Anatomy of an S+s line Of Business Application: Eugenio Pace

Download 15.2 Kb.
NameAnatomy of an S+s line Of Business Application: Eugenio Pace
A typeDocumentation
Anatomy of an S+S Line Of Business Application: Eugenio Pace

Eugenio Pace, Senior Technical Evangelist, led this session on the building of software and services. He encouraged participants to feel free to provide their questions and comments.
Eugenio Pace: I work in the architecture strategy team. You might have found our content when searching for software services and other places. Our portal is … We’ve been published for about two years now. Today, context where we’ve come from, combination of factors and relationships in S+S world. What are the challenges our partners and customers face? Feel free to interrupt.
Deep on one aspect of this. Building aspects of S+S. Vast space. Affects everything. Everything is software, everything is services. Need to focus down to the problem and differentiator between building software and what’s different now. I’ll leave you with pointers of our content and what’s coming in the future. We’d like feedback to guide our future content.
One possible world of participants and relationships. Two companies creating software. Business to develop and sell software. Independent software vendor sells to customers. Consumer spectrum to big Enterprises. In the middle, small and medium customers. Individual customers don’t want to pay for software. Other hand, Enterprises spend big bucks on software. In the middle, to the right of the network is the host. North wing hosting, a host. Relevant role. Provides infrastructure, networking. Added complexity in market places becoming more frequently found. Combination of hosting abilities. Bundling into richer offers. Finally, building block services and foundational element. Building blocks leverage services. Email, application enhancement services. Did this to identify main actors. This allows us to do what we’re here for, identifying architectural challenges, coming up with proven solutions. Both architectural and technological challenges. Many instances already. Four different areas to work on. We’d like your feedback if these are the right buckets.
Building, which is today’s topic. Building software that is necessary. How to run it. Composition challenges. Finally consumption, using, acquiring, running. Does not mean that we’re not worried about … that a big company doesn’t act as hoster and ISP. More about identifying challenges. Content for ISPs relevant to Enterprises. We’ll go deeper into what an application like this looks like. Most of us have used software that is a service. Today, we’ll look at what’s inside from the architecture perspective.
Our team. We write a lot. We recognize reading and translating or narrating is tough. Need to consume many things. We’ve switched not only from written content, to … We believe in the power of code. Creating reference applications. LitwareHR. What does it mean to build an S+S? Any application? We’ve been building for awhile. Focusing on what’s different. Security, creating productive development team. Many scenarios. When we look at successful companies, people who have attempted to do that, characteristics highlighted. What makes this environment tougher? Things you might encounter if you move into S+S? Three main areas which are quite extensive.
Designing backend as a service. Top in priorities. Multi tendency and isolation. Notion of application benefiting. Performance. On premise, software is scalable. If you sell scenario, with each customer having ten uses, many scenarios. Each customer has own server. Your application might be designed to handle ten uses and you’re fine. Move to an S+S world. Customers without any software installed. Hoster. Application designed to have ten customer usages, has 1,000. Big challenge. How to handle this load and scalability requirements? Multitenancy associated with application density. Tapping the economies of scale. Distribute among range of customers. I have 100 customers, ten users each. I can buy 100 servers. No incentive to do. Can’t distribute cost. Multi tenancy, maximize servers. Goes against isolation. Data sets, databases, codes naturally isolated due to physical boundary. Mixing of data. How to isolate.
Related to that, customization and notion of metadata and templatization. Ability to handle differences between customers without needing to rewrite. Enabling self-service. Consumer applications. Limited. Can’t change domain. Same thing here. How can same data and code base handle requirements allowing each customer to tweak as they want? Related to this is notion of two sets of services. Functional services. Things software supposed to be doing. HR system. Functional services related to recruitment and other HR processes. From design perspective, other applications. Exposing application and what its doing inside. Translates to set of web services. Submit resume, open position. Unfunctional would be description of entities. From entities to different components of applications. Enables consumption of service in scalable way. No need to change code each time.
Second aspect is notion of rich content. Associate fast applications with web applications. Not true. Limiting point of view. We believe new monitor of software and services. Need to pick the right UI for the right task. Critical. If targeting consumer market, user experience is critical. Within applications, if not appealing, people will view as too complex. Important aspect where you should invest. We build this application from end to end.
Final aspect, allows specialization. People without right SKUs to take advantage of things difficult to do on own. LitwareHR. Would like to focus energy on what they know the best. They don’t want to specialize in anything else. Need to take advantage of infrastructure on cloud. Allows people to be experts. Specialization that leads to optimization. Nobody buys an application because of its logging system. Not a differentiator. I buy HR systems. Need them to focus. LitwareHR is functionally incomplete. Microsoft isn’t in the business of HR systems. Design patterns appropriate to challenges. LitwareHR manages whole lifecycle of HR recruitment. Hosted somewhere. Same code base, web pages, client, database, all shared between customers. Whoever buys it, they share everything. We’ve optimized to maximizing density. Others will have isolation requirements. Maximizing density is tough. LitwareHR has moved ahead to different areas. Public website. Smaller customers use. Secure web client. Used to change look and feel. For more advanced users, a Smart Client. Interacts with LitwareHR services. Offline.
Finally, LitwareHR a system providing feature around video. Upload video resume. Don’t want to be media or streaming experts. Enhanced LitwareHR to use Silverlight streaming. Microsoft makes it available for everyone. Scalable, secure. Each company has own account in Silverlight.
Extending schema. Many ways to do. Obvious way. Language field. S+S has five different attributes. Will permeate rest of application. Effort of building something that works like this is quite high. Smart client. Functional services. Software downloaded and installed on network customer desktop. Guidance. Large amount of guidance available. We focused on regular smart client and the smart client aware of … want code base to be used by everyone. Key things to show with Smart Client. Ability to operate offline. Providing rich user experience. Able to handle bulk data, moving massive amounts. Biggest values. Synchronization. Schema changed, reflects to smart client as well. Paradox. Irony of smart clients that look and behave like web clients but work on both. Need to close then open application. Reminds of refreshing experience. Smart clients shouldn’t behave like web client. Same code base. Uploads immediately. Video is pointer to Silverlight content. Upload, task launched to upload, page content, in Silverlight content. Mega data created. Video content uploaded via Silverlight streaming. Smart client integrates experience. Richer experience.
Different question is that this is extensible, but how does it perform in real world. Proof. Data to back our claims in guidance package which is a performance and scalability test on database extensibility. For each table in application, fixed columns with changed meanings. Meanings depends on who you are. Might translate to description. Second way, is on top. Table extension. Implemented this in LitwareHR. One table using another to have extra fields and description there. That one is most flexible of all. As many fields as I want. Trade-off’s. XML data types. One column with XML types that changes so schema will change as requested. Extension fields, amount of information, providing customer with results of our tests. Publishing scripts and tests and data generators together with methodology that we follow so you can reproduce these tests. Content you can reuse with simple tweaks. Same report, same information. Theory behind methods in white paper we publish. Will be available soon.
You’ve been shown this many times. CRM Live. Midware. How CRM Live is different, under the hood.
Participant: Log on, credentials, Microsoft standardizing, authentication.
Eugenio Pace: Are we using our own method for authentication? STX in this particular example. I’m not a security expert. I don’t feel comfortable saying this is 100 percent standard basis. But if its not, it’s really close. We try to provide elements so you can use easily, as much as possible. Good principle in general. Well known, replaceable, you can use your own. Providing service. I’m not sure of status.
Microsoft Employee: Leverage … Windows Live … AD … authenticate … web based … coming future, AD … provide space … consumer space … credential on your own environment…vast entity … Enterprise … authenticate after … virtualization aspect …
Eugenio Pace: This represents building side of things. We are working on rest, what it means for architect of a hoster. Things between ISP and hoster. Today, 1:1 relationship, building infrastructure to support. We believe can be better than that. Content around hosting and building service-delivered platform. Working on marketplace aspects to create rich bundle of applications, leveraging common components. Consumer aspect, related to your question. Typical challenge in an Enterprise. Single sign on between my own infrastructure. Let HR logon or find way to provide single sign on. What’s required. Also consumption aspect. How to integrate software that is a service in an operational infrastructure. When something fails I know it. I can run through support processes. LitwareHR, no services. Across the boundaries of my company. I would like tat capability to be from a management perspective. How to do? We are working on content. Stay tuned.

Share in:


Anatomy of an S+s line Of Business Application: Eugenio Pace iconAnatomy of an Android Application

Anatomy of an S+s line Of Business Application: Eugenio Pace iconRequest for Proposal for a new Line of Business Solution

Anatomy of an S+s line Of Business Application: Eugenio Pace iconRequest for Proposal for a new Line of Business Solution

Anatomy of an S+s line Of Business Application: Eugenio Pace iconFigure. Each column represents a month. Each dark bar represents...

Anatomy of an S+s line Of Business Application: Eugenio Pace iconThe characters // denote my comments or explanations of what a line...

Anatomy of an S+s line Of Business Application: Eugenio Pace iconResearch demonstrates that effective flexibility can benefit an organization’s...

Anatomy of an S+s line Of Business Application: Eugenio Pace iconAbstract -this research evaluates product-line software engineering...

Anatomy of an S+s line Of Business Application: Eugenio Pace iconPrimary Instructional Mode: On-Line Asynchronous Canvas: On-line

Anatomy of an S+s line Of Business Application: Eugenio Pace iconThis transaction will be used by aafes to report application errors...

Anatomy of an S+s line Of Business Application: Eugenio Pace iconSpeech Mr. Dustin Pace

forms and shapes

When copying material provide a link © 2017