations even if they are not physically located on the same LAN segment.
Any switchport can belong to a VLAN, and unicast, broadcast, and multicast packets are forwarded and flooded only to end stations in that VLAN. Each VLAN is considered a logical network, and packets destined for stations that do not belong to the VLAN must be forwarded through a bridge or a router.
All ports, including the management port, are assigned to the default VLAN (VLAN1) when the device first comes up.
The devices support 4094 VLANs in accordance with the IEEE 802.1Q standard. These VLANs are organized into several ranges for different uses. Some of these VLANs are reserved for internal use by the device and are not available for configuration

Note Inter-Switch Link (ISL) trunking is not supported by the Cisco Nexus 1000V.

See Chapter 3, "VLAN Configuration" for complete information on configuring VLANs.

Private VLANs

Private VLANs (PVLANs) are used to segregate Layer 2 ISP traffic and convey it to a single router interface. PVLANs achieve device isolation by applying Layer 2 forwarding constraints that allow end devices to share the same IP subnet while being Layer 2 isolated. In turn, the use of larger subnets reduces address management overhead.

IGMP Snooping

The Internet Group Management Protocol (IGMP) snooping software examines Layer 2 IP multicast traffic within a VLAN to discover the ports where interested receivers reside. Using the port information, IGMP snooping can reduce bandwidth consumption in a multi-access LAN environment to avoid flooding the entire VLAN. The IGMP snooping feature tracks which ports are attached to multicast-capable routers to help the routers forward IGMP membership reports. The IGMP snooping software responds to topology change notifications. By default, IGMP snooping is enabled on the device. For more information, see Chapter 5, "IGMP Snooping Configuration."

Related Topics

The following documents contain related information:
Cisco Nexus 1000V Interface Configuration Guide, Release 4.0(4)SV1(1)
Cisco Nexus 1000V Port Profile Configuration Guide, Release 4.0(4)SV1(1)
Cisco Nexus 1000V Security Configuration Guide, Release 4.0(4)SV1(1)
Cisco Nexus 1000V System Management Configuration Guide, Release 4.0(4)SV1(1)

Useful drawing tools

  • https://www.draw.io
  • do not forget to listen to online music here http://www.dhingana.com/

Crowdfunding

A new buzz in tech industry in 2013 where an individual or a company can raise money online from social media contacts.

Funding needs just credibility and convincing idea ( may be creative )
Lately i hear scams and frauds in these where money floating across haywire...

REIT : Real Estate Investment Trust

Some useful links
  • http://www.youtube.com/watch?v=-38uPkyH9vI
  • http://www.youtube.com/watch?v=r7AxwR9mf-A
  • http://www.indiegogo.com/learn-how-to-raise-money-for-a-campaign
  • http://www.kickstarter.com/
  • http://www.ibtimes.com/crowdfunding-2014-forecasts-equity-crowdfunding-takes-1518870
  • https://www.realtymogul.com/

Wednesday, December 18, 2013

Cannot miss this blog

http://kendrickcoleman.com/index.php/Tech-Blog/standing-up-the-cisco-nexus-1000v-in-less-than-10-minutes.html


Monday, December 16, 2013

CiscoNexus and VMWare

I am trying to remove the VIB of VEM from ESXi

And this is how to do it.
  1. Enable SSH in your ESxi, login to ESXi Shell
  2. #vem stop
  3. #esxcli software vib list | grep cisco
  4. #esxcli software vib remove --vibname=cisco-vem-v162-esx --maintenance-mode
    Removal Result
       Message: Operation finished successfully.
       Reboot Required: false
       VIBs Installed:
       VIBs Removed: Cisco_bootbank_cisco-vem-v162-esx_4.2.1.2.2.1a.0-3.2.1
       VIBs Skipped:

With the 1000v, the intention is you're SPAN/ERSPAN traffic to your own sniffing device.  Either another switch (Cat6500), Wireshark Host/VM or other traffic capturing device.

The SPAN is mirroring source traffic on a single VEM host.  SPAN does not utilize the uplinks for traffic in this regard.  The amount of extra traffic in a SPAN/ERSPAN session is dictated by the source (vEth, Eth, Port Channel, Port Profile, entire VLAN etc.).  The destination for a SPAN session can be another vEthernet, Ethernet or Port Channel interface on that host.  Most customers setup a Wireshark VM, migrate it to whichever VEM host they want to run a SPAN on, and then just set their SPAN destination to the vEth port of the Wireshark VM.  Simple and free way to capture traffic.

Tuesday, December 10, 2013

Tech Entrepreneur

  • What is this?
  • I see lots of these buzz words lately more in different scale, how it is different from the dotcom days of bubble.
  • What is funding and who get the funding and how they fund?
  • B-School / Top tech school?
  • http://www.gv.com/lib/product-manager-meetup-san-francisco
  • http://thenextweb.com/insider/2013/10/12/product-managers-mini-ceos/

Tuesday, November 5, 2013

Startup watch

http://www.medianama.com/2012/10/223-gsf-accelerators-first-15-startups-zoomdeck-tushky-playcez-eduflix-more/

http://cubito.in/

Cloud Based IPTV

Oh!! man , IP based TVs are not a future technology since 10 years, but more and more network centric solutions amaze me.
When netflix was down a month back this year amazon web services were blamed and a glitch in their architecture caused the glitch in netflix since amazon EC2 and S3 are used to stream netflix (http://www.informationweek.com/cloud-computing/infrastructure/cloud-connect-netflix-finds-home-in-amaz/229300547)

There are more cloud based
  • Cisco videoscape
  • ActiveVideo H5 
  • Telus Optik TV

http://www.youtube.com/watch?v=X8NOrnL3-4Q
http://newsroom.cisco.com/press-release-content?type=webcontent&articleId=1280717

New Era of Whitebox Virtual Switches

whitebox hardware companies are the once who are building the hardware to run enterprise level network switch software.

Some good examples are
http://cumulusnetworks.com
  •  Interestingly most of the founders are ex-cisco
http://www.pica8.com/
  • This should be chinese version of this solution
 Arista's 7x switch
  • This week arista jumped and announced their new product 7000 series of switch ( note : cloud switch )
  • Single tier spline : in the rush to use the terminology before cisco insieme release
  • The Arista 7300 is comprised of three chassis: the Arista 7304, 7308 and 7316 with 4, 8 and 16 line card slots respectively. All three share a common resilient architecture that scales up to 512 ports of 40GbE or 2,048 ports of 10GbE, with wirespeed performance of 40Tbps (terabits per second) of throughput. X Series line card modules for the Arista 7300 include 10GBASE-T, SFP and QSFP configurations. Front-to-rear and rear-to-front airflow options along with platinum rated power supplies allow for improved efficiency and middle-of-row Spline configurations.
    Two Arista 7316 Series systems can fit in a single 42RU rack supporting over 4000 10GbE ports. With power consumption under 3W per 10GbE port and latency under 2 microseconds, a pair of 7300 series switches replaces two Catalyst 6509Es with more than ten times the scale, throughput, latency improvement and power efficiency.
    Complementing the 7300X series, the Arista 7250X Series is a high density solution delivering 64 ports of wire-speed 40GbE or up to 256 Ports of 10GbE in a compact and power efficient two rack unit fixed form factor with redundant and hot-swap power supplies and fan modules. Key features on the 7300X and 7250X Series:
    -- Unified Forwarding Table: The Layer 2 MAC table and Layer 3 routing tables dynamically expand based on the application to support up to 288K MAC entries, or 144K routes.
    -- New Duplex Fiber 40GbE Optics: Support for Arista's LRL4 QSFP optics requiring a single pair of single mode fiber, reducing the fiber requirements by 75% compared to existing choices.
    -- Network Telemetry: Real-time visibility into network congestion, buffer management and industry-standard sFlow support for traffic analysis.
    -- Physical-Virtual-Cloud Network: Arista VMTracer provisioning and native VXLAN support, workload mobility and migration at wire speed across multiple hypervisors.
    Pricing and Availability
    The Arista 7300 and 7250 Series are key components of the Arista Software Defined Cloud Network. The 7250QX-64 is available now and shipping for $1500 per 40GbE port. The 7300X series are available in Q1 2014 from $500 per 10GbE port. For more information, please visit http://www.aristanetworks.com/en/products/

Wednesday, October 30, 2013

Cisco Insieme - Nexus 9000 - Nexus 9K - N9K

I happened to read this post about CISCO Nexus 9K product a high speed high density switch at 40G http://www.networkworld.com/community/blog/cisco-nexus-9000-aimed-40g

The best part i liked is the comment section, scroll all the way down to the page and enjoy.

"Interesting considering the paint isnt even dry on the 7700. Given the 7700 launch, one might think they could have waited another year for the custom asics to be ready. I guess we will know for sure soon enough"

Monday, October 28, 2013

NFV (Network Function Virtualization)








I came across few commercial solutions to justify the buzz word NFV, where a merchant silicon function of FPGA,ASICs are moved on to the software and a very good example is in below video where Mobile Core Packet Network (EPC - Evolved Packet Core) is virtualized as software

http://www.intel.com/content/www/us/en/events/corporate/mobile-world-congress-2013/mwc-evolved-packet-core.html


http://www.sdncentral.com/companies/6wind-nfv-interview-2013/2013/08/

Some other network function virtualization i can recognize are
  • Any critical data plane traffic could be a good candidate for NFV
  • DPI ( Deep Packet Inspection) 
  • Switching (OVS is now part of linux)
    • See 6WINDGate ; Maximizing the switching performance of the virtual switch to provide high-bandwidth to the VNF running in VMs thereby increasing the aggregated bandwidth delivered to the VMs. Second, it accelerates the data plane performance of each VNF
  • A good use case is : In any traditional network each hardware does a separate function in a network, in the newer NFV based network multi-tenant solutions solved CAPEX and OPEX 
Another good example of open source project NFV-IMS (i.e)Virtualize the IP Multimedia subsystem network components see http://www.projectclearwater.org/

http://www.sdncentral.com/news/nfv-comes-life-sdn-openflow-world-congress/2013/10/

Friday, October 25, 2013

Point of sale - Solutions

Came across http://www.lightspeedretail.com/compare/

Looks cool

List of OpenFlow Software Projects

(I am trying to keep a running list of all OpenFlow-related software projects where either the bits or the source are available online. If you know of one that I'm missing, please e-mail me and I'll include it)
Switch Software and Stand-Alone OpenFlow Stacks

Open vSwitch: (C/Python) Open vSwitch is a an OpenFlow stack that is used both as a vswitch in virtualized environments and has been ported to multiple hardware platforms. It is now part of the Linux kernel (as of 3.3).
OpenFlow Reference: (C) The OpenFlow reference implementation is a minimal OpenFlow stack that tracks the spec.
Pica8: (C) An open switch software platform for hardware switching chips that includes an L2/L3 stack and support for OpenFlow.
Indigo: (C) Indigo is a for-hardware-switching OpenFlow implementation based on the Stanford reference implementation.
Pantou: (C) Pontou is an OpenFlow port to the OpenWRT wireless environment.
OpenFaucet: (Python) OpenFaucet is a pure Python implementation of the OpenFlow 1.0.0 protocol, based on Twisted. OpenFaucet can be used to implement both switches and controllers in Python.
OpenFlowJ: (Java) OpenFlow stack written in Java.
Oflib-node: (Javascript) Oflib-node is an OpenFlow protocol library for Node. It converts between OpenFlow wire protocol messages and Javascript objects.
Nettle: (Haskell) OpenFlow library written in Haskell.
Controller Platforms

POX: (Python) Pox as a general SDN controller that supports OpenFlow. It has a high-level SDN API including a queriable topology graph and support for virtualization.
MUL: (C) MÅ«L, is an openflow (SDN) controller. It has a C based muli-threaded infrastructure at its core. It supports a multi-level north bound interface for hooking up applications. It is designed for performance and reliability which is the need of the hour for deployment in mission-critical networks.
NOX: (C++/Python) NOX was the first OpenFlow controller.
Jaxon: (Java) Jaxon is a NOX-dependent Java-based OpenFlow Controller.
Trema: (C/Ruby) Trema is a full-stack framework for developing OpenFlow controllers in Ruby and C.
Beacon: (Java) Beacon is a Java-based controller that supports both event-based and threaded operation.
Floodlight: (Java) The Floodlight controller is Java-based OpenFlow Controller. It was forked from the Beacon controller, originally developed by David Erickson at Stanford.
Maestro: (Java) Maestro is an OpenFlow "operating system" for orchestrating network control applications.
NDDI - OESS: OESS is an application to configure and control OpenFlow Enabled switches through a very simple and user friendly User Interface.
Ryu: (Python) Ryu is an open-sourced Network Operating System (NOS) that supports OpenFlow.
NodeFlow (JavaScript) NodeFlow is an OpenFlow controller written in pure JavaScript for Node.JS.
ovs-controller (C) Trivial reference controller packaged with Open vSwitch.
Special Purpose Controllers

RouteFlow RouteFlow, is an open source project to provide virtualized IP routing services over OpenFlow enabled hardware. RouteFlow is composed by an OpenFlow Controller application, an independent RouteFlow Server, and a virtual network environment that reproduces the connectivity of a physical infrastructure and runs IP routing engines (e.g. Quagga).
Flowvisor (Java) FlowVisor is a special purpose OpenFlow controller that acts as a transparent proxy between OpenFlow switches and multiple OpenFlow controllers.
SNAC (C++) SNAC is an OpenFlow controller builton NOX, which uses a web-based policy manager to manage the network.
Resonance Resonance is a Network Access Control application built using NOX and OpenFlow.
Oflops (C) OFlops (OpenFLow Operations Per Second) is a standalone controller that benchmarks various aspects of an OpenFlow switch.
Misc

STS SDN Troubleshooting Simulator
FlowScale FlowScale is a project to divide and distribute traffic over multiple physical switch ports. FlowScale replicates the functionality in load balancing appliances but using a Top of Rack (ToR) switch to distribute traffic.
NICE-OF NICE is a tool to test OpenFlow controller application for the NOX controller platform.
OFTest OFTest is a Python based OpenFlow switch test framework and collection of test cases. It is based on unittest which is included in the standard Python distribution.
Mirage Mirage is an exokernel for constructing secure, high-performance network applications across a variety of cloud computing and mobile platforms. Apparently, it supports OpenFlow.
Wakame VDC (Ruby) IaaS platform that uses OpenFlow for the networking portion.
ENVI ENVI is a GUI framework that was designed as an extensible platform which can provide the foundation of many interesting OpenFlow-related networking visualizations.
NS3 (C++/Python) NS3 is a network simulator. It has openflow support built in to emulate an openflow environment and also it can be used for real-time simulations.

Wednesday, October 23, 2013

Network Visibility Controller



Virtualized network traffic Visibility
https://www.youtube.com/watch?v=pzSU7feB-I4

  • ·         Traffic on the physical host can be tapped from a physical port on the host or by spanning

  • ·         Netoptics phanthom vTap: monitoring at kernel level instead of SPAN or port mirror traffic from the port.







Saturday, October 19, 2013

FPGA vs Network Processor

http://www.nallatech.com/
http://www.aliathon.com/

vFabric VMWare

VMWare have a vFabric appinsignt product for network monitoring used for performance management
  • A probeVM  has to be installed into each host (ESXi) where the application component is installed
  • These probes send DPI information on network traffic and sends the information to AppInsight server
  • Login into app insight as admin and install probe into the host
  • Choose the proper switch and switch to monitor traffic

SOAP vs REST

My take on REST vs SOAP

REST
  • REST can be consumed by any client, web browser with ajax and javascript
  • REST is lightweight
    • No XML parsing required
    • Has less bandwidth, SOAP headers are heavy payload
  • HTTP GET operations are safe in REST
  • REST are mostly stateless 
  • Good for web services, light weight no xml parsing systems
  • Exposes data over internet
 SOAP

  • SOAP is more like a RPC, where REST is not
  • SOAP / WSDL  can be used to generate client side code unlike REST where HTTP code has to be developed
  • SOAP can be asynchronous in nature, REST are http based and need asynchronous implementation
  • HTTP/HTTPS layer  need not be developed like REST
  • REST have no contract - prone to go wrong in integration of systems
  • SOAP have contract - can be discovered easily
  • Good for enterprise services - system integrations
  • Asynchronous processing
  • Stateful with contract to systems

SOA - An Architect View

SOA [ Service Oriented Architecture ]
"SOA = Loose Coupling"

Trying to pen down my important markings of SOA and how to approach.
  1. Architects  mostly uses an existing product(s) to build solution to address an specific problem
  2. Said that, mostly uses existing products(s) , at times have to fill in the requirements from BA (business analyst ) and provide an overall high level needs for building newer products - which goes to the product engineering teams
  3. Data Layer is the most important of below - always insisted to use standardized data format for communication
  4. Technology Layers  SOAP/HTTP with JSON/XML brings values to standardize
  5. LOOSE COUPLING of systems and components are key here as i had quoted right above in the first
  6. Note : Key SOA solutions are not just ESBs
Read  carefully below diagram two layers of solution design, before that always remember



































Technology Layer Break down of components
  1. Service Container : where data is translated (if needed) to meaningful relational information an example is to translate multi-system information into standardized representation for services, Adapters are used for this translation of data
  2. Broker : is an adapter to ship data from one system to another
  3. Process coordinator : processing or warehousing of data and flow are handled here
Costly Mistake
  1. Many a times the processing of data like business logic are done by broker instead of service container
  2. Also at times broker used in place of process co-ordinator
Loose Coupling
  1. Data in one system means different in another system, always represent the data and the meta data for specific system
  2. Map domain data to message data
  3. Eliminate unnecessary dependencies

Monday, October 7, 2013

Following Arista

I have following arista for last year and jotting down their product scope and my understanding for future references.


  • Arista a spun out from cisco-Ex, have good presence in the market and heard quite a bit times often
  • Arista EOS is equivalent or better than Cisco IOS / Nx-OS
    • I have personally not any experienced the usage difference, but i am a Nx-OS guy
    • Will evaluate and report - when i get a chance
  • Artista moved from switch to relevant areas in market space similar and aligning to their business
  • Arista Introduces Next Generation 7050X Products,Provides Telemetry and Cloud Network Leaf
  • http://www.aristanetworks.com/en/news/pressrelease/630-pr-20131002-01

Bigdata Market share

"Every dog has it day", said that , here is the piece of pie for each big data , official report.

Splunk                 $186 million        Turns machine data into valuable insights
Opera Solutions $118                     Data-Science-as-a-Service
Mu Sigma            $114                       Data-Science-as-a-Service
Palantir                 $78                         Big data software
Cloudera              $61                         Apache Hadoop-based software, services and training
Actian                   $46                         Big data applications, analytics engine, and Ingres database
1010data              $37                         Cloud-based analytics
10gen                    $36                         MongoDB (open-source, document database)
Alteryx                 $36                         Big data analytics platform
Guavus                 $35                         Big data analytics solutions

Tuesday, September 24, 2013

Big data Analytics

Tricky and tricker

I was wondering what is happening to this era of technology and see i a paradigm shift.
Ha nope, today's cloud / distributed computing = 50 year old mainframe
And BDA big data is nothing different than the data in mainframes.

I see it as , when data getting bigger and bigger, there are cheaper memory and powerful hardware available these days, the way of storing data is changing , map based, graph based all in-memory.

Relation in data are not represented in relational databases , data persistence are claimed to be secured in remote cloud.

Telecom Analytics
http://www.guavus.com/

Log Analytics
http://www.sumologic.com/

AWS Amazon web services Analytics
http://www.attribo.com/


Smaller players
http://www.sevacall.com/

Thursday, September 12, 2013

Security SDDC / Hypervisor

Collecting meaningful information here about security in virtualization solution


http://www.virtualizationpractice.com/virtualization-security-team-ups-6437/
https://www.youtube.com/watch?v=haETMMXObUA&feature=em-subs_digest

Virtual network security: A vendor comparison

For all the benefits a virtualized data center offers -- cost savings, agility and scalability -- it presents an equal number of security challenges. As server virtualization projects mature, IT organizations find that they need a new way of protecting systems from threats both inside the network -- think empowered administrators -- and outside the network. Of course, the vendor community is ready to help with a new market of virtual network security products.
“Whatever application you put in the virtual data center, the security assurance for that application data has to be exactly the same as if that application was hosted on the physical infrastructure. So whatever security properties you have for those applications -- firewall laws, network segmentation, antivirus, data-level controls -- have to be replicated in the virtualized infrastructure,” said Chenxi Wang, vice president and principal analyst with Forrester Research.
But the network security controls and practices that network admins put to use in a physical infrastructure don’t carry over to the virtual data center. For example, in a physical infrastructure, admins can segment servers with a firewall, Wang explained. A Web server may be external facing and another server can sit on the internal network. When these two servers are moved to a virtual infrastructure, they may no longer run on two separate pieces of hardware. They could be two virtual machines (VMs) on the same server, but they still need to be segmented.
“The virtual technology today, with the exception of VMware, doesn’t have built in controls, so you have to add third-party technology or do things manually to make sure they’re segmented,” said Wang.
To further complicate matters, the vendors that are addressing virtual network security in the data center each approach the problem a little bit differently. Some vendors offer virtual network security products that deploy as software on a physical server, while other products come as appliances that are deployed directly in the data center, said Wang. “Different delivery models give different operational overhead,” she said.
Only a small number of vendors currently offer virtual network security products for the data center. Among the larger players are Cisco Systems, HP Networking, Juniper Networks and VMware. Smaller players include HyTrust, Vyatta and Catbird.
Cisco’s Virtual Security Gateway and ASA 1000V
Cisco’s virtual security architecture is comprised of the Virtual Security Gateway and the ASA (Adaptive Security Appliance) 1000V Cloud Firewall. Both products integrate with the Cisco Nexus 1000v distributed virtual switch and can run as virtual appliances on an ESX or on the Cisco Nexus 1010 Virtual Services Appliance, making Cisco’s product an option only for Cisco shops. However, the Nexus 1000V supports multiple hypervisors (VMware and soon Microsoft Hyper-V), a benefit for those IT organizations that are running a multi-hypervisor data center.
The Virtual Security Gateway (VSG) is a zone-based firewall designed to protect inter-VM communications within a particular tenant. It provides access control between VMs. VSG integrates with ASA 1000V, which is Cisco’s cloud version of its physical security infrastructure firewall, the Cisco Adaptive Security Appliance (ASA). The Cisco ASA 1000V Cloud Firewall secures the tenant edge.
HP TippingPoint Secure Virtualization Framework
HP addresses virtual network security with its Secure Virtualization Framework, which consists of the HP Virtual Controller (vController), Virtual Firewall (VFW), Virtual Management Center (VMC) and the HP TippingPoint N Series IPS. The Virtual Firewall creates trust zones and performs segmentation across VMs, clusters and application groups. The vController and VFW sit within each hypervisor and apply security policies to traffic going between VMs. Together they dictate which VMs can speak to each other. The vController also sends traffic to the intrusion prevention system (IPS). The TippingPoint N Series IPS inspects traffic and either sends it back to the virtual cluster or drops it, based on policies set within the VMC.
Juniper Networks vGW
Juniper Networks’ vGW is a firewall that sits within the hypervisor and performs security processing within the kernel. It is compatible only with VMware. It is integrated with VMware’s vCenter and managed through a management console, Security Design for vGW. In addition to stateful firewall functionality, vGW includes compliance, antivirus and monitoring and reporting functionalities..  It is based on technology from Altor Networks, a virtual network security specialist that Juniper acquired in late 2010.
VMware vShield
The VMware vShield product line includes vShield App, vShield Edge and vShield Endpoint. VShield Edge is a network and security gateway that protects the virtual data center perimeter. VShield App provides segmentation of inter-VM communications. It is designed to lock down applications to only those ports and services required to make them work. VShield Endpoint offloads antivirus functionality to a dedicated virtual appliance, thereby removing the antivirus agent footprint in VMs. The three software products can be deployed independently or together in a VMware infrastructure.
Vyatta Network OS
Vyatta Network OS, the company’s virtual router software, includes traditional network security functionality -- such as stateful firewall, IPsec and SSL-based VPNs, network intrusion prevention, Web filtering and dynamic routing -- as pre-packaged virtual machines. The software runs on the hypervisor and is compatible with VMware, Xen, XenServer and Red Hat KVM. Vyatta Network OS can be managed via the command line, Vyatta’s Web-based GUI or a third-party management system.
HyTrust Appliance
The HyTrust Appliance is a virtual appliance that is deployed within the VMware infrastructure. The software intercepts administrative requests in the VMware management plane and permits or denies the requests based on defined policy. HyTrust authenticates and verifies users’ identities to prevent unauthorized access to the virtual infrastructure. However, HyTrust also provides network layer protection by helping to enforce network-level policies. For example, if a network admin attempts to connect a VM to the wrong network segment, the HyTrust Appliance will prevent that request. HyTrust can be used with VMware vShield.
Catbird vSecurity
Catbird vSecurity is comprised of a virtual appliance that is deployed inside each virtual host and a Catbird Control Center that serves as the management console. There are four elements within the virtual appliance. VCompliance monitors and enforces compliance. Hypervisor Shield monitors the server and the network to protect the hypervisor against unauthorized access, incorrect configurations and bridging with the public network. VMshield protects the VMs themselves. If a VM’s configurations are not in accordance with policy, then it is quarantined from the rest of the network until it can be remediated. Finally, TrustZones enforces the security policies for individual machines, regardless of their location. TrustZones can be used to segment the network. Catbird vSecurity is available for VMware and XenServer environments.

bee-social