Talking Heads: TIBCO: Accommodating scale, granularity for instrumentation, and actionable, real-time intelligence
TIBCO Software Inc. provides real-time infrastructure software for companies to use on-premise or in cloud computing environments to capture the right information, at the right time, and act pre-emptively for a competitive advantage, what they call ‘the two-second advantage’.
Whether it’s processing millions of smart meter events, interactively analysing diagnostics generated by personal medical devices, or correlating numerous automobile operating parameters, environments, and locations, TIBCO seeks to leverage contextual awareness to the business environments of M2M.
Here, M2M Now’s editor, Jeremy Cowan talks to Inderjeet Singh, SVP and GM, Products and Technology at TIBCO, about how M2M service providers can help their customers act on timely intelligence (while managing exponential growth in sophisticated M2M devices and dynamic services) and provide diverse user communities with contextual mobility solutions.
M2M Now: Inderjeet, what do you see as the greatest challenge today in instrumenting the machine-tomachine (M2M) domain?
Inderjeet Singh: The greatest M2M challenges are associated with increasing scale on the one hand and granularity on the other. Particularly, the sheer number and range of devices, the volume of real-time events spawned by these devices, and the resulting number of composite services created through vast levels of instrumentation. Driving a contextual, real-time, actionable response to this scale of instrumentation remains the fundamental challenge within M2M environments.
M2M Now: In the past, developments in M2M have often focused on enhancing the capabilities of devices and platforms. But with the spread of services to new industry verticals, there seems to be a torrent of data that needs to be stored. It has to be analysed and acted on for M2M to be optimised. So how is TIBCO meeting this need?
IS: In addressing your question, let me give you some background on TIBCO’s vision with respect to ‘optimisation’, not only to the world of M2M, but to all real-time domains. Our focus and capability are in TIBCO’s contention that having all the information after a critical event has happened, has much less value than having correlated, actionable information before the critical event has happened, when it can be used to affect the outcome of a critical event rather than simply to plan better for the next event.
An ability to detect, understand, and react to patterns of critical events in real time drives our vision of operational intelligence or what we call ‘situational awareness’. By making sense of these ‘event-rich’ situations, we are able to provide an in-flight set of actionable directives either to a pair of eyeballs, another application, or an orchestrated process. This level of situational awareness, when brought to the world of M2M, provides the inherently missing context to deliver true value-added services to consumers across the spectrum of business.
TIBCO’s legacy, as you probably already know, has been as a real-time, integration company, connecting our customers’ mission-critical systems together. In the past five years, we have taken this capability one step further and have created an enterprise ‘event-enabled platform’. In the M2M world, this is yet another iteration of need, a crying out for our event-enabled platform.
We consume very large volumes of events, enrich these elements with other data sources, and manage the resulting ‘situations’ in an in-memory data grid. Unlike a normal application server, this in-memory grid, TIBCO’s event server, can apply methods, evaluate patterns, analyse data, and process directives or decision routines all in real time. Naturally, we couple our event server with our ultra high-performance enterprise service bus (ESB) to accommodate the volumes of events emanating across the M2M domain.
As we apply this architecture to the M2M world, we essentially connect devices over an ESB with billions of events exposed, managed, and administered within an in-memory data grid by master data management services. These data management services reconcile information, for example, device and customer information, location information, or delta change information, inside the data grid in real time. The events server applies dynamic, rule-based complex event processing (CEP) technology to impart understanding and correlate these events based on patterns of instrumentation. This ultimately creates actionable domain and business intelligence at the speed of events themselves, in real time.
Interestingly now, actionable events can be consumed and utilised by social technology environments – such as our own social platform called tibbr. In tibbr, users can subscribe to events in real time and also to ‘pattern-based’ subjects that track and analyse business assets, activities, or processes, using our active business process management (BPM) engines.
To close the ‘virtuous loop’ on leveraging the M2M environment, TIBCO’s interactive analytics visualisation tool set, Spotfire, allows business users to take large volumes of information and create interactive dashboards reflecting actionable intelligence within the high granularity of M2M environments. With visual analytics tools like Spotfire, the business user can get direct insight into the data created from complex infrastructure software without having to constantly gain intelligence through reports from the technical side of the business.
M2M Now: Have you got any use case examples?
IS: The Smart Energy environment is an area where TIBCO has been very active over the last few years. A number of global utilities combine the aspects of realtime M2M to give a picture of energy assets (distribution stations, feed profiles, energised and nonenergised devices) to network engineers, technicians, and those responsible for energy grid stability, resilience, and intelligent demand-response behaviours.
Now, with mobile activated, this has a huge impact across large geographic grid footprints, dynamic energy management processes and highly organisational infrastructures. We are now leveraging the same opportunities in the remote medicine and therapy fields where again, large number of diverse connected devices are ‘singing out’. These connected devices create real-time critical healthcare measurements with the concomitant requirement for real-time intelligent evaluation and decision-making on behalf of patients by their clinicians and primary care givers.
Now, the same TIBCO platform applies to other nonclassic M2M verticals. Let’s take an example in retail. You walk into a store, swipe your card and within milliseconds, the TIBCO event server receives a message. The event server knows who you are (from the pattern profile that’s been stored in the data grid): your behaviour, your age, your spouse, and so on. Naturally, this process only occurs on an opt-in basis. The server matches your profile information against the current business environment such as: what goods are sold fast, what goods are slow moving, inventory at respective stores by location, the current weather, and your purchase history against these dynamic patterns. The server appends bigger markers to this information such as: what life events are coming up – your wife or kid’s birthday, a vacation you are planning, or a graduation that’s impending.
Based on all these factors, marketing campaign rules kick in, figuring out the next best offer that has the highest propensity for success and can be intelligently delivered across the mobile device cloud. The net value of this for retailers is the ability to increase the response rate, a tremendously transformative IT operation, and increase business benefits for not only customers in terms of net sales, but also in more efficient optimisation across the entire supply chain.
M2M Now: Much of this data will be useless then, if it’s not dealt with immediately. How can M2M service providers ensure that their data delivery to clients and partners is timely and relevant?
IS: TIBCO’s platform is absolutely critical; the ability to change, to understand information in real time, to react in real time. We believe the social environment – like Facebook, and Twitter – have led to a realisation that disseminating this highly instrumented information through social media might actually be the best way to provide choice and decision advantage to a consumer base.
This is the province of tibbr, our enterprise social networking platform. With tibbr, you can not only monitor people-based situations, but also subjects that can be defined across a portfolio of devices defined in the M2M domain. tibbr can be run in a private or a public environment; you can secure it yourself, disseminating the right information to the right person, in the right context, in the channel you prefer.
M2M Now: We’ve talked about the platform a bit. TIBCO also talks about deriving context from M2M. Can you explain?
IS: Information today consists of what happened and where it happened. If you know the customer’s birthday is next week, or his anniversary is next month, that context is extremely important in regards to his purchasing decisions. If you were simply looking at an event in isolation you would not have this information. When you look at an event, surrounding information (historical, real-time, or enriched information from other sources) drives context and intelligent conversation with the consumer, which is very hard to do without context.
M2M Now: So, in the real world, it’s about how you turn raw data into actionable information?
IS: That’s right, yes. Filing a customer’s birthday or anniversary means nothing if you don’t use it in combination with purchasing history to make a very appealing offer and drive a sale.
M2M Now: With a growing number of partners in every M2M ecosystem, what steps should service providers take to ensure that everyone in their value chain can analyse service regarding customer and operational data?
IS: Service providers are usually reluctant to share this information, often for reasons beyond their control. We help service providers publish the right information over the right channel. That channel could be an application on their Android phone or iPhone, a web channel, or a social channel (like a user’s Facebook wall).
The wrong way to approach this is to have a one-sizefits-all strategy in which the telco provider can say, “Look, this is exactly what I’m going to do, and that’s it.” They really want to give control to their customers, who can tell the company, “This is the way that I want you to interact with me and this is the data that I want you to feed me.” Rather than being hesitant to give customers any control, telco providers will now have data on which channels individual customers will be most receptive, a hugely important piece of information based on personal preference.
M2M Now: And where in TIBCO’s experience is it being done well?
IS: Some telcos are in the early stages of doing this process well, some in the US, and some in EMEA. The TIBCO customers who realise social interactions with customers are valuable are the ones that are going to leapfrog the competition.
Simply looking at the provider and customer relationship in a one-channel or two-channel fashion won’t cut it. Consumers are looking for a social way of interacting with their services, which in the future will drive customer satisfaction, cross-selling, up-selling, and providing personalised offers with a very high probability of success.
Inderjeet Singh is senior vice president and general manager, Products and Technology at TIBCO Software, and responsible for engineering all TIBCO Infrastructure Products.
M2M Now: And lastly, how can TIBCO help in making M2M service data actionable, not just for the service providers and partners, but for corporate end users? And can you give examples of how this is being done?
IS: Extracting intelligence from this massive aggregation of information is important. For example, you might get hourly updates from a million devices, when what you really need is a spend analysis for a particular zip code in a particular period of time. That is something that a corporate user might like to pursue in assessing geographic metrics of success of different business approaches.
A visual analytics platform like Spotfire (which understands events through a large in-memory cache), becomes absolutely critical, as it can aggregate all those individual hourly updates and present patterns and trends in a visual fashion directly to the business user.
Here’s where it comes full circle. Extraction of intelligence from immense data streams by a corporate end user will result in additional rules, creating additional intelligence that will be fed back into the rules-based platform. For example, if a certain zip code has higher than average data usage on weekends, then a created rule can automatically re-route capacity accordingly. You don’t even have to know that weekend data usage is high in this zip code because it’s a tourist hotspot. You only need to know that capacity must be scaled up on weekends and down during the week, without extensive knowledge of the demographics of each zip code.
This gives you much better intelligence without going through any new IT projects or redeploying all of these logical software relationships. Simply have IT implement this new rule and dynamically change the behaviour of the application and your whole infrastructure.