Cloud Search Is Important, But Only A Piece Of The Enterprise Search Puzzle

by Barry Murphy

In an earlier post, I described the importance of having the ability to quickly search for information stored in the Cloud.  The post pointed out that Cloud search is somewhat more complicated than one might think at first glance because the speed of search is affected by how close the index lives to the actual data in the Cloud infrastructure.  One comment I received was that Cloud search can be fast and simple if the Cloud vendor promises a certain service level for query times and results.  That can address part of the issue around search (although IaaS providers – what we are truly talking about when we say “Cloud” – are typically not interested in guaranteeing SLAs for things like search because they allow customers to provision their infrastructure set to enable fast search with products like X1 Rapid Discovery).  Even if a Cloud vendor were to guarantee phenomenal search SLAs, the issue of unified enterprise search of all information still remains.

The reality is that enterprises and government agencies store information in “hybrid” environments that encompass on-premise systems within corporate data centers, virtualized systems that companies operate, and Cloud-based repositories.  Research firm Gartner predicts that by 2017, half of mainstream enterprises will have a hybrid cloud.  And, research from NetApp shows that organizations will be managing data across multiple cloud environments, not just a single provider, per se.

Click image to enlarge

Click image to enlarge

These are exciting developments.  As organizations embrace more modern infrastructures, there are many benefits to be had.  What we need to remember, however, is that business professionals still need to quickly find and take action on their information assets to do their jobs.  As that information gets further scattered, enterprise search will take on increased importance.  Workers don’t care if their data is stored on-premise or in the Cloud as long as they can quickly find it in an easy-to-use interface.

The challenge for today’s organizations is that information now lives in multiple infrastructures – on-premise, virtual, Cloud, or most frequently, a hybrid of all of these.  Current approaches to including Cloud-based data in enterprise search and eDiscovery require downloading a copy of the data to search so that it resides alongside other local content.  Unfortunately, that defeats the purpose of storing the data in the Cloud in the first place.

This takes me back to my original point:  Cloud search is very important.  But, Cloud search cannot simply exist in a vacuum.  An effective enterprise search solution will combine on-premise search capabilities that can talk to search in the Cloud – without requiring downloading the cloud-based information in order to search across all data.

Leave a comment

Filed under Cloud Data, Enterprise Search

Cloud Search: Not As Simple As You Think

By Barry Murphy

Corporations and Government agencies are moving data to the Cloud in droves.  No matter which analyst firm you look to on Cloud storage adoption, you will find consistent results:

  • Forrester Research reports that 40% of enterprises surveyed indicated they have already rolled out workloads on public clouds or have near-term plans to do so and that the number will increase to 50% this year.
  • IDC predicts that from 2013–2017 public IT cloud services will have a compound annual growth rate (CAGR) of 23.5%, five times that of the IT industry as a whole.
  • Gartner says Cloud Computing Will Become the Bulk of New IT Spend by 2016 and that spending on public Cloud services will have a CAGR of 17.7% from 2011 – 2016, with spending on Infrastructure-as-a-Service (IaaS) itself will have a CAGR of 41.3% in that time period.
  • In eDJ Group’s recent Cloud services adoption fast poll, Greg Buckles found that less than 5% of respondents reported that all information is kept on-premise on company infrastructure and cloud services are not being actively considered.

Cloud-icon_magnifying-glassNo matter where data is being stored, though, the fact remains that the ability to search that data will be critically important.  Workers still demand unified access to email, files, and SharePoint information, and they want fast-as-you-type search results regardless of where the data lives.  In addition, Legal teams require that search queries and collections execute within specific time-frames.  But, Cloud search is slow, as indexes live far from the information.  This results in frustrated workers and Legal teams afraid that eDiscovery cannot be completed in time.

Lest you think this is not a big deal, consider the following story.  When I was at eDJ, we worked with a very large enterprise client that wanted to move its collaboration system to the Cloud.  The problem was that the Cloud system the client was contracting with could not meet the Legal Department’s requirements for speed of query results and collection.  This significantly slowed down the movement to the Cloud until the client had worked with the Cloud vendor to ensure that search and collection could execute at the necessary speeds.  The delay frustrated an IT team anxious to reap the promised benefits of the Cloud and cost the project team significant man-hours.

This story highlights the need to granularly define search and eDiscovery requirements before moving data to the Cloud.  Most “cloud search” solutions pass queries through connectors, and then the Cloud vendor needs to figure out where in its vast data center the index lives, find the content, return the query result, and then the customer will need to download all the content.  The result is a slow search and another copy of the data downloaded on premise, which basically defeats the purpose of moving to the Cloud in the first place.

If a customer wanted to speed up search, it would have to essentially attach an appliance to a hot-air balloon and send it up to the Cloud provider so that the customer’s index could live on that appliance (or farm of appliances) in the Cloud providers data center, physically near the data.  There are many reasons, however, that a Cloud provider would not allow a customer to do that:

  • Long install process
  • Challenging pre-requisites
  • 3rd party installation concerns
  • Physical access
  • Specific hardware requirements
  • They only scale vertically

The solution to a faster search is a cloud-deployable search application, such as X1 Rapid Discovery.  This creates a win-win for Cloud providers and customers alike.  As enterprises move more and more information to the Cloud, it will be important to think about workers’ experiences with Cloud systems – and search is one of those user experiences that, if it is a bad one, can really negatively affect a project and cause user revolt.

 

Leave a comment

Filed under Cloud Data, Enterprise eDiscovery, Enterprise Search, Information Access, Virtualized Environment

Search as a Desktop Virtualization Enabler

Desktop_virtualizationby Barry Murphy

 

Too often, search is taken for granted.  When I first started doing research on eDiscovery in the cloud, the prevailing attitude was, “as long as information is searchable, eDiscovery is taken care of.”  Sadly, many organizations have learned the hard way that it is not that easy.  There is much more to search than meets the eye.  But, most organizations do not figure that out until it is too late – until search does not work in the desired manner or at the required speed.

eDiscovery is not the only area where search is overlooked and becomes an issue.  In fact, search is a critical function for today’s knowledge worker.  Despite the importance of information access, unified search of workers’ most critical assets (email, files, desktop content, and SharePoint) is not always a huge requirement of IT organizations.  It is to end-users, however, and that is one of the reasons that X1 has had such success with the Search 8 product – it has a user-friendly interface that provides simple, fast access to the information assets users need the most.

The lesson that I have taken away from being involved in the search market is that search as a standalone application may not seem sexy, but it provides a real return on investment.  It also allows organizations to ensure that investments in other technologies are optimized.  This fact can be seen especially in virtual desktop (VDI) environments.  Desktop virtualization promises many benefits: lower IT costs; streamlined administration of IT assets; and end-user flexibility in terms of accessing the desktop from anywhere.  Given the popularity of BYOD, the consumerization of IT, and the need for mobility to support telecommuting, VDI is becoming more and more important.

It is the little details of IT projects, however, that can have big impact on results.  Some organizations find that the cost savings anticipated from VDI are less than expected because of high disk resources necessary to support Windows indexing on the virtual desktop.  Or, best practice is followed and Windows indexing is turned off – and then users are unable to search for information on their desktops.  There are two possible outcomes from this, and both are bad:  either users are rendered unproductive because they cannot easily find information or they simply reject the virtual desktop and find ways around the system.

In order to ensure that VDI deployments meet expectations, organizations can build unified search into requirements early on.  At the very least, this will help to ensure that end-users are more receptive to the virtual desktop and allow them to remain productive.  Getting end-users to buy in is often half the battle when deploying new technology.  As I mentioned, though, search is often an afterthought – an issue that only comes up after a VDI deployment where end-users complain or reject the solution outright.  That is why it is important to make search a requirement early on.

When it comes to VDI environments, a good search solution must decouple the search UI from the indexing service.  Otherwise, indexing will require virtual desktop computing resources and cut into VDI cost savings.  The goal is to minimize the RAM usage and search client footprint on the virtual desktop.   It sounds simple, but traditional search solutions are not architected for this.  We at X1 are doing a webinar with Citrix on this very issue – enabling lightning-fast search in VDI environments.  The webinar is on April 10, 2014 at 1pm ET / 10am PT.  Please click here if you would like to join us to learn how to use search to enable optimization of desktop virtualization deployments.

Leave a comment

Filed under Desktop Search, Enterprise eDiscovery, Information Access