Xen FAQs
- What is Xen?
- Who created Xen?
- Who uses Xen?
- What technology vendors offer products supporting Xen?
- What are the business problems solved by Xen?
- How does Xen address these problems?
- What are the business benefits of adopting Xen?
- How does Xen differ from other virtualization technologies?
- What version of Xen is currently available?
- What’s new in Xen 3.0?
- What operating systems are supported by Xen?
- What is the Xen approach to security?
- How fast is Xen?
- What are the technological differences between Xen and other virtualization products?
- What is a virtual server?
- What is paravirtualization?
Xen is an open-source hypervisor that enables increased server utilization and server consolidation by enabling multiple operating system images to simultaneously run on a single physical server. Xen provides resource guarantees to virtual servers to ensure that application layer SLAs are met, including CPU, memory and I/O guarantees. Xen is the industry’s fastest and most secure infrastructure virtualization software technology, and has been endorsed and adopted by over 20 of the industry’s major vendors, including AMD, Dell, Egenera, Hewlett-Packard, IBM, Intel, Mellanox Technologies, Network Appliance, Novell, Red Hat, Sun Microsystems, TopSpin, Unisys and Voltaire. Xen is licensed under the General Public License (GPL), and is available free for download. It is offered by XenSource and other vendors as a supported enterprise software product.
[ back to FAQs ]
Xen was created in 2003 at the University of Cambridge Computer Laboratory in what’s known as the Xen Hypervisor project led by Ian Pratt. Additional team members include Keir Fraser, Steven Hand, and Christian Limpach. This same team founded XenSource with experienced Silicon Valley technology entrepreneurs Nick Gault and Simon Crosby.
[ back to FAQs ]
Xen is increasingly being deployed in enterprise data centers with the goal of increasing server utilization, enabling server consolidation, and improving total cost of ownership. Xen is widely used in by application service providers and hosting companies because it offers precise control of system resources and allows users to host more virtual servers per physical device than competitive technologies. Xen is increasingly being used in test and development activities across the industry – presently for Linux based applications, but with increased attention to Windows applications. In this use case, virtualization allows developers of multi-tier applications to host multiple virtual machines on a single physical server to check that the application works – saving expense on additional infrastructure testing. Moreover, test hardware can be instantly re-purposed to support the test/dev activities of another application, by simply instantiating new Virtual Servers containing the images to be tested. Finally, applications that have been tested and are ready for roll-out in production can be migrated directly from the Xen test/dev environment into production, by simply re-locating the virtual machine containing the application.
[ back to FAQs ]
OEMs are developing products that include Xen as a supported platform component on every server. System vendors are including support for Xen in their infrastructure management solutions for enterprise data centers, and many embedded vendors are using Xen as a tool to allow rapid development of complex new appliance products. We expect many of these products to reach market in 2006. Vendors developing solutions for Xen include AMD, Dell, Egenera, Hewlett-Packard, IBM, Intel, Mellanox Technologies, Network Appliance, Novell, Red Hat, Sun Microsystems, TopSpin, Unisys and Voltaire.
[ back to FAQs ]
Enterprise data centers are running out of room, because the success of scale-out computing using industry standard servers has led to a proliferation of devices, each of which typically runs a single application workload. This proliferation is expensive: for each dollar spent on hardware, an enterprise typically spends between $5 and $7 per year on associated costs: provisioning, power and cooling, “real estate” for the device, and (most importantly) the human cost of managing the device. Depending on the industry sector, the cost per server per year ranges from $8,000 to $15,000. Most of these servers run at about 10-15% utilization – a tremendous waste of money.
[ back to FAQs ]
Xen reduces sever sprawl by enabling increased server utilization. This enables data centers to consolidate servers, and reduce wasted capital costs and personnel management expenses. It also reduces data center management complexity. For every dollar saved on CapEx on servers, another $5-6 per year on soft costs is also saved. One of XenSource’s F100 customers estimates savings of the order of $100M over 5 years, using Xen virtualization.
[ back to FAQs ]
Adopting virtualization using open source Xen provides a number of business benefits, including:
- Increased server utilization: Data centers can achieve dramatic improvements in server utilization and enable server consolidation, reducing wasted capital costs and personnel management expenses.
- Reduced burden on IT: By enabling server consolidation, Xen reduces the cost, complexity, and personnel time required for data center server management. XenSource’s virtualization infrastructure software solutions provide both Xen based virtualization for servers and management, and control and automation tools that allow IT organizations to “operationalize” virtualization in a seamless, low cost fashion.
- Lowest TCO: Today virtualization has yet to enter the heart of the data center. There are two reasons: poor performance of proprietary hypervisors and their tremendous cost. Xen is the best performing hypervisor, and it’s free. That radically changes the economics of the adoption of virtualization.
[ back to FAQs ]
Unlike other virtualization technologies, only Xen is entirely open source. This brings a number of benefits over proprietary solutions, including improved functionality, better performance, and greater extendibility. Xen is without doubt the highest performing hypervisor in the industry – with typically 10x less overhead than competitive proprietary offerings. Xen’s unique performance benefits accrue from its pioneering and industry leading paravirtualization technology, which allows hosted virtual servers to collaborate with the hypervisor to achieve the best performance for enterprise applications. Other vendors are now rushing to implement paravirtualization, but are at least three years behind Xen. Xen also optimally uses the hardware virtualization capabilities of Intel’s VT and AMD’s Pacifica processors. Unlike other proprietary hypervisors which rely on dated, software-only virtualization, Xen is the industry’s first supported software base for Intel VT. Xen runs unmodified guests such as Windows, on “the bare metal” at native processor speed on Intel VT enabled hardware. Paravirtualization in this case provides I/O performance that Intel VT cannot provide, while still using the best in hardware support for accelerated performance of virtualization. Finally, since Xen is free, it results in a much lower total cost of ownership.
[ back to FAQs ]
Xen 3.0 was released in December 2005. The previous version, Xen 2.0, was released in November 2004.
[ back to FAQs ]
Xen 3.0 supports key new features required for data center applications in large enterprise environments. Features new to this release include support for hardware virtualized guests on Intel® VT-x hardware virtualization technology, up to 32-way SMP guest workloads, and support for 32 bit, Physical Address Extensions (for servers with more than 4GB memory), and 64 bit processors – making it the fastest, leanest, and most secure virtualization technology available.
[ back to FAQs ]
Currently Xen supports Linux 2.4, 2.6, and NetBSD 2.0. XenSource demonstrated Windows XP running on Xen at the Intel Developer Forum (IDF) in August 2005, and will deliver a commercial support offering for this in early 2006.
[ back to FAQs ]
Xen supports absolute resource isolation between domains meaning it has the highest level of separation and security possible in i386 class hardware. You won't, for example, be able to tcpdump on a virtual host and see traffic intended for other virtual hosts. Additionally, Xen’s code base is very small – under 40,000 lines for the core hypervisor. This allows the security community to verify its security continually. More importantly, Xen can use hardware security capabilities, such as Trusted Platform Modules (TPMs) to build a layer of attestation and trust up from the hardware, through the software. XenSource demonstrated a secure hypervisor at Intel Developer Forum in August 2005. The secure solution is an integration of the Xen hypervisor with the market leading open source Snort Intrusion Detection System. By embedding security capabilities into the hypervisor, users receive a powerful new ability to implement the same security policies across the virtualized enterprise, independent of the operating system. Moreover, the hypervisor can ensure that even legacy guests that have not been patched will be protected. Xen can even prevent a compromised virtual machine from attacking other virtual or physical servers in the enterprise by blocking its network traffic.
Finally, XenSource, IBM and Intel are collaborating on a project to deliver a key security capability using Xen. So-called multi-layer secure systems (MLS) allow the hypervisor and its security to be independently managed, monitored and controlled from that of the guests – effectively providing yet another layer of independent security, outside the guest operating system.
[ back to FAQs ]
Xen is without doubt the highest performing hypervisor in the industry – with typically 10x less overhead than competitive proprietary offerings. Xen’s unique performance benefits accrue from its pioneering and industry leading paravirtualization technology, which allows hosted virtual servers to collaborate with the hypervisor to achieve the best performance for enterprise applications.
[ back to FAQs ]
Xen is designed to function more like the mainframe partitioning systems found on multi million dollar systems from vendors like IBM and less like the pseudo-separated systems that are available today from several other vendors such as VServer and UML. VMWare would be the closest analog to what Xen does. The main difference is that operating systems running on Xen know they're running in a virtual server, and not on the raw iron.
[ back to FAQs ]
A virtual server is simply an instance of a server operating system, such as Red Hat Advanced Server, or SUSE SLES, or Microsoft Windows Server 2003, together with its application workload, running on the Xen hypervisor. Rather than controlling hardware directly, the OS instance accesses hardware through the hypervisor, which also has the ability to share the hardware resources with other virtualized OS instances and applications.
[ back to FAQs ]
Paravirtualization is key to Xen's ability to achieve drastically better performance than alternative approaches to virtualization available on the market today. Paravirtualization involves making the virtual server OS aware of the fact that it is being virtualized, and enabling the two to collaborate to achieve optimal performance. On Linux, BSD, Solaris x86 the paravirtualized guest sees Xen as an idealized hardware layer – a new form of hardware. Indeed Xen is simply an idealized hardware architecture for the kernel.org Linux tree maintained at the OSDL. For Windows and other guests that are unaware of Xen, the hardware virtualization of Intel VT, combined with paravirtualizing device drivers in Windows, allows Xen to achieve the same high level of performance as virtualized Linux guests.
[ back to FAQs ]
|