One of the technology issues in this case is that some email archives maintain many non-overlapping, atomic search indexes. Often times, one or more of these indexes will become corrupt or otherwise unavailable. When this happens, data handled by that index may simply not be returned as part of an E-Discovery search. Organizations engaged in E-Discovery may now wish to request an index corruption log listing inaccessible search indexes as well as the email covered by those indexes as part of their E-Discovery protocol.

Another question is whether the email metadata was accessible via another method such as a SQL query search. Many email archives use different technologies for fulltext indexes (e.g. Lucene, dtSearch, Alta Vista, etc.) and for relational data (e.g. MS SQL, Oracle RDBMS, etc.). Even if the fulltext index failed, a SQL query may be able to show how many emails were in the archive during a certain time period and associated with certain custodians and subjects.

There is a limited amount of information available in this case; however, the players and technology issues make it with studying. A brief timeline of this case has been assembled based on the available, but incomplete, public information for a quick overview of the case.

SpringCM is adding customers and winning acclaim (KMWorld’s Trend-Setting Products of 2008); however, SaaS ECM is still a very minor part of the overall industry and major players are entering the market. Gartner estimates SaaS and open source solutions combined account for less than 5% of the software license revenue for ECM. One reason for the slow penetration is the need to integrate ECM solutions into existing infrastructure which can be difficult with SaaS. However, with Google, Microsoft, IBM, EMC, and others entering the market, SaaS ECM may get some legs. For now, SpringCM will continue to differentiate itself through departmental and industry vertical solutions.

SpringCM just released version 4.5 with some new features, notably in the process mangement and security areas. To get a feel for their product you can sign up for a live demo or watch a number of their canned screencast demos available on their website.

Overheard:

• Dan Carmel, CEO
  • “As economic conditions have deteriorated, companies are challenged with improving business productivity without sacrificing compliance. Software-as-a-Service is a proven model for these times because it enables companies to deploy software solutions faster and at dramatically lower cost than traditional, installed software.” (Source)
• CMSWatch ECM Suites Report 2009
  • “If you were to believe the marketing hype that surrounds SpringCM, you would start to believe that the likes of EMC and IBM were quaking in their boots…” (Source)

Website: www.springcm.com

Management: Dan Carmel, CEO

Funding: $20M (latest: $10M, April 2008)

Investors: North Bridge Venture Partners, Foundation Capital

Guidance Software is a top ranked E-Discovery company in the 2008 Socha-Gelbmann Survey (excerpt below); however, this survey comes with a heavy disclaimer. Additionally, a competitor, AccessData, is now run by many ex-Guidance staff.

Although there is much speculation that the E-Discovery market may survive the current economic downturn, this announcement has had little effect on the Guidance stock (chart courtesy of Yahoo!):

There are several interesting features with this product. First, the data is crawled and the search index is created for initial analysis without the need to first collect the ESI. The ESI is collected and placed on hold only after the initial assessment. Secondly, individual messages can be extracted from container files, e.g. individual email messages can be extracted from PST files. While many e-mail archive vendors are working with organizations to eliminate PST files, this product can search and collect individual files within PSTs.

Notably, this product does not appear to handle Legal Hold Notification Management as offered by exterro, Autonomy, and now Guidance Software.

There are some open questions:

• Early Case Assessment: Andrew writes that “Counsel identifies the custodians (that is, the people that have potentially relevant data in their possession), data sources, and date ranges. The Insite software connects to the data sources, scans and indexes files that match the search criteria, and brings back the information in HTML format.” This raises two questions:
  1. How ideal is it to narrow the ESI to custodians, data sources, and date ranges? Recommind mentions there’s no need to copy data to a central repository; however, central repositories can be extremely scalable and allow Early Case Assessment without the limitations of custodians and date ranges. Combined with fast search capabilities, this can provide more comprehensive Early Case Assessment. Can InSite also work with large amounts of data without the need to restrict data before it is searched?
  2. Is there an option to bring back data in native format?
• Structured Data: The product page mentions an ability to capture structured data; however, the data sources listed are limited to “laptops/desktops, file share servers, email servers, email archives, enterprise document management systems such as Livelink™, and Microsoft SharePoint™.” Typically, these repositories hold unstructured data. Structured data is normally applied to database systems, ERP, CRM, etc. What structured data can InSite handle?
• Third-Party Repositories: Andrew writes that InSite will move the collect the content and move it to third-party repositories. Which back-end repositories are supported?

The product looks interesting for reducing collection costs, but it is also behind in certain areas. Other products in this space can perform remote forensics to locate deleted files and institute legal hold on disconnected laptops (presumably when the user does not have administrator privleges). I’m interested to hear how comprehensive the Early Case Assessment and structured data capabilities are.

Website: recommind.com
Management: Robert Tennant, CEO
Funding: $7.5M (Latest: $7.5M, October 2008)
Investors: Kennet Partners

There have been a number of recently reported customer complaints with regards to EAS support with some customers recommending getting support from resellers instead of Autonomy. With the closure of the EAS office, this may accelerate. Notably, Gartner mentions the following in the 2007 Magic Quadrant for Information Access Technology:

• Autonomy still has a deserved reputation for complexity and sophistication, and it lacks the passion to simplify the product for unsophisticated administrators who may nevertheless have strategic needs to fulfill….
• Autonomy must improve its ability to support existing customers when they face
  challenges in strategic implementations.

Discussion on AXS-One is available at InvestorVillage. Recent posts there and on Yahoo! Finance have not been positive and some are looking for AXS-One to file for bankruptcy.

Notably, they announced their first contract for their Dynamic Data Migrator product which they announced back in June of this year.

1. Functionality: this product is focused on the left hand side of the EDRM model
   • Search, culling, legal hold, export, and audit trail capabilities are included
   • Review, analysis, and tagging are not included
2. Data Source coverage:
   • E-mail ingestion from Microsoft Exchange, Lotus Domino, PST files, NSF files
   • ECM integration with Microsoft SharePoint, EMC Documentum, and OpenText
   • Integration with email archives via APIs
3. Required Components: eDiscovery Manager requires one of the following IBM ECM products:
   • IBM DB2® Content Manager and IBM CommonStore
   • IBM FileNet® P8 and IBM FileNet Email Manager

The ECM integration makes sense since IBM has the base integration capabilities through IBM WebSphere Information Integrator Content Edition, a product from the 2004 Venetica acquisition. Some have highlighted the issue that while IBM’s product can handle third-party ECM products, it cannot natively handle third-party e-mail archive products. This is especially important as many organizations have moved long term archival of e-mail to e-mail archiving solutions, and not their production mail servers. Customers with third-party e-mail archiving solutions will likely wish to use the eDiscovery search, culling, and legal hold features that come in those products instead of moving large amounts of e-mail to an intermediate IBM archiving solution. Notably, Oracle Universal Records Management (URM), a product from the $440M 2006 Stellent acquisition, claims to handle Symantec Enterprise Vault in addition to SharePoint, Documentum, OpenText, etc. The ECM integration space for E-Discovery will likely see more activity as CMIS takes hold and ECM solutions start providing a common interface.

Spokespersons for IBM eDiscovery Manager include:

• Ken Bisconti, vice president, product and strategy, IBM Enterprise Content Management
  • “The explosion of electronic content presents challenges for organizations to retain and produce information efficiently and accurately when needed,” said Ken Bisconti, vice president, product and strategy, IBM Enterprise Content Management. “Our eDiscovery offerings enable customers to have insight into knowing what information exists, where it is stored, how long it must be kept and how to locate it.” (source)
• Aaron Brown, program director for IBM’s content discovery segment
  • “Out of the box today, we can’t pull e-mail out of a [third-party] archive.” (source)

Website: www.ibm.com … ediscovery.html

Gartner MarketScope Rating: Caution

Market Cap: $111B

From reading the press release and some articles, the capabilities appear to be:

1. VMware support
   • Support is mentioned for ESX and ACE (on XP and Vista). Support for ESXi, Server, Workstation, and Fusion are not mentioned. VMware management products are not supported.
   • It is not mentioned if Kazeon can handle VMware images that are on, paused, and/or off. One may assume all are supported, but it would be nice if this was made explicit. Additional information on whether Kazeon needs to turn on or off images would also be useful information.
   • Additionally, although the software can find VMs on servers, laptops and desktops are mentioned as the primary targets. It would be interesting to know if using this software is sub-optimal for file servers, e.g. if it causes a high amount of server load which could slow performance.
   • VMware forensics would be an interesting feature for locating locating deleted files on VMware images. This is not mentioned and likely not included.
2. Preservation / Legal hold
   • The software claims the ability to enforce legal holds on data in a VM. Andrew Conry-Murray reports “The software can enforce legal holds on information stored in a virtual machine, including making it read-only for the local user, or denying end user access to that information entirely.” A requirement for this to work may be the presence of administrator privileges on the OS, granting of such privileges to Kazeon, and the lack of such privileges assigned to the custodians.
3. Collection
   • Kazeon claims the ability to collect data and files from the VMs to a central repository without changing meta data, including access and modify times.
4. “Unknown or hidden virtual instances”
   • The press release makes a lot of mention of this; however, there is no mention of what they mean by this. This appears to be marketing hype for VMs that the organization isn’t aware that their employees have installed. However, people don’t usually say Microsoft Office documents are “unknown or hidden” simply because they are saved on a local hard disk without updating a central repository. Until there’s some clarification, it seems safe to assume this is marketing hype and the VM instances and images are not actually hidden, say using root kit type technologies to hide it from the file system and process monitoring system.

Overall this sounds like a very interesting feature and there is definitely a need to identify, preserve, and collect data on VMs. I look forward to learning more about it.

Website: kazeon.com

Pricing: Kazeon Information server starting from $4.30/GB (Source)

Management: Muddu Sudhakar, CEO

Funding: $44M (Latest: $21M, August 2006)

Investors: Redpoint Ventures, Clearstone Venture Partners, Goldman Sachs, Focus Ventures

Clearwell has posted an informative Flash demo which can be reached from their homepage which is worth watching. PlanetEIM has taken some screenshots and put together short overview.

In discussing this new feature, Kamal Shah, Clearwell vice president of marketing, gave an interesting quote to Dave Raffo of SearchStorage.com indicating web search needs to scale but the same requirement may not be necessary for enterprise search or e-discovery. He says,

Saying one approach can be applied by all is like saying the same person can win the 100 meters and marathon in the Olympics. They have different objectives and business drivers. For enterprise search, it’s speed and simplicity. For the Web, it’s speed, simplicity and scalability. For e-discovery, it’s defensibility – can you defend best practices and show it in court and minimize nonrelevant results because of cost per results?

With enterprises facing many terabytes of ESI and billions of records, scalability requirements for enterprise search and e-discovery are becoming increasingly important. For e-discovery, this requirement is primarily upstream from Clearwell where other solutions will be used to perform an initial cull to make the number of records more manageable.

Overall, the features bundled with Transparent Search look very compelling and are just part of the many capabilities Clearwell offers. Organizations facing high stakes E-Discovery should consider the Clearwell E-Discovery Platform for their needs.

Website: clearwellsystems.com

Pricing: E-Discovery Platform is $65,000 per 100 GB plus maintenance and support

Management: Aaref A. Hilaly, President & CEO

Funding: $29+M (C: $17M, August, 2007; B: $12M, April, 2006; A, January, 2006?)

Investors: DAG Ventures, Redpoint Ventures, Sequoia Capital

1) Multi-lingual support

The ability to support multiple languages is becoming more important for E-Discovery and Axcelerate eDiscovery addresses this with:

• automated language detection
• double-byte language support
• ability to index, cull, filter, search, review, and code documents in different languages

For automated language detection, it would be useful to have some information on how Recommind handles multi-lingual documents. Additionally, it would be useful to know what kind of word segmentation techniques they are using for character-based, compounding, and agglutinative languages. This is useful in establishing defensibility of the search as some techniques such as morphological analysis may lead to false negatives which would be problemaitc for eDiscovery.

2) Filtering Improvements

Filters can be automatically populated, providing suggestions of what to use. Selecting documents across filters can also be performed. To save time, type ahead technology, which has been showing up in many products, including Google’s homepage, is also included.

3) Native file viewing

Hundreds of file types are mentioned but the feature is not described. Certainly hundreds of text-based file types can be viewed in a text editor, but does viewing other file types require the original application, e.g. does viewing a MS Word doc require MS Word or other third-party application.

4) Processing improvements

Processing improvements include the ability to filter, cull and de-duplication within Axcelerate.

5) Indexing performance improvements

Faster indexing performance is mentioned without specifics.

Spokespersons and journalists:

The follow people have provided quotes and/or covered Axcelerate 3.0:

• Quotes
  • Robert Tennant, CEO, Recommind
  • Craig Carpenter, VP Marketing, Recommind
  • Vivian Tero, Program Manager, Compliance Infrastructure, IDC
• Journalists
  • Phil Muncaster, vnunet.com

Website: recommind.com

Management: Robert Tennant, CEO

Funding: $7.5M (Latest: $7.5M, October 2008)

Investors: Kennet Partners