Back to Basics: The Many Modes of Proxies
The simplicity of the term "proxy" belies the complex topological options available. Understanding the different deployment options will enable your proxy deployment to fit your environment and, more importantly, your applications. It seems so simple in theory. A proxy is a well-understood concept that is not peculiar to networking. Indeed, some folks vote by proxy, they speak by proxy (translators), and even on occasion, marry by proxy. A proxy, regardless of its purpose, sits between two entities and performs a service. In network architectures the most common use of a proxy is to provide load balancing services to enable scale, reliability and even performance for applications. Proxies can log data exchanges, act as a gatekeeper (authentication and authorization), scan inbound and outbound traffic for malicious content and more. Proxies are a key strategic point of control in the data center because they are typically deployed as the go-between for end-users and applications. These go-between services are often referred to as virtual services, and for purposes of this blog that's what we'll call them. It's an important distinction because a single proxy can actually act in multiple modes on a per-virtual service basis. That's all pretty standard stuff. What's not simple is when you start considering how you want your proxy to act. Should it be a full proxy? A half proxy? Should it route or forward? There are multiple options for these components and each has its pros and cons. Understanding each proxy "mode" is an important step toward architecting a suitable solution for your environment as the mode determines the behavior of traffic as it traverses the proxy. Standard Virtual Service (Full Application Proxy) The standard virtual service provided by a full proxy fully terminates the transport layer connections (typically TCP) and establishes completely separate transport layer connections to the applications. This enables the proxy to intercept, inspect and ultimate interact with the data (traffic) as its flowing through the system. Any time you need to inspect payloads (JSON, HTML, XML, etc...) or steer requests based on HTTP headers (URI, cookies, custom variables) on an ongoing basis you'll need a virtual service in full proxy mode. A full proxy is able to perform application layer services. That is, it can act on protocol and data transported via an application protocol, such as HTTP. Performance Layer 4 Service (Packet by Packet Proxy) Before application layer proxy capabilities came into being, the primary model for proxies (and load balancers) was layer 4 virtual services. In this mode, a proxy can make decisions and interact with packets up to layer 4 - the transport layer. For web traffic this almost always equates to TCP. This is the highest layer of the network stack at which SDN architectures based on OpenFlow are able to operate. Today this is often referred to as flow-based processing, as TCP connections are generally considered flows for purposes of configuring network-based services. In this mode, a proxy processes each packet and maps it to a connection (flow) context. This type of virtual service is used for traffic that requires simple load balancing, policy network routing or high-availability at the transport layer. Many proxies deployed on purpose-built hardware take advantage of FPGAs that make this type of virtual service execute at wire speed. A packet-by-packet proxy is able to make decisions based on information related to layer 4 and below. It cannot interact with application-layer data. The connection between the client and the server is actually "stitched" together in this mode, with the proxy primarily acting as a forwarding component after the initial handshake is completed rather than as an endpoint or originating source as is the case with a full proxy. IP Forwarding Virtual Service (Router) For simple packet forwarding where the destination is based not on a pooled resource but simply on a routing table, an IP forwarding virtual service turns your proxy into a packet layer forwarder. A IP forwarding virtual server can be provisioned to rewrite the source IP address as the traffic traverses the service. This is done to force data to return through the proxy and is referred to as SNATing traffic. It uses transport layer (usually TCP) port multiplexing to accomplish stateful address translation. The address it chooses can be load balanced from a pool of addresses (a SNAT pool) or you can use an automatic SNAT capability. Layer 2 Forwarding Virtual Service (Bridge) For situations where a proxy should be used to bridge two different Ethernet collision domains, a layer 2 forwarding virtual service an be used. It can be provisioned to be an opaque, semi-opaque, or transparent bridge. Bridging two Ethernet domains is like an old timey water brigade. One guy fills a bucket of water (the client) and hands it to the next guy (the proxy) who hands it to the destination (the server/service) where it's thrown on the fire. The guy in the middle (the proxy) just bridges the gap (you're thinking what I'm thinking - that's where the term came from, right?) between the two Ethernet domains (networks).
Bidirectional Forwarding Detection (BFD) Protocol Cheat Sheet
Definition This is a protocol initially described inRFC5880and IPv4/IPv6 specifics inRFC5881. I would say this is an aggressive 'hello-like' protocol with shorter timers but very lightweight on the wire and requiring very little processing as it is designed to be implemented in forwarding plane (although RFC does not forbid it to be implemented in control plane). It also contains a feature called Echo that further leaves cpu processing cycle to roughly ZERO which literally just 'loops' BFD control packets sent from peer back to them without even 'touching' (processing) it. BFD helps routing protocol detects peers failure at sub-second level and BIG-IP supports it on all its routing protocols. On BIG-IP it is control-plane independent as TMM that takes care of BFD sending/receivingunicastprobes (yes,no multicast!) and BIG-IP's Advanced Routing Module® being responsible only for its configuration (of course!) and to receive state information from TMM that is displayed in show commands. BIG-IP's control plane daemon communicates with TMM isoamd. This daemon starts when BFD is enabled in the route domain like any other routing protocol. BFD Handshake Explained 218: BFD was configured on Cisco Router but not on BIG-IP so neighbour signals BIG-IP sessionstate is Down and no flags 219: I had just enabled BFD on BIG-IP, session state is now Init and only Control Plane Independent flag set¹ 220: Poll flag is set to validate initial bidirectional connectivity (expecting Final flag set in response) 221: BIG-IP sets Final flag and handshake is complete² ¹Control Plane Independentflag is set because BFD is not actively performed by BIG-IP's control plane. BIG-IP's BFD control plane daemon (oamd) just signals TMM what BFD sessions are required and TMM takes care of sending/receiving of all BFD control traffic and informs session state back to Advanced Routing Module's daemon. ²Packets 222-223are just to show that after handshake is finished all flags are cleared unless there is another event that triggers them. Packet 218 is how Cisco Router sees BIG-IP when BFD is not enabled. Control Plane Independent flag on BIG-IP remains though for the reasons already explained above. Protocol fields Diagnostic codes 0 (No Diagnostic): Typically seen when BFD session is UP and there are no errors. 1 (Control Detection Time Expired):BFD Detect Timer expired and session was marked down. 2 (Echo Function Failed):BFD Echo packet loop verification failed, session is marked down and this is the diagnostic code number. 3 (Neighbor Signaled Session Down):If either neighbour advertised state or our local state is down you will see this diagnostic code 4 (Forwarding Planet Reset):When forwarding-plane (TMM) is reset and peer should not rely on BFD so session is marked down. 5 (Path Down):Ondemand mode external application can signal BFD that path is down so we can set session down using this code 6 Concatenated Path Down): 7 (Administratively Down):Session is marked down by an administrator 8 (Reverse Concatenated Path Down): 9-31:Reserved for future use BFD verification 'show' commands ³Type IP address to see specific session Modes Asynchronous(default): hello-like mode where BIG-IP periodically sends (and receives) BFD control packets and if control detection timer expires, session is marked as down.It uses UDP port3784. Demand: BFD handshake is completed but no periodic BFD control packets are exchanged as this mode assumes system has its own way of verifying connectivity with peer and may trigger BFD verification on demand, i.e. when it needs to use it according to its implementation.BIG-IP currently does not support this mode. Asynchronous + Echo Function: When enabled, TMM literally loops BFDecho-specificcontrol packetson UDP port 3785sent from peers back to them without processing it as it wasn't enough that this protocol is already lightweight. In this mode, a less aggressive timer (> 1 second) should be used for regularBFD control packets over port 3784 and more aggressive timer is used by echo function.BIG-IP currently does not support this mode. Header Fields Protocol Version:BFD version used. Latest one is v1 (RFC5880) Diagnostic Code: BFD error code for diagnostics purpose. Session State: How transmitting system sees the session state which can be AdminDown, Down, Init or Up. Message Flags:Additional session configuration or functionality (e.g. flag that says authentication is enabled) Detect Time Multiplier:Informs remote peer BFD session is supposed to be marked down ifDesired Min TX Intervalmultiplied by this value is reached Message Length(bytes):Length of BFD Control packet My Discriminator:For each BFD session each peer will use a unique discriminator to differentiate multiple session. Your Discriminator:When BIG-IP receives BFD control message back from its peer we add peer's My Discriminator to Your Discriminator in our header. Desired Min TX Interval(microseconds):Fastest we can send BFD control packets to remote peer (no less than configured value here) Required Min RX Interval(microseconds):Fastest we can receive BFD control packets from remote peer (no less than configured value here) Required Min Echo Interval(microseconds):Fastest we can loop BFD echo packets back to remote system (0 means Echo function is disabled) Session State AdminDown:Administratively forced down by command Down: Either control detection time expired in an already established BFD session or it never came up.If probing time (min_tx) is set to 100ms for example, and multiplier is 3 then no response after 300ms makes system go down. Init:Signals a desire to bring session up in the beginning of BFD handshake. Up:Indicates session is Up Message Flags Poll:Pool flag is just a 'ping' that requires peer box to respond with Final flag. In BFD handshake as well as in Demand mode pool message is a request to validate bidirectional connectivity. Final:Sent in response to packet with Poll bit set Control Plane Independent:Set if BFD can continue to function if control plane is disrupted¹ Authentication Present:Only set if authentication is being used Demand:If set, it is implied that periodic BFD control packets are no longer sent and another mechanism (on demand) is used instead. Multipoint:Reserved for future use of point-to-multipoint extension. Should be 0 on both sides. ¹ This is the case for BIG-IP as BFD is implemented in forwarding plane (TMM) BFD Configuration Configure desired transmit and receive intervals as well as multiplier on BIG-IP. And Cisco Router: You will typically configure the above regardless of routing protocol used. BFD BGP Configuration And Cisco Router: BFD OSPFv2/v3 Configuration BFD ISIS Configuration BFD RIPv1/v2 Configuration BFD Static Configuration All interfaces no matter what: Specific interface only: Tie BFD configuration to static route:HTML5 Web Sockets Changes the Scalability Game
#HTML5 Web Sockets are poised to completely change scalability models … again. Using Web Sockets instead of XMLHTTPRequest and AJAX polling methods will dramatically reduce the number of connections required by servers and thus has a positive impact on performance. But that reliance on a single connection also changes the scalability game, at least in terms of architecture. Here comes the (computer) science… If you aren’t familiar with what is sure to be a disruptive web technology you should be. Web Sockets, while not broadly in use (it is only a specification, and a non-stable one at that) today is getting a lot of attention based on its core precepts and model. Web Sockets Defined in the Communications section of the HTML5 specification, HTML5 Web Sockets represents the next evolution of web communications—a full-duplex, bidirectional communications channel that operates through a single socket over the Web. HTML5 Web Sockets provides a true standard that you can use to build scalable, real-time web applications. In addition, since it provides a socket that is native to the browser, it eliminates many of the problems Comet solutions are prone to. Web Sockets removes the overhead and dramatically reduces complexity. - HTML5 Web Sockets: A Quantum Leap in Scalability for the Web So far, so good. The premise upon which the improvements in scalability coming from Web Sockets are based is the elimination of HTTP headers (reduces bandwidth dramatically) and session management overhead that can be incurred by the closing and opening of TCP connections. There’s only one connection required between the client and server over which much smaller data segments can be sent without necessarily requiring a request and a response pair. That communication pattern is definitely more scalable from a performance perspective, and also has a positive impact of reducing the number of connections per client required on the server. Similar techniques have long been used in application delivery (TCP multiplexing) to achieve the same results – a more scalable application. So far, so good. Where the scalability model ends up having a significant impact on infrastructure and architectures is the longevity of that single connection: Unlike regular HTTP traffic, which uses a request/response protocol, WebSocket connections can remain open for a long time. - How HTML5 Web Sockets Interact With Proxy Servers This single, persistent connection combined with a lot of, shall we say, interesting commentary on the interaction with intermediate proxies such as load balancers. But ignoring that for the nonce, let’s focus on the “remain open for a long time.” A given application instance has a limit on the number of concurrent connections it can theoretically and operationally manage before it reaches the threshold at which performance begins to dramatically degrade. That’s the price paid for TCP session management in general by every device and server that manages TCP-based connections. But Lori, you’re thinking, HTTP 1.1 connections are persistent, too. In fact, you don’t even have to tell an HTTP 1.1 server to keep-alive the connection! This really isn’t a big change. Whoa there hoss, yes it is. While you’d be right in that HTTP connections are also persistent, they generally have very short connection timeout settings. For example, the default connection timeout for Apache 2.0 is 15 seconds and for Apache 2.2 a mere 5 seconds. A well-tuned web server, in fact, will have thresholds that closely match the interaction patterns of the application it is hosting. This is because it’s a recognized truism that long and often idle connections tie up server processes or threads that negatively impact overall capacity and performance. Thus the introduction of connections that remain open for a long time changes the capacity of the server and introduces potential performance issues when that same server is also tasked with managing other short-lived, connection-oriented requests. Why this Changes the Game… One of the most common inhibitors of scale and high-performance for web applications today is the deployment of both near-real-time communication functions (AJAX) and traditional web content functions on the same server. That’s because web servers do not support a per-application HTTP profile. That is to say, the configuration for a web server is global; every communication exchange uses the same configuration values such as connection timeouts. That means configuring the web server for exchanges that would benefit from a longer time out end up with a lot of hanging connections doing absolutely nothing because they were used to grab standard dynamic or static content and then ignored. Conversely, configuring for quick bursts of requests necessarily sets timeout values too low for near or real-time exchanges and can cause performance issues as a client continually opens and re-opens connections. Remember, an idle connection is a drain on resources that directly impacts the performance and capacity of applications. So it’s a Very Bad Thing™. One of the solutions to this somewhat frustrating conundrum, made more feasible by the advent of cloud computing and virtualization, is to deploy specialized servers in a scalability domain-based architecture using infrastructure scalability patterns. Another approach to ensuring scalability is to offload responsibility for performance and connection management to an appropriately capable intermediary. Now, one would hope that a web server implementing support for both HTTP and Web Sockets would support separately configurable values for communication settings on at least the protocol level. Today there are very few web servers that support both HTTP and Web Sockets. It’s a nascent and still evolving standard so many of the servers are “pure” Web Sockets servers, many implemented in familiar scripting languages like PHP and Python. Which means two separate sets of servers that must be managed and scaled. Which should sound a lot like … specialized servers in a scalability domain-based architecture. The more things change, the more they stay the same. The second impact on scalability architectures centers on the premise that Web Sockets keep one connection open over which message bits can be exchanged. This ties up resources, but it also requires that clients maintain a connection to a specific server instance. This means infrastructure (like load balancers and web/application servers) will need to support persistence (not the same as persistent, you can read about the difference here if you’re so inclined). That’s because once connected to a Web Socket service the performance benefits are only realized if you stay connected to that same service. If you don’t and end up opening a second (or Heaven-forbid a third or more) connection, the first connection may remain open until it times out. Given that the premise of the Web Socket is to stay open – even through potentially longer idle intervals – it may remain open, with no client, until the configured time out. That means completely useless resources tied up by … nothing. Persistence-based load balancing is a common feature of next-generation load balancers (application delivery controllers) and even most cloud-based load balancing services. It is also commonly implemented in application server clustering offerings, where you’ll find it called server-affinity. It is worth noting that persistence-based load balancing is not without its own set of gotchas when it comes to performance and capacity. THE ANSWER: ARCHITECTURE The reason that these two ramifications of Web Sockets impacts the scalability game is it requires an broader architectural approach to scalability. It can’t necessarily be achieved simply by duplicating services and distributing the load across them. Persistence requires collaboration with the load distribution mechanism and there are protocol-based security constraints with respect to incorporating even intra-domain content in a single page/application. While these security constraints are addressable through configuration, the same caveats with regards to the lack of granularity in configuration at the infrastructure (web/application server) layer must be made. Careful consideration of what may be accidentally allowed and/or disallowed is necessary to prevent unintended consequences. And that’s not even starting to consider the potential use of Web Sockets as an attack vector, particularly in the realm of DDoS. The long-lived nature of a Web Socket connection is bound to be exploited at some point in the future, which will engender another round of evaluating how to best address application-layer DDoS attacks. A service-focused, distributed (and collaborative) approach to scalability is likely to garner the highest levels of success when employing Web Socket-based functionality within a broader web application, as opposed to the popular cookie-cutter cloning approach made exceedingly easy by virtualization. Infrastructure Scalability Pattern: Partition by Function or Type Infrastructure Scalability Pattern: Sharding Sessions Amazon Makes the Cloud Sticky Load Balancing Fu: Beware the Algorithm and Sticky Sessions Et Tu, Browser? Forget Hyper-Scale. Think Hyper-Local Scale. Infrastructure Scalability Pattern: Sharding Streams Infrastructure Architecture: Whitelisting with JSON and API Keys Does This Application Make My Browser Look Fat? HTTP Now Serving … EverythingPersistent and Persistence, What's the Difference?
The English language is one of the most expressive, and confusing, in existence. Words can have different meaning based not only on context, but on placement within a given sentence. Add in the twists that come from technical jargon and suddenly you've got words meaning completely different things. This is evident in the use of persistent and persistence. While the conceptual basis of persistence and persistent are essentially the same, in reality they refer to two different technical concepts. Both persistent and persistence relate to the handling of connections. The former is often used as a general description of the behavior of HTTP and, necessarily, TCP connections, though it is also used in the context of database connections. The latter is most often related to TCP/HTTP connection handling but almost exclusively in the context of load-balancing. Persistent Persistent connections are connections that are kept open and reused. The most commonly implemented form of persistent connections are HTTP, with database connections a close second. Persistent HTTP connections were implemented as part of the HTTP 1.1 specification as a method of improving the efficiency Related Links HTTP 1.1 RFC Persistent Connection Behavior of Popular Browsers Persistent Database Connections Apache Keep-Alive Support Cookies, Sessions, and Persistence of HTTP in general. Before HTTP 1.1 a browser would generally open one connection per object on a page in order to retrieve all the appropriate resources. As the number of objects in a page grew, this became increasingly inefficient and significantly reduced the capacity of web servers while causing browsers to appear slow to retrieve data. HTTP 1.1 and the Keep-Alive header in HTTP 1.0 were aimed at improving the performance of HTTP by reusing TCP connections to retrieve objects. They made the connections persistent such that they could be reused to send multiple HTTP requests using the same TCP connection. Similarly, this notion was implemented by proxy-based load-balancers as a way to improve performance of web applications and increase capacity on web servers. Persistent connections between a load-balancer and web servers is usually referred to as TCP multiplexing. Just like browsers, the load-balancer opens a few TCP connections to the servers and then reuses them to send multiple HTTP requests. Persistent connections, both in browsers and load-balancers, have several advantages: Less network traffic due to less TCP setup/teardown. It requires no less than 7 exchanges of data to set up and tear down a TCP connection, thus each connection that can be reused reduces the number of exchanges required resulting in less traffic. Improved performance. Because subsequent requests do not need to setup and tear down a TCP connection, requests arrive faster and responses are returned quicker. TCP has built-in mechanisms, for example window sizing, to address network congestion. Persistent connections give TCP the time to adjust itself appropriately to current network conditions, thus improving overall performance. Non-persistent connections are not able to adjust because they are open and almost immediately closed. Less server overhead. Servers are able to increase the number of concurrent users served because each user requires fewer connections through which to complete requests. Persistence Persistence, on the other hand, is related to the ability of a load-balancer or other traffic management solution to maintain a virtual connection between a client and a specific server. Persistence is often referred to in the application delivery networking world as "stickiness" while in the web and application server demesne it is called "server affinity". Persistence ensures that once a client has made a connection to a specific server that subsequent requests are sent to the same server. This is very important to maintain state and session-specific information in some application architectures and for handling of SSL-enabled applications. Examples of Persistence Hash Load Balancing and Persistence LTM Source Address Persistence Enabling Session Persistence 20 Lines or Less #7: JSessionID Persistence When the first request is seen by the load-balancer it chooses a server. On subsequent requests the load-balancer will automatically choose the same server to ensure continuity of the application or, in the case of SSL, to avoid the compute intensive process of renegotiation. This persistence is often implemented using cookies but can be based on other identifying attributes such as IP address. Load-balancers that have evolved into application delivery controllers are capable of implementing persistence based on any piece of data in the application message (payload), headers, or at in the transport protocol (TCP) and network protocol (IP) layers. Some advantages of persistence are: Avoid renegotiation of SSL. By ensuring that SSL enabled connections are directed to the same server throughout a session, it is possible to avoid renegotiating the keys associated with the session, which is compute and resource intensive. This improves performance and reduces overhead on servers. No need to rewrite applications. Applications developed without load-balancing in mind may break when deployed in a load-balanced architecture because they depend on session data that is stored only on the original server on which the session was initiated. Load-balancers capable of session persistence ensure that those applications do not break by always directing requests to the same server, preserving the session data without requiring that applications be rewritten. Summize So persistent connections are connections that are kept open so they can be reused to send multiple requests, while persistence is the process of ensuring that connections and subsequent requests are sent to the same server through a load-balancer or other proxy device. Both are important facets of communication between clients, servers, and mediators like load-balancers, and increase the overall performance and efficiency of the infrastructure as well as improving the end-user experience.
SDN: An architecture for operationalizing networks
As we heard at last week’s Open Networking Summit 2014, managing change and complexity in data centers is becoming increasingly difficult as IT and operations are constantly being pressured to deliver more features, more services, and more applications at ever increasing rates. The solution is to design the network for rapid evolution in a controlled and repeatable manner similar to how modern web applications are deployed. This is happening because it is no longer sufficient for businesses to deliver a consistent set of services to their customers. Instead, the world has become hyper-competitive and it has become necessary to constantly deliver new features to not only capture new customers but to retain existing customers. This new world order poses a significant conflict for the operations side of the business as their charter is to ensure continuity of service and have traditionally used careful (often expensive) planning efforts to ensure continuity of service when changes are needed. The underling problem is that the network is not operationalized and instead changes are accomplished through manual and scripted management. The solution for operations is to move towards architectures that are designed for rapid evolution and away from manual and scripted processes. Software Defined Networking address these challenges by defining a family of architectures for solving these types operational challenges and operations teams are latching on with a rarely seen appetite. The key to the success of SDN architectures is the focus on abstraction of both the control and data planes for the entire network via open APIs (not just the stateless Layer 0-4 portions). The first layer of abstraction allows for a centralized control plane system called an SDN Controller, or just Controller, that understands the entire configuration of the network and programmatically configures the network increasing the velocity and fidelity of configurations by removing humans from configuration loop – humans are still needed to teach the Controller what to do. These Controllers allow for introspection on the configuration and allow for automated deployments. As Controllers mature, I expect them to gain the capabilities of a configuration management system (CMS) allowing for network architects to rapidly revert changes virtually instantaneously. The second layer of abstraction allows for network architects or third parties to programmatically extend the capabilities of a data path element. This can be as seemingly simple as adding a match-and-forward rule to a switch (e.g., OpenFlow) or as seemingly complex as intercepting a fully parsed HTTP request, parsing an XML application object contained within, and then interrogating a database for a forwarding decision (e.g., LineRate and F5) based on the parsed application data. However, realizing the fully operational benefits of SDN architectures requires that the entire network be designed with SDN architectural principles including both the stateless components (e.g., switching, routing, and virtual networking) and the stateful components (e.g., L4 firewalls, L7 application firewalls, and advanced traffic mangement). Early on SDN proponents, as SDN evolved from a university research project, proposed pure stateless Layer 2-3 networking ignoring the complexities of managing modern networks that call for stateful L4-7 services. The trouble with this approach is that every additional operational domain disproportionately increases operational complexities, as the domains need to be “manually” kept in sync. Recognizing this need, major Layer 2-4 vendors, including Cisco, have formed partnerships with F5 and other stateful Layer 4-7 vendors to complement their portfolios. With the goal of helping customers operationalize their networks, I offer the following unifying definition of SDN for discussion: “SDN is a family of architectures (not technologies) for operationalizing networks with reduced operating expenses, reduced risks, and improved time to market by centralizing control into a control plane that programmatically configures and extends all network data path elements and services via open APIs.” Over the next few months I’ll dig deeper into different aspects of SDN – stay tuned!Totally Unscientific and Informal SDN Adoption Poll Results
#SDN During a session at Gartner Data Center 2013 (Software Defined Networking? Not a Question of If, A Matter of When and How) the question of SDN adoption was asked by speakers Joe Skorupa and Andrew Lerner of an approximately 100 person audience. These are the (very informal and totally unscientific results).The Disadvantages of DSR (Direct Server Return)
I read a very nice blog post yesterday discussing some of the traditional pros and cons of load-balancing configurations. The author comes to the conclusion that if you can use direct server return, you should. I agree with the author's list of pros and cons; DSR is the least intrusive method of deploying a load-balancer in terms of network configuration. But there are quite a few disadvantages missing from the author's list. Author's List of Disadvantages of DSR The disadvantages of Direct Routing are: Backend server must respond to both its own IP (for health checks) and the virtual IP (for load balanced traffic) Port translation or cookie insertion cannot be implemented. The backend server must not reply to ARP requests for the VIP (otherwise it will steal all the traffic from the load balancer) Prior to Windows Server 2008 some odd routing behavior could occur in In some situations either the application or the operating system cannot be modified to utilse Direct Routing. Some additional disadvantages: Protocol sanitization can't be performed. This means vulnerabilities introduced due to manipulation of lax enforcement of RFCs and protocol specifications can't be addressed. Application acceleration can't be applied. Even the simplest of acceleration techniques, e.g. compression, can't be applied because the traffic is bypassing the load-balancer (a.k.a. application delivery controller). Implementing caching solutions become more complex. With a DSR configuration the routing that makes it so easy to implement requires that caching solutions be deployed elsewhere, such as via WCCP on the router. This requires additional configuration and changes to the routing infrastructure, and introduces another point of failure as well as an additional hop, increasing latency. Error/Exception/SOAP fault handling can't be implemented. In order to address failures in applications such as missing files (404) and SOAP Faults (500) it is necessary for the load-balancer to inspect outbound messages. Using a DSR configuration this ability is lost, which means errors are passed directly back to the user without the ability to retry a request, write an entry in the log, or notify an administrator. Data Leak Prevention can't be accomplished. Without the ability to inspect outbound messages, you can't prevent sensitive data (SSN, credit card numbers) from leaving the building. Connection Optimization functionality is lost. TCP multiplexing can't be accomplished in a DSR configuration because it relies on separating client connections from server connections. This reduces the efficiency of your servers and minimizes the value added to your network by a load balancer. There are more disadvantages than you're likely willing to read, so I'll stop there. Suffice to say that the problem with the suggestion to use DSR whenever possible is that if you're an application-aware network administrator you know that most of the time, DSR isn't the right solution because it restricts the ability of the load-balancer (application delivery controller) to perform additional functions that improve the security, performance, and availability of the applications it is delivering. DSR is well-suited, and always has been, to UDP-based streaming applications such as audio and video delivered via RTSP. However, in the increasingly sensitive environment that is application infrastructure, it is necessary to do more than just "load balancing" to improve the performance and reliability of applications. Additional application delivery techniques are an integral component to a well-performing, efficient application infrastructure. DSR may be easier to implement and, in some cases, may be the right solution. But in most cases, it's going to leave you simply serving applications, instead of delivering them. Just because you can, doesn't mean you should.The Bifurcation of the Network
#SDN #Cloud #Virtualization #devops Emerging technology is bifurcating not only operations, but the network itself A funny thing happened on the way to the next generation data center architecture... the network split. Not just operationally, with devops taking on more and more operational responsibility for application-focused infrastructure, and not just topologically. There is a bifurcation of the network occurring that's more than logical, it's physical - with boundaries and everything. Rather than become more fluid, as one would expect with such liquefying technology such as virtualization and cloud, a core network is emerging that is encased in a firm shell through which there is no passage for application-focused elements. In this blog we shall henceforth refer to the two networks as the "core" network and the "business" network. Driven by cloud and virtualization, application architectures have become what developers refer to as a "singleton" pattern. It's a Highlander pattern. You know, "there can be only one". An "application" is becoming isolated - along with its supporting infrastructure. That in itself is not a bad thing, because it recognizes that there are infrastructure elements and services that are an integral component of an "application" such as load balancing, caching, and application-specific security. But it does act as a forcing factor on the network architecture that results in a more tenant (and by tenant I mean application) oriented architecture. Core network services, on the other hand, remain shared. Organizational security focused on preventing network attacks from succeeding in denial of service or infection by malware remains shared; it is an organizational concern that remains part of the core network. But specific application security and infrastructure services are increasingly moving inward in more isolated pocket architectures within a business network to be deployed and managed along with the applications they support. The business network may exist primarily in the virtual domain, being wholly topological in nature and, increasingly, software defined. Comprised of virtual machines hosting applications, infrastructure services, and leveraging virtual switches and services, the business network is beginning to fall into the demesne of devops and business owners rather than that of the traditional network professional. Meanwhile, the core network remains firmly in the domain of the network professional. Physical routers and switches and infrastructure, whether traditional or SDN-enabled, remain the norm here in the core network, owing to the shared nature of the services and the need for extreme scalability, reliability and simply, bandwidth. You can see the impact of this change occurring when you investigate how developers and devops spend much of their time. They spend it on troubleshooting, configuring, automating, and deploying. And it's increasingly common to see developers spending time on what are traditionally "network" responsibilities. Electric Cloud conducted a developer survey on productivity in March 2013. The results? Folks in application development indicated they spend 2-5 hours per week on environmental configuration. While this has implications for devops and, in particular, speaks to the need for automation, it also tells a story; a story of a more involved application development organization in managing the "environment" in which an application is ultimately deployed. There is a shift in the power balance of the data center, and it is leaning toward devops and developers. It is because the network is so critical to the business that this shift is occurring. Isolation of core connectivity and security from the increasingly agile, continuous development and deployment cycles supporting the business. This shift in power and responsibility is game-changing and ultimately it's going to result in a lot of very interesting changes across the entire industry.1024 Words: Application Health Monitoring at a Glance
It's not enough to measure connectivity or even status codes… See also Axiom #1 of the Three Axioms of Application Delivery 1024 Words: Why Devops is Hard 1024 Words: The Theory of (Performance) Relativity Filling the SLA Gap in the Cloud Visibility: Keystone to Rapid Recovery Everywhere All Your Packets Are Belong to … You? The Three Axioms of Application Delivery
Bare Metal Blog: Introduction
#f5 #baremetalblog Introducing the geekiest bits of F5 hardware in a new blog series: Bare Metal Blog. Just about every high-tech vendor on the planet has some form of secret sauce that makes them believe they’re the best option for your IT needs. For some it is price, for others service, yet others point at integration. All have some level of support for all of the others, but believe they excel at one more than the others. I’m not talking about the bit they tell you when you ask, I’m talking about the bit they talk about when you’re not around. F5 does really well at all of the above, but the secret sauce, the thing that makes all of the above important, is the integration of rock-solid software running on a VM or hardware designed internally. By controlling all phases of development, the systems are purpose-designed from end-to-end to deliver applications fast and enable ADC functionality beyond the norm. By extension, our Virtual Editions are cool and work really well, especially in hybrid cloud environments, but for sheer geek madness, our hardware platforms are insane. Since we don’t often talk much about F5 hardware and how it comes to be what it is, this blog series is part of an effort to give you just a peek inside our hardware development environment. We keep a cadre of rock stars stashed away designing ever more powerful solutions, and even if you are the most devoted user of F5 gear, you probably do not know any of their names or realize what they do to bring high performing, high reliability hardware to your office. We’re going to change just a bit of that here, by talking about a topic a week, and sometimes about the rockstars that make these topics work. When it comes down to it, there are instances where hardware – particularly finely tuned special purpose hardware – just cannot be beat for performance and scalability. When that hardware is generally utilized to improve performance or add security without impacting performance, the difference can be amazing. By way of example, hardware-based compression for application acceleration is implemented throughout our line of physical ADCs. The performance edge that offers is often unknown or underestimated. In the course of this blog series, we’ll delve into how the hardware is tested, what benefits hardware offers that software or our Virtual Editions cannot, how to test hardware to get real-world test results that are also apples-to-apples across vendors, myths that actually sell more hardware but you can use for capacity planning, MTBF and what it means to your organization, and, well, a ton of other things that I have to choose from. From time to time others who have knowledge of this topic will post using the same #baremetalblog hashtag, so you can search DevCentral for baremetalblog to get the entire array of blogs, no matter which of us wrote them. The goal (aside from talking about the cool stuff we don’t normally discuss) is to arm you with questions you should be asking any networking vendor about performance, quality, sourcing of parts, and integration. We’ll keep it up for as long as we have content, which is likely to be months. I’m looking forward to writing it, hopefully it is as informative and evocative as I expect it to be. Related Articles and Blogs: When Did Specialized Hardware Become a Dirty Word? Hardware Acceleration Critical Component for Cost-Conscious Data ... Red Herring: Hardware versus Services The Conflation of Pay-as-you-Grow Hardware with On-Demand The (hopefully) definitive guide to load balancing Lync Edge Servers ... Audio White Paper - Application Delivery Hardware A Critical ... Sometimes It Is About the Hardware Cloud Lets You Throw More Hardware at the Problem Faster LTM HA - network or hardware failover? - DevCentral - DevCentral ...
