A Nonprofit Case Study: OpenShift, JBoss, and Amazon Web Services

In the real estate industry, buyers and sellers are constantly researching listing property information such as status, price, and location. A Multiple Listing Service (MLS) is a localized database which association members, salespersons and broker salespersons, access to update their clients with relative property information. However, this process has become slow and outdated because of the Internet and the Web.

The Global Real Estate And Technology Consortium (GR8C) is a scientific nonprofit located in Orlando, Florida. Its mission is to promote the education, operation, and use of the World Wide Multiple Listing Service (WWMLS) to its fullest potential by developing protocols, specifications, and standards for the benefit of all people throughout the world.

This case study will provide a low-level look at the Multiple Listing Service Network Protocol (MLSN), a new request-response messaging protocol developed by the GR8C for querying property information on user reserved port 32801. The session will discuss how OpenShift is used to build, deploy, and monitor configured JBoss Application Servers as MLSN Servers to Amazon’s Web Services Platform. In addition, the presenter will share key lessons learned and technical issues dealing with OpenShift. Finally, an MLSN client to MLSN server messaging demonstration will be provided.

–Corey

Title: MLSN: A Request and Response Internet Protocol for Real Estate Property Listings

Abstract:

Sellers list their properties with real estate professionals who enter property information into private databases called Multiple Listing Services (MLS). These MLS’s attempt to store and index data listings with serialized numbers that only members of the local real estate association have access to query, insert, and update. In the United States alone, there are over 900 MLS’s which unfortunately create redundant, localized MLS numbers causing confusion over specific properties and their respective listing information on the Internet. These duplicated MLS numbers can be anywhere from five to ten digits long or more relaying no meaningful or geographical information to real estate professionals, sellers, and buyers alike.

This paper proposes and discusses a new internet protocol called Multiple Listing Service Network (MLSN) Protocol for registering, assigning, and distributing property listings on the Internet. Similar to the WHOIS protocol, MLSN is an application publicly available on the Internet, however instead of port 43, MLSN is accessible from reserved port 32801. The numbering scheme is comprised of two parts separated by an ‘@’ symbol: (1) a triple-dotted notation of three decimal integers ranging from 1 to 65,535 and (2) a double-dotted notation number composed of a postal code and country code standardized by ISO 3166-1. Taking in consideration for expired listings, MLSN offers a reusable feature which allows assigning a new, double-dotted notation number to a preexisting triple-dotted number thereby creating a new MLSN property listing number. Together with a standard protocol and unique numbering schema, MLSN provides a uniform resource locator (URL) to users for conducting a more efficient and convenient real estate property searching on the Internet.

Keywords: internet, multiple listing service, protocol, internationalization

–Corey

Abstract

Sellers list their properties with real estate professionals who enter property information into private databases called Multiple Listing Services (MLS). These MLSs attempt to index listings with serialized numbers that only members of the local real estate association have access to query, insert, and update. In the United States alone, there are over 900 MLSs which unfortunately create redundant, localized MLS numbers causing confusion over specific properties and their respective listing information on the Internet. These duplicated MLS numbers can be anywhere from eight digits or more relaying no meaningful or geographical information to real estate professionals and real estate buyers alike.

In this paper I introduce a public, standardized numbering system and protocol called Multiple Listing Service Network (MLSN). Similar to the Web, MLSN is an application publicly available on the Internet, but instead of port 80, MLSN is accessible from port 32801. The numbering scheme is comprised of two parts separated by an ‘@’ symbol: (1) a triple-dotted notation of three decimal integers ranging from 1 to 65,535 and (2) a double-dotted notation number composed of a postal code and country code defined by ISO 3166-1. Taking in consideration for expired listings, MLSN numbers have a reusable feature which allows assigning a new, double-dotted notation number to a preexisting triple-dotted number thereby creating a new MLSN number. Together with a standard protocol and unique numbering schema, a Uniform Resource Locator (URL) is available to users for searching for property listings on the Internet.

–Corey

One of the challenges for buyers in real estate is knowing the most current listing information. The MLSN Protocol resolves this issue for public access to listing information, but it is based upon a request/response protocol. A user must manually query an mlsn server to find out the latest increase or decrease price for a listing. Whereas, the publish/subscribe could be used to receive listing information automatically over the Internet.

In the diagram above, users subscribe to topic message queues using the PubSubListing Protocol. When a listing’s price changes, this update is sent to the messaging server which updates the subscribed topic queues, respectively. The pricing update is broadcasted over the PubSubListing Protocol to the users who then view the new price on their laptop, desktop machine, or mobile device. This broadcast pattern continues whenever the listing’s price changes. Users have the ability to continue to receive future broadcasted messages by remaining subscribed or unsubscribing to the topic queues.

In summary, the obvious benefits of the PubSubListing protocol are the automated messages of changes and the ability to subscribe or unsubscribe to topic queues. The next step for the protocol is to apply for a reserved port with IANA.

One of the major benefits of MLSN URL’s is the ability to reuse assigned MLS Numbers. Using an simple scenario, I’ll go in detail of how reusability works with MLSN URL’s.

In the diagram above, Alice is an owner of an accredited OpenMLS registrar. Her service is located at alicesmls.com.

Bob the seller, needs to sell his house in Orlando, Florida. He registers it on Alice’s MLS for 60 days and receives a unique MLSN URL, mlsn://1.1.1@3203.840.

Unfortunately, Bob’s house does not sell within the 60 days and expires. When this happens, the assigned 1.1.1 or triple short number returns to Alice’s block of MLS numbers to be reused again.

A new seller, Cristy, registers her house in Atlanta, Georgia on alicesmls.com and receives her unique MLSN URL, mlsn://1.1.1@30329.840.

As you can see, 1.1.1 was used once with Bob’s listing and then again with Cristy’s. Each MLSN URL represents two completely different listings, which demonstrates reusability.

By the way, the number 840 in the MLSN URL’s represents the iso country code for the United States.

–Corey

A Simple User Session

Scenario: User wants to know the status of a property listing in the downtown Orlando zip code, 32801, located in the United States.

A user starts a session on a client desktop. He enters the following MLS Number URL at the prompt:

user>mlsn 1.1.1@32801.840/status

The mlsn client application sends the MLSN URL via port 32801 over tcp through the internet to an MLSN Messaging Server which accepts the text string via port 32801.

The MLSN Server routes the URL text string to an application server which parses and creates a query for the database which returns a result.

The result is sent from the application server back to the MLSN Messaging Server which sends the result via port 32801 over tcp back through the Internet cloud.

Finally, the client desktop receives the result via port 32801 and displays the results to the user:

user>active

–Corey

One of the problems in the Real Estate Industry revolves around the numbers assigned to listings by Multiple Listing Services. Six, seven, eight, or more numerical digits may be assigned to an individual MLS listing number along with possible alphanumeric characters preceding the number.

From my research, listed below are ten randomly selected, real world MLS numbers I extracted from a daily Google Alert listening for text string ‘MLS#’:

• MLS# 829710
• MLS# 90035087
• MLS# 4240855
• MLS# H740907
• MLS# 9088503
• MLS# 100006059
• MLS# M1366719
• MLS# MC5555362
• MLS# H741473
• MLS# 4334588

Lacking Visual Meaning

The problem with these numbers and others is that they have no visual value or meaning to a user. MLS numbers are stored and generated within datatbases by MLS providers. Since these numbers are required to be unique, a field type of serial is assigned to the field which increments the next unique MLS number by 1 to enforce a primary key’s uniqueness. However, this system does not benefit the user from deciphering where the listing is located nor any other details.

Duplication Misery

Another problem is that there are hundreds of MLS’s in existence and the probability of the same number assigned to two different listings in two different MLS databases is highly probably. If a user only has the number to reference, which listing number is which? And how much time is lost trying to figure out which listing corresponds to which MLS database?

Zero Reusability

Since MLS numbers are simply incremented by 1, the opportunity for number reuse is eliminated. A provider’s MLS numbers will continue to increase by 1 which will increase the number of digits of the MLS numbers over time. Having no delimiters, MLS numbers become humanly impossible to remember since the average human is said to only have the capacity to remember up to 7 consecutive digits. But looking at the random MLS numbers I collected above , a high percentage of real world MLS Numbers either equal or are greater than 7 digits.

The Solution

The Scheme

The MLSN URL scheme is used to designate Internet MLS listings accessible using MLSN (Multiple Listing Service Network). The scheme adheres to IETF’s RFC1738 which describes syntax for strings representations or URLs via the Internet.

The Registrar Assigned Number

After a user posts a listing to a registrar, a 3-octet number is assigned to a listing just like an IP address is statically assigned to a machine or device. But in the case of the MLSN protocol, 3-octet numbers are dynamically assigned to allow for reusability. After a listing expires, the 3-octet number returns to the registrar’s pool to be later reassigned to alternate user’s listing.

For example:

User registers listing with AlicesMLS and receives the following MLS Number URL Address:

mlsn://1.1.21@32821.840

The listing expires after 30 days and 1.1.21 returns to the registrar’s pool of mls numbers and is reassigned to another user’s listing:

mlsn://1.1.21@32819.840

The new MLS URL Address reuses 1.1.21, but is a different listing from the previous listing.

Note: 1.x.x is designated as a reserved Class A block of numbers and used for testing purposes.

The Postal Code

By adding a local postal code to the URL, the listing becomes visually meaningful. Just by looking at the URL, a user can tell if the MLS listing is located in an ideal or non-ideal postal code area. In addition, postal codes are used by the majority of the world’s countries which makes MLSN URL’s extensible to potentially cover any property in the world.

The Country Code

Keeping the global MLS market in mind, a 3-digit country code follows the postal code. Every country is designated with an ISO 3166-1 3-digit number.

This number enhances an MLS Number URL for reusability and visual meaning when coupled with the postal code.

The Port

Two weeks ago, I secured internet port 32801 with IANA on its list of registered ports.

mlsn            32801/tcp   Multiple Listing Service Network
mlsn            32801/udp   Multiple Listing Service Network
#                           Corey Leong  27 January 2010

The purpose of this new port will be to send and receive request and response messages from client machines to designated mlsn servers. This type of messaging mirrors the whois service which enables a user to lookup domain names over port 53.

The Resource

Using a RESTful approach, a resource appended to the URL string permits lookup of a particular attribute of the listing.

For example:

A client sends the following request to an mlsn server.

mlsn://1.1.17@32802.840/status

The mlsn server returns “active” to the client, then the connection is severed.

Other potential resource lookup scenarios include beds, baths, square footage, price, etc.

Conclusion

The above information is a high-level perspective of the Multiple Listing Service Network (MLSN) Protocol, URL, and Port. MLSN is in development mode. Eventually, MLSN will be used by any user to query MLS Listings from registrars over the Internet.

I intend to continue testing a multitude MLS Number URL strings until this protocol reaches full 1.0 version.

–Corey

This past weekend I submitted our online application for a new user registered port with IANA. Here is a simple client server model depicting requests and responses.

The protocol, Multiple Listing Service Network (MLSN), will involve online users and applications querying a real estate database of listings.

I’ll post more diagrams and information regarding this protocol/port as we draw closer to our application being approved.

–Corey